arc456.files.wordpress.com · Contents Authors and Title Page Preface v Foreward vi M. Feroze Ahmed and K. Nurul Ashfaque Sanitation and Solid Waste Management in Dhaka City During

Engineering Concerns of Flood A 1998 Perspective

M. Ashraf Ali Department of Civil Engineering Bangladesh University of Engineering and Technology, Dhaka, Bangladesh Salek M. Seraj Department of Civil Engineering Bangladesh University of Engineering and Technology, Dhaka, Bangladesh Sohrabuddin Ahmad Department of Civil Engineering Bangladesh University of Engineering and Technology, Dhaka, Bangladesh

Bangladesh University of Engineering and Technology, Dhaka-1000, Bangladesh

Published by: Directorate of Advisory, Extension and Research Services Bangladesh University of Engineering and Technology, Dhaka, Bangladesh August 2002 ISBN 984-823-002-5 Copyright © 2002 by the Publisher. All rights reserved. No part of this publication may be reproduced, stored in a retrieval system or transmitted, in any form or by any means – graphic, electronic, mechanical, photocopying, recording or otherwise, without prior permission of the Publisher. Book and Cover Design by M. Ashraf Ali and Salek M. Seraj

ii

Contents Authors and Title

Page

Preface

v

Foreward

vi

M. Feroze Ahmed and K. Nurul Ashfaque Sanitation and Solid Waste Management in Dhaka City During the 1998 Flood

1

Farooque Ahmed, M. A. Jalil, A.B.M. Badruzzaman and M. Ashraf Ali Development of Low-cost Technologies for Treatment of Contaminated Water in Flood Affected Areas

15

Md. Mujibur Rahman, Md. Mafizur Rahman and Md. Delwar Hossain Assessment of Water Quality in Flood Affected Areas of Dhaka City

35

A.B.M. Badruzzaman and Md. Rezwan Khan Purification of Floodwater by Electrolysis

49

A.S.M. Abdul Quium and S.A.M. Aminul Hoque The Completeness and Vulnerability of Road Network in Bangladesh

59

S. J. Md. Yasin, Md. Hossain Ali, Tahmeed M. Al-Hussaini, Eqramul Hoque and Sadik Ahmed Effect of Submergence on Subgrade Strength

77

M. J. B. Alam and M. Zakaria Design and Construction of Roads in Flood Affected Areas

91

M.A. Ansary and A.M.M. Safiullah Effect of Flood on Earth Structures: A Case Study

101

Abu Siddique, Md. Hossain Ali, Md. Shariful Islam and Md. Monwarul Islam Evaluation of Performance of Rajshahi Town Protection Embankment

115

S. J. Md. Yasin, Md. Hossain Ali, Eqramul Hoque, Sadik Ahmed and Rezaul Hoq Grain Size Distribution of Materials Deposited Over Floodplain Due to Embankment Failure

131

Moazzem Hossain, Alamgir M. Hoque and M. Zakaria Assessment of Flood Damage to Roads in and around Dhaka City and Remedial Measures

141

M. J. B. Alam, Alamgir M. Hoque and Md. Mazharul Hoque Assessment of Economic Loss Caused by Flood Damaged Transportation Network

151

Abdur Rahim and M. Reaz H. Khondoker Assessment of Flood Damages to Inland Water Transport Sector of Bangladesh

161

M. J. B. Alam and Md. Hossain Ali Concept of Flood Shelter to Cope with Flood

175

iii

K. M. Maniruzzaman and B. M. Alam A Study on the Disaster Response for Shelters During the 1998 Flood in Dhaka City

187

M. Mozzammel Hoque, Sujit K. Bala, Syed Mohib Uddin Ahmed, M. Anisul Haque and Saifullah Al Mamun Impact of the 1998 Flood on the Morphology of Rivers Around Bridges

201

M. Mozzammel Hoque, Syed Mohib Uddin Ahmed and Md. S. A. Hossain Impact of the 1998 Flood on Groundwater Recharge in Dhaka

213

A.K.M. Saiful Islam and Jahir Uddin Chowdhury Hydrological Characteristics of the 1998 Flood in Major Rivers

227

Anisul Haque, Mashfiqus Salehin and Jahir Uddin Chowdhury Effects of Coastal Phenomena on the 1998 Flood

241

A.F.M. Saleh, S.M.U. Ahmed, M. Mirjahan, M.R. Rahman, M. Salehin and M.S. Mondal Performance Evaluation of FCD/FCDI Projects During the 1998 Flood

253

Mohammad Rezaur Rahman and Jahir Uddin Chowdhury Experiences with Flood Management Practices During the 1998 Flood

267

M.J.B. Alam, M. H. Rahman and Md. Mujibur Rahman Remote Sensing Imagery to Assess the Environmental Impacts of Flood

281

Sarwar Jahan The Socio-Economic Impacts of the 1998 Flood in Dhaka City

289

Mohammad A. Mohit and Shakil Akther Delineation of Flood Damaged Zones of Dhaka City Based on the 1998 Flood by Using GIS

303

Md. Ehsan, Md. Imtiaz Hossain, Md. Nasir Uddin Miah and Md. Abu Sayed The Role of Small Diesel Engines in Rural Bangladesh During the 1998 Flood

319

Abu Md. Azizul Huq, Md. Imtiaz Hossain, A.K.M. Sadrul Islam, Md. Arif Hasan Mamun, Mosfequr Rahman and Md. Obaidul Gani Damage and Productivity Loss in Industries During the 1998 Flood

331

Md. Quamrul Ahsan and M. Alam Solar Powered Lantern for Flood Affected Areas

343

Salek M. Seraj and Md. Rezaul Karim Effect of the 1998 Flood on Non-Engineered Structures

363

iv

PREFACE

This book Engineering Concerns of Flood has been designed as a reference work on various engineering aspects related to the causes and consequences of flood, which is a very common natural hazard in Bangladesh. Bangladesh is a small country of 147570 sq. km., but has a very large population of about 130 million, majority of whom are poor. The economic condition of Bangladeshi people makes them very vulnerable to the adverse effects of natural forces. It is, indeed, the vulnerability of the poor people of Bangladesh, coupled with the regularly occurring natural hazards like floods, that causes disasters during and at the aftermath of such events. Lack of proper planning and preparedness, poor quality of infrastructure and their inadequate maintenance also add to peoples’ sufferings and magnitude of loss.

This book is an outcome of a systematic research initiative that was undertaken by the academics of Bangladesh University of Engineering and Technology (BUET) during and after the 1998 flood that lasted for several months causing havoc to almost every spheres of life. All the research works reported in this book were funded by the Committee for Advanced Studies and Research (CASR) of BUET. A total of 28 research projects covering varied aspects of flood were successfully completed and this book is an edited and summarized compilation of these research works.

The papers presented in this book cover a wide range of topics encompassing almost all the possible engineering, socio-economic and planning aspects of flood. Papers addressing both urban and rural issues have been included in this book. Environmental issues like water quality, water purification, sanitation and solid waste management have been covered. Different aspects of communication and transport sector have been covered including the vulnerability of existing road network, design aspects of roads including effect of inundation on subgrade strength, and economic losses in both road and water transport sectors. Papers analyzing the effects of flood on flood protection embankments, groundwater recharge, morphology of rivers, and performance of flood control and drainage projects have been included in this volume. Damages and losses in various sectors including those in industrial sector have been presented. Engineering concerns and prospects related to non-engineered rural houses, alternate power supply as well as role of remote sensing and imagery to assess the environmental impacts of flood have been covered in this book.

Flood is a major problem for Bangladesh. The present book comes out at a difficult time, as the country appears to be heading towards another flood. The lessons learnt from the 1998 flood through the research initiative at BUET, as summarized in this book, may help engineers, planners and policy makers to face future floods more effectively and minimize the disastrous effects.

M. Ashraf Ali, Salek M. Seraj and Sohrabuddin Ahmad, Editors

v

FOREWARD I am very happy to note that the Directorate of Advisory, Extension and Research Services (DAERS) is publishing the compilation of research works conducted at BUET on various engineering as well as socio-economic aspects of flood. Right after the devastating flood in 1998, the BUET authority under the then Vice Chancellor Professor Dr. Iqbal Mahmud launched a research program on various facets of flood. At the end of the respective studies, researchers of BUET submitted a total of 28 research reports. Engineering Concerns of Flood: A 1998Perspective is the edited and summarized version of all the studies that were conducted under the funding provided by the Committee for Advanced Studies and Research (CASR) of BUET. I have my special thanks to Professor Iqbal Mahmud for initiating the research on such an important issue like flood, which affects the life and economy of Bangladesh in many ways. My thanks go to the editors who have successfully completed an otherwise very difficult job.

I am confident that the contents of Engineering Concerns of Flood will form a strong foundation for further research and help Bangladesh combat and cope with natural calamities in the future. I wish this book will be of immense benefit to the people and to the Government of Bangladesh. Professor Dr. Nooruddin Ahmed Vice Chancellor Bangladesh University of Engineering and Technology Dhaka 1000, Bangladesh

vi

M. A. Ali, S. M. Seraj and S. Ahmad (eds): ISBN 984-823-002-5

Sanitation and Solid Waste Management in Dhaka City During the 1998 Flood

M. Feroze Ahmed and K. Nurul Ashfaque

Department of Civil Engineering Bangladesh University of Engineering and Technology, Dhaka-1000, Bangladesh Abstract The degradation of environmental quality in the city of Dhaka was a major concern during the flood of 1998. The damages and disruption of sanitation and solid waste disposal facilities were mainly responsible for the deterioration of environmental quality of the city. This paper presents an overview of the sanitation and solid waste management facilities in Dhaka city, functioning of these essential facilities during flood and the alternative arrangements adopted by flood affected people. Water pollution, drinking water contamination and prevalence of diarrhoeal diseases due to failure of sanitation and solid waste management systems during the 1998 flood have also been presented in this paper. INTRODUCTION Flood in Bangladesh is an annual event, but flood of 1998 had been the most severe and devastating one in recent years. The peak flood levels of 1988 around Dhaka city were higher than those of 1998 but the severity of the flood 1998 was due to its longer duration. The maximum duration of flood above danger level at Mirpur in 1988 was 30 days, but in 1998 the duration was 69 days at the same location. The danger level at Dhaka is considered as 6 m above mean sea level as most of the urbanized areas of Dhaka city lie at elevations between 6 and 8 m. The area between the flood embankment and floodwall in the west and Pragoti Sarani in the east is considered as flood protected. But in real situation a large area within the flood-protected area was flooded due to poor management of

Engineering Concerns of Flood 1


2 Engineering Concerns of Flood

flood protection and drainage facilities. An estimated 150 km2 out of the total 275-km2 area of Dhaka Metropolitan area was inundated during the 1998 flood.

Sanitation and solid waste management are essential service facilitates in densely populated urban centers. Safe and hygienic disposal of wastes is required for healthy living. Proper sanitation and solid waste management improve environmental quality and lack of these facilities results in a filthy condition detrimental to health and human well being. The population coverage by proper sanitation has been correlated with prevalence of disease, death and debility, particularly mortality of children and is considered as an indicator of development. Inundation of the sanitation and solid waste management facilities cause severe water pollution. Contaminated water in the surroundings due to damage and disruption of these essential facilities opens many routes of transmission of diseases. As a result, concern of outbreak of diseases, particularly of diarrhoeal diseases, is expressed in every flood and activities related to preventive and curative interventions are geared up towards effective control of such incidences.

The devastating flood of 1998 caused severe damages and disruptions to infrastructure. Disruption of these essential service facilities had serious implications on degradation of environmental quality and public health in the densely populated Dhaka city. Submergence of sewerage facilities, back flow of sewage and decomposition of garbage were mainly responsible for degradation of environmental quality particularly for deterioration of physico-chemical and bacteriological quality of floodwater and contamination of drinking water supplies. The extent of damage and disruption of sanitation and solid waster management system, alternative arrangements adopted by flood-affected people and its consequences were major environmental issues of the flood. This paper attempts to make an evaluation of some of these environmental concerns of the 1998 flood. MATERIALS AND METHODS The assessment of sanitation and solid waste management situation in flood affected areas of Dhaka City during the 1998 flood involved data collection from secondary sources, questionnaire survey to acquire primary data and analysis of data to synthesize the major findings. The available data from relevant organizations like DWASA, DOE, DCC, etc. were collected and reviewed for relevant information. These data primarily reflected the pre-flood conditions. The news items on flood in the news media were most important source of information about flood and these were regularly reviewed during the flood.

Sanitation and Solid Waste Management During Flood

A questionnaire survey was conducted to acquire data regarding prevailing situation during the flood. The questionnaire originally prepared for the survey was tested in the field and some modifications were made on the basis of field inputs. In view of time constraint, the survey was mainly conducted in the severely affected eastern part of Dhaka city. Since this survey was conducted mainly in the flood-affected areas, the sanitation situation during the flood has been reflected in the findings.

The questionnaires survey to acquire information on solid waste disposal practices of 68 households during and before flood was conducted in the flood affected areas. The most important water quality parameters of samples collected during the flood and post-flood conditions were analyzed. The Faecal Coliform (FC) counts of 116 drinking water samples collected from various locations during the flood were compared with pre-flood data. The Biochemical Oxygen Demand (BOD), Dissolved Oxygen (DO) and Faecal Coliform counts of a large number of floodwater samples were analyzed. The main purpose of analysis of drinking and floodwaters was to assess the degree of water pollution by the discharge of sewage and solid waste in floodwater. SANITATION

Sanitation Facilities and Practices

The sanitation system of the city of Dhaka is comprised of waterborne sanitary sewerage system, combined sewer system, small bore sewer (SBS) system, septic tank with soak well and pit latrine. The areas under different sanitation system have been shown in Fig.1. The sewer system of Dhaka was originally designed as separate systems i.e., sanitary sewer system for conveyance of wastewater from toilet (sanitary sewage) and storm sewer system for the discharge of storm and surface drainage (storm sewage). In the combined system both sanitary sewage and storm waters are conveyed through same sewer line. The SBS system receives effluent from septic tank for disposal by suitable means. At present, as shown in Fig.1, the sewered sections of the city is being covered by separate, combined and SBS systems.

The sanitary sewage is collected and transported by a network of sanitary sewer system having 23 sewage lifting stations to convey the sewage to treatment plant at Pagla. To prevent bacteriological contamination, the effluent of the treatment plan is chlorinated before disposal in the river Buriganga. The Waste Stabilization Ponds system at Pagla is capable of partial removal of pollution load from sewage. In the combined sewer area, sanitary sewage is discharged in the storm sewer in the absence of sanitary sewers. The sewage conveyed by combined sewer is discharged in water bodies in low-lying areas, natural khals and rivers with storm water for natural degradation. The effluent of the SBS




system is also discharged in the low-lying area like storm drainage. Septic tank with soak well system is a very good sanitation initiative at the household level and covers largest number of population in the city.

Figure 1: Area covered by different sanitation systems in Dhaka


The top pre-consolidated clay layer beneath Dhaka City is not good for soak away of septic tank effluent. Many inhabitants have connected the effluent to surface drains or storm sewers where available to avoid overflow within the premises. This practice contributes to the deterioration of environmental quality in the locality and pollution of surface water sources. Sanitary pit latrines are low-cost sanitation system practiced in the low-income squatters and slums

Sanitary sewerage system covers about 18 % of the population and in the un-sewered section septic tank system covers 40 percent and sanitary pit latrines cover 15 percent of the population. The population coverage by SBS has been reported under septic tank system. The remaining 27 percent people of Dhaka city mainly living in urban slums do not have any acceptable sanitation system (MoLGRDC, 1994). Katcha Latrine is the common method of excreta disposal in the slum areas. It has been reported that some 87 percent of the inhabitants in slums and squatters in Dhaka city rely on low quality katcha latrines (CUS, 1988). The transient nature of settlement and poor awareness of health-sanitation relationship deter building of sanitary latrines in squatters and slums. Limited public sanitation facilities have been constructed by Dhaka City Corporation for use on payment basis. About 18 such facilities with latrine, shower and lavatories exist in densely populated areas in Dhaka city.

Impact of Flood on Sanitation

In the flooded area, all sanitation facilities were not functional. The sewer and storm water drainage systems were not functioning because of negative hydraulic gradient which in some cases caused back flow of sewage mixed flood water in the flood protected low lying areas. The only Sewage Treatment Plant (STP) situated at Pagla within the DND flood protected area has been reported to be functioning with a small sewage flow from old Dhaka. The main sewer trunk line leading bulk of sewage to STP was closed down to avoid flooding of DND area and damages to this large diameter brick sewer line. A total of 13 sewage-lifting stations out of 23 were closed because of entry of floodwater. Flood also caused severe damages to sewer lines and manholes. The repair and rehabilitation cost of DWASA sewerage system was estimated to be Tk. 620 million (DWASA, March, 1999). In the flooded areas, the contents of the pit latrines were mixed up with the floodwater and the flooded septic tanks and soak wells were oozing out their additional loads in floodwater. Back flow of faecal matter in flooded ground floor through toilets was common in the flooded areas. The upper floor residents in multistoried buildings were able to use the toilets but the additional loads caused overflow of sewers or septic tanks in the flooded areas.

During the flood, the problems encountered by the different socio-economic groups of the people were different. The sanitation facilities of the households surveyed during and before the flood are presented in Fig.2. It has been observed that sanitation facilities of all the households were more or less affected but 64




percent of the households were able to continue their usual practice by elevating level or adopting alternative facilities. About 26 percent of the household adopted the practice of open defecation and 10 percent used community latrines and other temporary facilities erected during the flood.

64%

26%

10%UsualSanitation

OpenDefecation

CommunityLatrine

13%

33%44%

7% 3% SewerSystemSeptic Tank

Pit Latrine

No FixedPlacesCommunityLatrine

Sanitation before flood Sanitation during flood

Figure 2: Sanitation practices during and before flood

In medium-income areas where floodwater depth was such that the ground floor was submerged, the toilets were not useable at ground floor level. The residents of the ground floor of multistoried houses were compelled to use toilets of the upper floors. A sanitation practice by erecting temporary toilet on the roof due to flooding of ground floor is shown in Fig. 3. In low-income areas, the people were using pit latrines. These people had made temporary arrangement by erecting elevated platform for defecation thus releasing the faecal matter directly in floodwater.

Figure 3: Temporary latrine on roof in flooded area


SOLID WASTE MANAGEMENT

Solid Waste Management System in Dhaka City

The quantity of solid waste generated in the Dhaka Metropolitan area is about 3,500 Metric tons at the rate of 0.5 kg. per capita per day (Bhuiyan, 1999). About 49 percent of the waste originates from domestic source, 21 percent from commercial, 24 percent from industrial and 6 percent from hospital and clinical sources in the city. The responsibility of the individual households is to put their waste in the roadside dustbins, containers or garbage accumulation centers for collection and transportation to final disposal points by Dhaka City Corporation (DCC). To improve the situation local individuals and groups have come forward to organize their own collection system. There are about 130 local initiatives in 90 wards of the city to provide primary waste collection services from 10,000 households (Kazi, 1999). The garbage under these initiatives is collected from individual households and carried to the nearest DCC bins/containers.

Extensive unorganized recycling takes place in the city of Dhaka at various stages of collection and disposal system. In the first stage, the households set aside the waste of market value and sale it to the street Hawker and in the subsequent stages of recycling all wastes of immediate market value are picked up by Tokais from the dustbins and disposal sites. The salvaged wastes are processed and sold to the old materials' shops for recycling.

The waste accumulated in the dustbins is collected by DCC trucks for final disposal at the dumping sites. Dhaka City Corporation (DCC) can collect only about 52 percent of the total waste generated in the city. About 11 percent of the waste generated is recycled and the remaining waste goes into backyard, road sides drains and open spaces. The quantities of solid waste disposed off or recycled by different routes are shown in Fig.4. The solid waste collected by DCC is finally dumped at landfill sites located in low-lying area by open dumping. The three waste dumping sites at Matuali, Mirpur and Lalbagh were in operation during pre-flood periods. DCC practices crude dumping of solid waste at landfill sites in order to keep the disposal cost low at a great risk to soil and groundwater pollution. Solid Waste Management During Flood

Solid waste is collected and transported by DCC in open trucks, closed trucks and container carriers. In addition to stationary dustbins, DCC has presently employed demountable hauled container system. Solid waste collection and transportation operations were totally disrupted in the inundated areas during the 1998 flood. The operation of vehicles over the flooded roads and within congested areas became increasingly difficult and finally solid waste collection and transportation from deeply flooded areas were discontinued. With the increase in flood depth, the operation of container system of collection was also




suspended and the water soaked garbage in containers remained in place. A completely filled container located at the side of a road is shown in Fig 5.

Dumped by DCC

Backyard & Landfill

Road, Drain & Openspaces

Recycle by Tokai

Recycle by Households1800 ton900 ton

100 ton300 ton

400 ton

Figure 4: Solid waste disposal and recycling routes in Dhaka City

Figure 5: A completely filled solid waste container

The area of Dhaka City inundated by the 1998 flood is about 150 km2.

Sanitation and solid waste management facilities in these flooded areas were practically non-functional. The solid waste management service facility of DCC is extended over 350 km2. It is estimated that about 43 percent of solid waste could not be collected and transported by DCC during peak flood period

Primary collection involving accumulation of solid waste in the nearest dustbin is the responsibility of households. Hence it is completely dependent on people’s attitude and participation. The practice of placing the solid waste in the right manner and in the right place had never been satisfactory. The dumping of solid waste in floodwater became indiscriminate because of absence of solid


waste collection system during flood. Soon garbage and polyethylene bags become a part of floodwater in many places. Dropping of garbage filled polyethylene bags in floodwater from upper floors of multistoried buildings was a common sight in flood-affected areas.

The results of the survey conducted in flood affected areas on solid waste disposal practices before and during flood are shown in Fig. 6. The results show that before the flood, 44 percent of the people used municipal dust bins for solid waste disposal, 12 percent used the good practice of van collection system, while the remaining people had the habit of open and backyard disposal. The survey results in Fig. 6 also show that during the flood the unhygienic practices of open dumping and throwing in flood were the modes of solid waste disposal of 82 percent of the people. Solid waste was also found to be dumped on a submerged access road to waste disposal site at Kazla.

In some residential areas, people selected an open plot of land with boundary walls as temporary garbage dump for the locality. The plot was used as a large bin for accumulation of solid wastes throughout the flood period for collection by DCC after the flood. This practice of solid waste accumulation innovated and adopted by local people helped prevention of littering of waste in the locality. However, the occurrence of septic condition in floodwater in the waste accumulation center by decomposing waste could not be avoided.

Practices before flood Practices during flood

Figure 6: Solid waste disposal practices before and during flood DEGRADATION OF ENVIRONMENTAL QUALITY

Pollution Problems

The organic fraction of the solid waste dumped in the stationary as well as hauled containers under favorable temperature and water content started decomposing within days. The anaerobic process of decomposition produced bad smell in the surrounding areas and the leachate produced by the soaked waste was released in

9%3%

9%

73%

6% MunicipalDustbin

VanCollection

OpenDisposal

Throwing inWater

DistantDisposal

44%

12%

41%

3% MunicipalDustbin

VanCollection

OpenDisposal

Others




the floodwater. The accumulated waste in confined area also decomposed through anaerobic process causing blackish color and bad odor to water. The waste discharged in open floodwater caused depletion of dissolved oxygen in floodwater. Similar effects were exerted by sewage and human excreta that entered into floodwater due to disruption of sanitation facilities. The faecal bacterial load in floodwater was primarily added to flood water by human excreta.

The failure of sanitation system caused severe pollution of floodwater. Faecal matters in sewage and decomposing garbage were mainly responsible for contamination of floodwater. The degree of contamination was assessed by physical, chemical and bacteriological analysis of floodwater from different locations. The DO and BOD of floodwater are presented in Figs. 7 and 8 respectively (Ahmed and Ashfaque, 1999). It may be observed that 38 percent of the samples showed a DO lower than 4 mg/l of which septic condition prevailed in 4 percent of the samples. About 20 percent of the samples showed a BOD higher than 50 mg/l, which is equal to one-sixth of the usual strength of domestic sewage in Dhaka City.

0102030405060708090

100

0 2 4 6 8 10Dissoled Oxygen, mg/l

Perc

ent l

ess

than

the

stat

ed

conc

entr

atio

n

Figure 7: Dissolved oxygen content of floodwater

The bacteriological quality of floodwater indicated by Faecal Coliform (FC) count is shown in Fig.9 (Ahmed and Ashfaque, 1999). The FC distribution indicates that 50 percent of the sample had a count of more than 180/ ml. The minimum FC was found to be 5/ml or 500/100ml. The presence of FC in high concentrations is a definite indication that floodwater was polluted by faecal matters of domestic sewage. The bacteriological quality of water of the DWASA water supply before and during flood indicated by Faecal Coliform (FC) count has been shown in Fig.10. Contamination of drinking water supplied by Dhaka


WASA occurs due to leakage in the pipeline. Contaminants enter into the water distribution main through leaks due to negative (suction) pressure during non-supply hours (Ahmed, 1998). The Dhaka Water supply as shown in Fig. 10 is contaminated to a certain extent but the situation further deteriorated during flood. The bacterial contamination of water was within 0 – 100 FC/100 ml before flood but the number in some of the samples was found to be innumerable to count by Membrane Filter Technique. The median level of FC count increased from 20/100ml before flood to 65/100ml during flood. The bacteria-free water samples decreased to 4 percent during flood from 20 percent during pre-flood period.

0

20

40

60

80

100

120

0 50 100 150 200Biochemical Oxygen Demand, mg/l

Perc

ent H

ighe

r tha

n th

est

ated

val

ue

Figure 8: Dissolved oxygen content of floodwater

0

10

20

30

40

50

60

70

0 50 100 150 200 250 300 350 400

FC Count, Number/ml

Perc

ent l

ess

than

the

stat

ed

Con

cent

ratio

n

Figure 9: FC counts of floodwater in Dhaka City




0102030405060708090

100

0 25 50 75 100 125 150 175 200FC Count, Number/100mg

Perc

ent l

ess

than

the

stat

ed

Con

cent

ratio

n

Pre-flood

Duringflood

Figure 10: FC count of DWASA water supply during and before flood

Diarrhoeal Diseases

The prevalence of diarrhoeal diseases in the flood-affected areas is a major concern. The incidence of diarrhoeal diseases in the flood affected areas showed that about 26 percent of the people suffered from diarrhoeal diseases during flood, of which 22 percent had diarrhea once. At the national level, death among the diarrhoeal patient was reported to increase to 0.09 percent i.e., 0.9 death per 1000 cases of diarrhoeal attack. In an epidemic situation the death rate is 1 percent according to international standard (The Independent, 18 September, 1998). It appears that outbreak of large scale diarrhoeal epidemic was under control due to awareness of the people, massive health campaign and distribution of safe drinking water and good management of diarrheas.

CONCLUSIONS The sanitation and solid waste management systems of Dhaka City were in most deplorable conditions during the 1998 flood. The sewer system, sewage treatment and disposal facilities and solid waste collection, transportation and disposal systems of Dhaka city were severely disrupted. All sanitation facilities in the flooded areas and partly in non-flooded areas were non-functional. The katcha and pit latrines were flooded, sewage was oozing out of septic tanks, soak wells and manholes. Solid waste collection and disposal in the deeply flooded areas were discontinued

Repair and rehabilitation of 10 sewage lifting stations, a part of the trunk line leading to Pagla sewage treatment plant, 500 km sewer line and large number of


manholes in the flooded areas were required after the flood. The estimated cost of reconstruction and rehabilitation of infrastructures of DWASA sewerage system is Tk. 620 Million. The removal of decomposed and partly decomposed scattered solid wastes in the flooded areas was a gigantic post-flood task for DCC. The unused hand trolleys and waste containers were damaged mainly due to corrosion during the flood.

People in the flood-affected areas adopted different sanitation and solid waste disposal practices. About 26 percent of people adopted the practice of open defecation and 10 percent used community latrines and other temporary facilities erected during the flood. Most sanitation practices released part of the human excreta in floodwater. People practiced indiscriminate discharge of solid waste in floodwater to control accumulation in fixed places during the flood.

The waste released in floodwater caused severe pollution problem. The pollution of floodwater by overflow of sewage and decomposing garbage could be fully prevented. Proper covering of septic tanks and manholes, installation of community latrine and accumulation of solid wastes in fixed places could reduce water pollution to some extent. It has been observed that secondary interventions such as provision of safe water and hygiene education campaign and management of diarrheas can help avoid spreading of water-related diseases.

REFERENCES Ahmed, M. F., (1998) “Problems of Water Supply in Dhaka Metropolitan City”,

Presented at the Workshop on Problems and Prospect of Water Supply in Dhaka Metropolitan City, Dhaka, 17 January.

Ahmed, M. F. and Ashfaque, K. N., (1999) Assessment of Sanitation and Solid Waste Managemnet Situation in Flood Affected Area of Dhaka City, BUET, November.

Bhuiyan, M.S.H. (1999) Solid Waste management of Dhaka City, Paper presented at the seminar on Solid Waste Management, Dhaka, 9 October, 1999.

Centre for Urban Studies (CUS) (1988) “Slums and Squatters in Dhaka City” Report on a Survey Conducted for Dhaka City Corporation.

The Independent, (1998), 18 September DWASA (1999), Detailed Assessment of Damages Occurred and Rehabilitation

Required in Sewer System of Dhaka WASA Due to Recent Devastating Flood, Final Report, March.

Kazi, N. M., (1999), Capacity Building for Primary Collection of Solid Waste, Citizens Guide for Dhaka, Environment and Development Associates.

MoLGRDC, UNDP, UNICEF, UNDP-World Bank Water and Sanitation Program, (1994) “Bangladesh Situation Analysis Water Supply and Sanitation, September.



Development of Low-cost Technologies for Treatment of Contaminated Water in Flood

Affected Areas

Farooque Ahmed, M. A. Jalil, A. B. M. Badruzzaman and M. Ashraf Ali Department of Civil Engineering

Bangladesh University of Engineering and Technology, Dhaka-1000, Bangladesh Abstract Flood is a yearly event in Bangladesh and people in the flood-affected areas suffer from waterborne diseases due to non-availability of potable water. During floods, people are advised to treat water through boiling, alum dosing and chlorination. However, there are no clear guidelines for proper application of these processes. For the development of suitable water treatment technologies, five simple technologies were investigated. These were alum coagulation, chlorination, alum coagulation-chlorination, boiling and storing. The quality of locally available chlorine tablets and bleaching powders were also evaluated. Water samples were collected from street taps, hand-pump tubewells, stagnant floodwater and rivers during the 1998 flood. The experimental results showed that: (i) heating water just upto the boiling point was sufficient to destroy pathogens, (ii) chlorination was quite effective in killing pathogens at a chlorine dose of 0.5 to1.0 mg/L with 30 minutes contact time, (iii) alum dosing of 200 mg/L effectively reduced the turbidity of floodwater but it could not make the water bacteria free even at much higher doses, (iv) colored and turbid floodwater could be treated with simultaneous application of chlorination and alum coagulation. The required alum dose was 200 mg/L and the chlorine dose was 0.5 mg/L with 30 minutes settling time/contact time, and (v) the quality of the available chlorine disinfectant varied widely and the local chlorine tablets were better than foreign tablets. Based on these findings, recommendations were made for practical applications of the developed technologies.


Farooque Ahmed, M. A. Jalil, A. B. M. Badruzzaman and M. Ashraf Ali


INTRODUCTION Flooding and consequent human sufferings and loss of property is a common phenomenon in Bangladesh. Although flooding is a yearly phenomenon in Bangladesh, in some years, it becomes devastating causing widespread havoc and human misery. The 1998 flood brought such widespread havoc in much of Bangladesh. Human misery caused by flooding that year had probably surpassed all previous records because of its prolonged duration, reaching about two months for some parts of Bangladesh. Lack of potable water and sanitation facilities is responsible for the greatest human miseries in the flood-affected areas. In fact, most of the deaths during and immediately after floods have been linked to water borne diseases such as diarrhea, cholera, typhoid, etc.

In most areas of Dhaka, water is supplied through an age-old distribution system of the Dhaka Water Supply and Sewerage Authority (DWASA). During flooding, contaminated water from overflowing sewers and other sources enter into the water supply lines through numerous leaks in the distribution system. In addition, water in the underground reservoirs at many households becomes contaminated with overflowing polluted water. Many tubewells also become inoperable. Absence of proper facilities for excreta disposal, and disposal of solid and liquid wastes make the water quality situation even worse.

During flooding, people in the flood-affected regions were advised not to use DWASA supplied water or water from any other source (e.g., tubewell water and pond-water) without treatment. Common water treatment methods used by people in flood affected areas include (a) use of water disinfectants available in tablet forms in the market, and (b) boiling of water.

Effectiveness of the disinfectant tablets and their recommended doses has been widely questioned during flood episodes. Wide ranges of such tablets are available in the market and there is an urgent need to evaluate the effectiveness of these tablets. In almost all areas of Bangladesh, commercial alum and bleaching powder are available at the local markets. Commercial alum has been used in the past for purifying/disinfecting water although its effectiveness as a disinfectant and its optimum dose are not clearly known. Commercial bleaching powder is a common household disinfectant. These products could be effectively used for disinfecting water in the flood-affected areas. Thus there is an urgent need to determine the effectiveness and required dose of commercial alum and bleaching powder in disinfecting water during floods.

Boiling is a very popular means of disinfecting water at household level, especially where fuel is available. Often it is recommended that water be kept at boiling condition for at least 30 minutes before use. However, availability of fuel becomes a real problem during flooding and people usually find it difficult to boil water for a long time. In order to determine a reasonable time of water

Technologies for Treatment of Contaminated Water

boiling for disinfecting, there is a need to check the effect of boiling time and temperature on the quality of the treated water.

The overall objective of the project was to develop low-cost technologies for treating water at household or small community levels in flood affected areas. For this purpose, water samples were collected from (a) WASA supply line (from taps in household and street hydrant), (b) hand pump tubewells, and (c) from open water bodies in flood affected areas. The specific objectives of this study were: (i) to evaluate the effectiveness and to determine the required optimum dose of alum in treating different types of polluted water, (ii) to evaluate the effectiveness and to determine the required optimum dose of chlorine in treating different types of polluted water, (iii) to evaluate the effectiveness of combined alum coagulation and chlorination in treating different types of polluted water, (iv) to determine the effect of boiling time and temperature on bacteriological quality of water, (v) to evaluate the natural decay of pathogens during storage of water, and (vi) to evaluate the strength of commercial chlorine disinfectants available in the market. SIMPLE WATER TREATMENT METHODS Virtually all the surface water sources and the flood-contaminated supply water contain easily detectable turbidity. Smaller suspended particles cannot be efficiently removed by plain sedimentation. Colloidal suspensions, which usually constitute the major part of a flood-contaminated water, are more stable as they do not agglomerate naturally. Coagulation is performed to make these colloidal particles unstable and force them to agglomerate. Although coagulation, settling and filtration can remove approximately 90% of the bacteria and viruses, disinfection is performed to kill or render harmless, pathogenic microorganisms. Coagulation

The stability of colloidal suspension is principally governed by the large surface-to-volume ratio resulting from their very small size. In most natural waters the colloidal surfaces are negatively charged. When two colloids come in close proximity there are two forces acting on them. The electrostatic potential created by the halo of the counter ions surrounding each colloid reacts to repel the particles, thus preventing contact. The second force, an attraction force called the van der Waals force, supports contact. This force is inversely proportional to the sixth power of the distance between the particles and also decays exponentially with distance. It decreases more rapidly than the electrostatic potential, but is a stronger force at close distances.

In water treatment plants, chemical coagulation is usually accomplished by the addition of trivalent metallic salts such as Al2(SO4)3 (Aluminum Sulfate) or




FeCl3 (Ferric Chloride). Coagulation is accomplished by four mechanisms namely, ionic layer compression, adsorption and charge neutralization, entrapment in a flocculent mass, and adsorption and inter-particle bridging. Ionic layer compression: The quantity of ions in the water surrounding a colloid has an effect on the decay function of the electrostatic potential. A high ionic concentration compresses the layers composed predominantly of counter ions towards the surface of the colloid. If this layer is sufficiently compressed, then the van der Waals force will be predominant across the entire area of influence, so the net force will be attractive and no energy barrier will exist. Adsorption and charge neutralization: The nature of the ions is of prime importance in the theory of adsorption and charge neutralization. The ionization of aluminum sulfate (alum) in water produces sulfate anions (SO4

2-) and aluminum cations (Al3+). The sulfate ions may remain in this form or combine with other cations. However, the Al3+ cations react immediately with water to form a variety of aqua-metallic ions and hydrogen ions.

Al3+ + H2O AlOH2+ + H+

Al3+ + 2 H2O Al(OH2)- + 2 H+-

7 Al3+ + 17 H2O Al7(OH)174+ + 17 H+

Al3+ + 3 H2O Al(OH)3 + 3 H+

The aqua-metallic ions thus formed become part of the ionic cloud surrounding the colloid and, because they have a great affinity for surfaces, are adsorbed onto the surface of the colloid where they neutralize the surface charge. Once the surface charge has been neutralized, the ionic cloud dissipates and the electrostatic potential disappears so that contact occurs freely. Overdosing with coagulants may result in re-stabilizing the suspension. Sweep coagulation: The last product forming in the above equation is aluminum hydroxide, Al(OH3). The Al(OH3) forms amorphous, gelatinous flocs that are heavier than water and settle by gravity. Colloids may become entrapped in a floc as it is formed, or they may become enmeshed by its sticky surface as the flocs settle. The process by which the colloids are swept from suspension in this manner is known as sweep coagulation. Intra-particle bridging: Large molecules may be formed when aluminum or ferric salts dissociate in water. Synthetic polymers may also be used instead of or in addition to metallic salts. These polymers may be linear or branched and are highly surface reactive. Thus several colloids may become attached to one


polymer and several of the polymer-colloid groups may become enmeshed resulting in a settleable mass.

Generally, Jar test is performed to determine the optimum coagulant dose for particular water. Thus, it should be repeated if there is a significant change in the quality of water. The Jar tests are usually performed using a series of glass containers those hold at least 1 L and are uniform in size and shape. Normally, six jars are used with a stirring device that simultaneously mixes the contents of each jar with a uniform power input. Each of the six jars is filled to the 1-L mark with the water whose turbidity, pH, and alkalinity have been predetermined. One jar is used as a control while the remaining five are dosed with different amounts of coagulant at different pH values until the minimum values of residual turbidity are obtained. After chemical addition the water is mixed rapidly for about 1 minute to ensure complete dispersion of chemicals and then mixed slowly for 15 to 20 minutes to aid in the formation of flocs. The water is next allowed to settle for approximately 30 minutes, or until clarification has occurred. Portions of the settled water are then decanted and then tested to determine the remaining turbidity. Test results are then used to determine the quantity of coagulant to be used in water treatment. Chlorination

A good disinfectant must be toxic to microorganisms at concentrations well below the toxic thresholds to humans and higher animals. Additionally, it should have a fast rate of kill and should be persistent enough to prevent re-growth of organisms in the distribution system. Disinfectants include chemical reagents such as halogen groups, ozone, and irradiation with gamma waves or ultraviolet lights.

Chlorination may be applied to water in gaseous form (Cl2) or as salts [Ca(OCl)2, NaOCl]. The reactions in water are as follows:

Cl2 + H2 O H+ + HOCl + Cl-

Ca(OCl)2 Ca2+ + 2 OCl-

NaOCl Na+ + OCl-

The hypochlorous acid (HOCl) and the hypochlorite ion (OCl-) in the above equations are further related by the following relationship which is governed by pH and temperature. The sum of HOCl and OCl- is called the free chlorine residual and is the primary disinfectant employed.

HOCl H+ + OCl-




Boiling

Boiling is considered to be an easy and effective method for removal of pathogens from water. One of the most widely used household techniques, boiling involves heating the water for sometime at or around 100°C. Since almost all the pathogens in water survive and thrive in a temperature range of approximately 25-50°C, high temperature induces death in almost all the pathogens. However, there are some microorganisms that form spores and are resistant to high temperature. Thus, it is suggested that the water be boiled for a prolonged period. Storage for Natural Decay of Pathogens

Intestine of warm-blooded animals provide the optimum conditions for survival, growth and reproduction of pathogens. As soon as they are excreted and find their way in aquatic environment, they face adverse environmental conditions for their survival and natural death occurs. If water is kept in a container thus blocking the supply of food and nutrient, a higher death rate should occur. EXPERIMENTAL APPROACH Sampling Program

In this study, water samples were collected from the available water sources in the flood-affected regions in and around Dhaka city. The sources were WASA supply line (taps in household and street hydrants), hand-pump tubewells, and open water bodies in flood-affected areas. Development of Technologies

For treating water in flood-affected areas, 5 low-cost technologies were experimentally investigated to evaluate their performance. The technologies were (i) alum coagulation, (ii) chlorination, (iii) chlorination-alum coagulation, (iv) boiling, and (v) natural decay. The experimental procedures followed are described below. Alum Coagulation

Dry alum [Al2(SO4)3. 14H2O] was dissolved in distilled water to obtain a stock solution of alum of 5 mg/mL strength. To determine the optimum alum dose, 3 samples were collected from different locations of Dhaka city.


The usual procedure of establishing the optimum alum dose is to conduct Jar tests. But in reality, the results of Jar tests are not fully applicable in case of water treatment by individuals in flood-affected areas. Hence a simplified method was established and followed in determining the optimum dose for water treatment. The procedure involves the following steps: (i) Taking 500 ml water sample in each of 5 beakers, (ii) Adding alum solution to the beakers to obtain alum dosing of 100, 200, 300, 400 and 500 mg/L, (iii) Stirring vigorously with a bamboo/wooden stick for about 1 minute for uniform mixing, (iv) Stirring slowly for about 1 minute to facilitate floc formation, (v) Providing 30 minutes settling time for the flocs, (vi) Taking supernatant from each beaker at the end of the settling time and measuring turbidity, pH and faecal coliform of the collected supernatant.

After determining the optimum alum dose from these experimental results; another set of experiment was carried out to show the effect of contact time on removal of bacteria applying the optimum dose. Two sets of experiments were also conducted to evaluate the effect of contact time on bacterial removal at 500-mg/L alum dose using 2 samples. The contact time was varied from 15 minutes to a maximum of 120 minutes. Chlorination

Collected samples from tap and hand pump tubewells from the flood-affected areas showed the presence of coliform but the water looked clean. Hence only chlorination was applied to these water samples to kill the pathogens in order to make them suitable for drinking. One such sample was treated with different chlorine doses. Bleaching powder (25% strength) was used to prepare a stock solution of chlorine. The dosing varied from 0.5 to 2.0 mg/L. After adding the doses to the beakers, each containing 200 ml water sample, the content was vigorously mixed with a glass rod. Allowing a contact time of 15 minutes, the treated water samples were tested for the presence of coliform and residual chlorine. The optimum dose of chlorine was determined from the experimental results. Alum Coagulation-Chlorination

For clean water (i.e., water without suspended particulate, color, etc.) in flood-affected areas, chlorination was sufficient to produce drinking water. But depending on the magnitude of flood, clean water was not available in many areas and the floodwater was to be treated for drinking. As the floodwater contained turbidity, color and pathogens, chlorination and alum coagulation were simultaneously applied to the water. A total of 5 water samples were considered for chlorination-alum coagulation. The turbidity, color and faecal coliform concentrations of the water samples were measured. Then for each water sample,




500 ml was taken in a beaker and 1.0 mg/L of chlorine and 200-mg/L of alum (the optimum alum dose) were added and the same procedure as employed in the case of alum coagulation was followed. The turbidity, color, faecal coliform and residual chlorine concentrations of the treated water were determined. Boiling

Chemicals may not be available to treat water in the flood-affected regions. So boiling may be the only option to kill the pathogens of the drinking water within a short time. Since fuel is scarce in the flood-affected areas, the minimum boiling required was investigated to save fuel. To fulfill the objective, a total of six water samples were boiled to reveal the effect of temperature and time on death rate of coliform. Floodwater, stagnant water and WASA water were heated upto boiling. For each case, the temperature of the water was measured and water sample was collected at different times. The water samples were analysed for coliform count in order to determine the effectiveness of temperature and boiling time on disinfection. Water was heated both in glass beaker and aluminum cooking pot (dekchi). For heating in beakers, multiple tube fermentation technique was used to determine the presence/absence of coliform. Membrane filtration method was used to enumerate the faecal coliform in the water samples. Storage for Natural Decay of Coliform

Two flood water samples were tested for natural decay having initial faecal coliform count of 10,000 and 37,500 per 100 ml. The samples were kept open in the laboratory under natural environment. Faecal coliform count was determined on different days to determine their decay rate. Strength Measurement of Locally Available Disinfectants

In this study, a survey was conducted on the availability of disinfectants in the local markets. It was found that disinfectants were available in two forms - chlorine tablet and bleaching powder. Since people of flood-affected area normally use these disinfectants, their strength (chlorine content) was determined. The commonly available disinfectants are described below. Chlorine Tablets

In this study, different types of disinfectants, available in the market in “tablet” forms, were purchased and the strength of each type in treating water was determined separately. The brand names of the chlorine tablets procured from the market were (i) Halotab, (ii) Halazone, (iii) Hydroclonazone, and (iv) Wasserent


keinun. Each type of tablet was dissolved in distilled water and the chlorine content was determined by standard procedure. Bleaching Powders

Various types of bleaching powder available in the local markets were procured and the chlorine content of each type of the bleaching powder was determined by standard method after dissolving it in distilled water. The tested bleaching powders include Navy bleaching powder, Azad bleaching powder and SP Company’s bleaching powder. RESULTS AND DISCUSSION Alum Coagulation

To determine the optimum alum dose, 3 water samples of flood-affected zones were collected and tested for alum coagulation. The pH, turbidity and coliform concentrations of the collected samples are shown in Table 1.

Figure 1 presents the effect of alum dose on residual turbidity of treated water samples. From the figure, the optimum alum dose was found to be about 200 mg/L for all the three different samples tested. Table 1: Characteristics of Water Samples used for Determining Optimum Alum Dose

Sample No.

Source of water pH Turbidity (NTU)

Faecal Coliform (#/100 mL)

1 Flood water 7.2 147 6,000 2 Buriganga River,

Nawabganj 6.4 19 60,000

3 Buriganga River Sawarighat

5.8 23 70,000

The pH change of the treated water samples with the variation of alum dose

is shown in Fig. 2. The pH values decreased due to the formation of hydrogen ion as a result of chemical reaction with water. The higher was the alum dose, the more hydrogen ions were produced, resulting in lower pH value. In case of floodwater, the pH change was smaller than that of the river waters. It indicates that the floodwater had higher alkalinity value in comparison with that of the river waters.




0 100 200 300 400 500

0

5

10

15

20

25

Sample -1 Sample -2 Sample -3

Turb

idity

(NTU

)

Alum Dosage (mg/L)

Figure 1: Effect of alum dose on residual turbidity of water

0 100 200 300 400 5003

4

5

6

7

8 Sample -1 Sample -2 Sample -3

pH V

alue

Alum Dosage (mg/L)

Figure 2: Change of pH value with the variation of alum dose


The relationship between the removal of faecal coliform and the alum dose for 30-minute contact time is shown in Fig. 3. The removal was higher with the increase of alum dose. It appears that the removal was greatly affected by the composition of the raw water. In case of significant reduction in pH value due to the addition of alum, a high degree of removal was attained. It appears that the acidic condition resulted in higher removal efficiency. However, complete removal of faecal coliform was not attained in any case.

100 200 300 400 5000

500

1000

1500

2000

2500

3000

Sample -1 Sample -2 Sample -3

Faec

al C

olifo

rm (N

o./1

00 m

l)

Alum Dosage (mg/L)

Figure 3: Effect of alum dose on removal of faecal coliform

The effect of settling time on bacteria removal at two different alum dosing is

presented in Fig. 4. It shows that the removal rate was very high initially but then slowed down. The larger and heavier flocs settled quickly and removed most of the microorganisms. Settling of micro-flocs occurred slowly and longer contact time at acidic condition was mainly responsible for subsequent removal. It was also noticed that there was a difference in the removal rate between the two water samples at the same alum dose. It appears that the bigger change in pH value produces higher removal rate. However, long settling time (2 hours) and high alum dose (500 mg/L) could not remove the faecal coliform completely.




0 20 40 60 80 100 1200

100200300400500600700800

Sample -2 (200 mg/L Alum) Sample -2 (500 mg/L Alum) Sample -3 (500 mg/L Alum)

Faec

al C

olifo

rm (N

o./1

00 m

l)

Settling Time (minutes)

Figure 4: Effect of settling time on the removal of faecal coliform Chlorination

The effect of chlorination on tap water is presented in Fig. 5. The initial faecal coliform count of the raw water was 210 no./100-mL and chlorine doses of 0.5 to 2.0 mg/L were applied. A contact time of 20 minutes was maintained. The figure shows that the residual chlorine increased with higher chlorine dose without showing any break point. No faecal coliform was detected at these applied doses. It should be noted that drinking water should not contain more than 0.2 mg/L of chlorine in order to avoid possible adverse health effects. Application of 0.5 mg/L of chlorine dose produced 0.08 mg/L of residual chlorine. Hence, from the figure, the allowable dosing of chlorine can be taken as 0.5 to 1.0 mg/L. The low dose of 0.5 mg/L should be applied to apparently clear water and the high dose of 1.0 mg/L should be applied to highly turbid floodwater.

A

lum Coagulation-Chlorination

To treat floodwater having high turbidity, color and pathogens, both chlorination and alum coagulation were applied. The turbidity, color and faecal coliform concentrations of the raw water samples are presented in Table 2.


0.5 1.0 1.5 2.0 2.50.00.10.20.30.40.50.60.7

Res

idua

l C

hlor

ine

(mg/

L)

Chlorine Dosage (mg/L)

Figure 5: Effect of chlorine dose on residual chlorine

Table-2: Characteristics of Floodwater Treated by Chlorination-Alum Coagulation

Sample Designation

Turbidity (NTU)

Color (TCU)

Fecal Coliform (No. / 100 ml)

A 90.0 123 37500 B 12.2 49 12500 C 20.0 56 34000 D 30.0 115 20000 E 57.0 125 8000

The results of turbidity removal are shown in Fig. 6. It is evident that the

turbidity in the range of 12 to 90 NTU was effectively removed (less than 5 NTU, WHO guide line value), resulting in highly transparent water. It should be mentioned that the drinking water standard of turbidity is 10 NTU according to Environment Conservation Rules (GOB, 1997).

The color removal through the process of chlorination-alum coagulation is shown in Fig. 7. The color removal was very effective for 2 samples (sample A and B), while in the cases of other 3 samples, the process failed to bring the color of the treated water below 15 TCU, the Bangladesh standard for color (GOB, 1997).




0

20

40

60

80

100

A B C D E

Turb

idity

(NTU

)

Sample Designation

Raw Water Treated Water

Figure 6: Removal of turbidity through alum coagulation-chlorination

0

20

40

60

80

100

120

140

A B C D E

Col

our (

Pt.C

o.U

nit)

Sample Designation


Figure 7: Effect of chlorination-alum coagulation on color removal


The faecal coliform removal and residual chlorine present in the treated water after the coagulation-chlorination process are presented in Table 3. Table 3: Removal of Faecal Coliform by Chlorination- Alum Coagulation

Faecal Coliform (No. / 100 mL)

Sample Designation


Residual Chlorine (mg/L)

A 37,500 0 0.58 B 12,500 0 0.51 C 34,000 0 0.73 D 20,000 0 0.65 E 8000 0 0.47

Note: Cl2 = 1 mg/L, Alum = 200 mg/L, Contact Time/Settling Time = 30 min.

From Table 3, it is evident that irrespective of the concentration of faecal coliform and other constituents initially present in the raw water samples, the process removed faecal coliform from the water samples completely. However, in all the treated samples the residual chlorine concentrations were more than 0.2 mg/L (drinking water standard). It can be concluded that the chlorine dose should be 0.5 mg/L when applied with alum coagulation. Since alum coagulation alone reduced the faecal coliform significantly, low dose such as 0.5 mg/L of chlorine may be adequate for floodwater. Boiling

The effect of boiling time on the temperature for two water samples (flood water and tap water) is shwon in Fig. 8. The floodwater started to boil at a temperature of 85° C after heating the water for 35 minutes. The tap water began to boil at 87°C after 35 minutes of heating. For both the samples, test results showed the presence of coliform at 40°C temperature, however at 80°C temperature, coliform was not detected. The corresponding heating times were 17 minutes and 23 minutes, respectively.

Fig. 9 shows the relationship between the boiling time and the temperature of water for boiling in aluminum cooking pot (dekchi) using gas burner for two water samples (flood water and tap water). It was observed that the samples started to boil at around 98° - 100°C after 20 minutes of heating. A comparison between Figs. 8 and 9 reveal that rate of temperature rise in an aluminum pot was much higher than that in a glass beaker. This was simply because aluminum is a good conductor of heat while glass is a bad conductor and the aluminum pot was under direct heating while the glass beaker was placed in a water bath.




0 10 20 30 40 502030405060708090

100

Flood Water WASA Tap Water

Tem

pera

ture

(

0 C

)

Boiling Time (minutes)

Figure 8: Effect of boiling time on water temperature

0 10 20 3020

40

60

80

100

Flood Water WASA Tap Water

Tem

pera

ture

( o C

)

Boiling Time (Days)

Figure 9: Effect of boiling time on water temperature


The effect of temperature of water on the removal of faecal coliform for the two samples is presented in Table 4. It is seen that faecal coliform was not detected in any sample when the temperature was higher than 60°C. The corresponding boiling time was 6 to 7 minutes. In practical application, it is difficult to monitor either the temperature of water or the duration of heating. So it is safe to boil a water sample upto the boiling point to ensure total removal of coliform. Table 4: Variation of Faecal Coliform Concentration with Temperature (T) of Water.

Faecal Coliform Count (No./100mL) Type of water T = 30.5oC T = 45oC T = 60oC T = 80oC T = 100oC

Tap water 10 144 0 0 0 Flood water 80,000 80,000 0 0 0

Natural Decay of Coliform

The results of natural decay for two samples are presented in Fig.10. It was observed that the decay rate was very high initially and then slowed down with time. It is seen that storing of water under natural environment even for 7 days was not sufficient for 100% removal of the faecal coliform. It proves that natural decay is a very slow process and is not a practical method to obtain pathogen-free water. Strength of Available Disinfectants

A total of four different types of chlorine tablets and three kinds of bleaching powder were tested for the presence of total chlorine. The results are shown in Table 5. Table 5: Chlorine Content of Commercial Disinfectants

Type of disinfectant

Trade Name Total Chlorine (mg per tablet)

Total Chlorine (%)

Halotab 3.0 - Halazone 0.92 - Hydroclonazone 1.11 -

Chlorine Tablet

Wasseren keinun 0.02 - SP company - 0.4% Navy - 1.4%

Bleaching

powder Azad - 0.85%




0 1 2 3 4 5 6 7 80

5000

10000

15000

20000

25000

30000

35000

40000

Sample-F Sample-G

Faec

al C

olifo

rm (N

o./1

00m

l)

Detention Time (Days)

0 1 2 3 4 5 6 7 8

Figure 10: Effect of detention time on natural decay of faecal coliform

It is seen from Table 5 that the chlorine content of the tablets varied from

0.02 to 3.0 mg per tablet. Manufacturing and expiry dates were not mentioned for any of the brands tested. Since chlorine is volatile, the shelf life of the tablets was the prime reason for such a wide variation in strength. Other reasons might be the variations in constituents, production method and packaging of the tablets. Halotab, a local product, was the best in terms of strength, possibly because it was recently produced.

The bleaching powders showed a large variation in chlorine strength, which varied from 0.4 to 1.4 %. Manufacturing and expiry dates were not written on the packets. Constituents as well as shelf life are the possible reasons behind the variation. It appears that the available disinfectants are not reliable for proper dosing. Smell of chlorine-based disinfectant may be a crude indicator of the strength - the more intense the odor, the higher the chlorine content.

CONCLUSIONS A variety of simple techniques were studied to develop low-cost technologies for water treatment in flood-affected areas. Boiling was found to be the most effective method of killing almost all types of pathogens. Heating upto the boiling point was found to be sufficient for this purpose. Chlorination was found to be quite effective in killing pathogens at a chlorine dose of 0.5 -to1.0 mg/L


and 30-minute contact time. Alum dosing at 200 mg/L could effectively reduce the turbidity of floodwater but it could not make the water bacteria-free even at much higher doses. Colored and turbid floodwater could be treated with application of chlorination and alum coagulation simultaneously. For this purpose, the required alum dose was found to be 200 mg/L and the chlorine dose 0.5 mg/L. The settling time/contact time to be provided was found to be 30 minutes. The quality of the available chlorine disinfectants varied widely mainly due to variation in shelf life. The strength may roughly be assessed by smelling. The quality of local chlorine tablets was found to be better than that of foreign tablets.

The low-cost technologies developed through this study can be conveniently used to treat contaminated water during floods/post flood period. The application procedures can be summarized as follows: (a) Heat water upto the boiling point. There is no need to heat the water beyond the starting of boiling. It is the most convenient method of killing pathogens at household level; (b) Add chlorine tablet or bleaching powder solution to tubewell water or water from distribution line or clean floodwater. Mix thoroughly and wait for 30 minutes. The water is then suitable for drinking. The chlorine dosing requirement is 0.5 - 1.0 mg/L, depending on the turbidity of the water. For apparently clear water, use 1 fresh Halotab (a local chlorine tablet containing 3.0 mg chlorine) for 6 L water or 1-teaspoon (5-mL) bleaching powder solution for 10 L water. The solution is to be prepared by dissolving 1-teaspoon (5 gm) medium strength bleaching powder (chlorine content = 10%) in 2 glasses (0.5 L) of water. Preserve this solution in a green/brown bottle with stopper; (c) Treat the water of underground reservoir by adding 5 L bleaching powder solution (10 teaspoon of 10% strength) per 10,000 L water and mix uniformly. Wait for about 1 hour and then use the water; (d) Complete removal of pathogens from water can not be achieved by alum treatment. Treat turbid/colored floodwater by adding both alum and chlorine. Dissolve 1-teaspoon (5 gm) alum powder in 25 L floodwater in a suitable container (e.g., a bucket fitted with a tap). Then dissolve 4 fresh Halotab in the water or add 2.5 teaspoon bleaching solution prepared as described earlier and preserved in a bottle. Stir the floodwater vigorously for 1 minute and then slowly for 1 minute. Wait for 30 minutes and collect/decant the clear water leaving sludge at the bottom of the container. This water is suitable for drinking; (e) The hand-pump tubewells in flood-affected areas need a post flood treatment. Prepare 10 L bleaching powder solution (5 teaspoon of 10% strength) for one tubewell. Open the seat (foot) valve of the tube well and add the bleaching solution. Wait for 12 hours and then start pumping. Discard all the water as long as the odor of chlorine is detected in the water. Then the tubewell water is suitable for drinking.




REFERENCES

GOB (1997), Environmental Conservation Regulation, Government of the Peoples Republic of Bangladesh.

WHO (1996), Guide Lines for Water Quality Standards 1996, World Health Organization, Geneva.


Assessment of Water Quality in Flood Affected Areas of Dhaka City

Md. Mujibur Rahman, Md. Mafizur Rahman and Md. Delwar Hossain

Department of Civil Engineering Bangladesh University of Engineering and Technology, Dhaka-1000, Bangladesh Abstract The impacts of 1998 flood on water quality in and around the eastern part of Dhaka city have been assessed. The floodwater has been found to be heavily polluted by organic wastes. High concentration of BOD5 and COD have been found in the stagnant floodwater indicating accumulation of organic pollutants from a variety of sources including submerged sewerage system and direct discharge of human excreta and other household solid wastes. As the communication system of the affected areas was totally disrupted, people were directly exposed to such polluted floodwater that caused them to suffer from various skin diseases. The drinking water supply was also found to be heavily polluted by fecal coliform, which resulted in the outbreak of diarrhoea and other waterborne diseases.

BACKGROUND

Bangladesh, a fertile deltaic region, criss-crossed by numerous rivers, is subject to periodic and occasionally catastrophic flooding. The three large international rivers – the Ganges, the Brahmaputra and the Meghna and their tributaries and distributaries, dominate the hydrology of Bangladesh. These river systems constitute a large catchment area of about 1.72 million square kilometers lying mostly in India, China, Nepal and Bhutan and only 8% of the catchment area lies within Bangladesh (Nishat, 1993). As a result, a huge uncontrolled inflow of water enters into Bangladesh,




which in combination with other factors, e.g., heavy monsoon rainfall over Bangladesh, low floodplain gradient, congested drainage channels, influence of tides and storm surges greatly contribute to the recurring flooding in Bangladesh.

The flood in 1998 was significant due to its prolonged duration (Islam, 1999) above danger level (see Fig. 1). The 1988 flood in Dhaka rose to the highest level on record with peak water level of 7.58 m above sea level and the water above the danger level mark of 6.0 m for 23 days (SWMC, 1998). This is comparable to the 1954 flood with 7.06 m peak water level that lasted for 46 days, and the 1955 flood with 7.09 m peak water level that remained above the danger level mark for 31 days. The 1998 flood on the other hand rose to a peak level of 6.70 m and remained above the danger level for more than 80 days. This prolonged flood which affected some 100,000 sq. km. areas out of the total area of Bangladesh of 148,000 sq. km., destroyed basic infrastructures like roads, bridges, houses, standing crops and killed birds, animals and cattle heads.

The long duration of the 1998 flood caused immense sufferings of the people in the affected areas. Disruptions of drinking water supply, sanitation, waste disposal, and disease transmission were among the major adverse impacts of the 1998 flood. This flood heavily affected the eastern part of the capital city, Dhaka. The residents of this part of the city suffered the consequences of this devastating flood. Embankments constructed after the devastation of 1988 flood however, protected the western part of the city. Most of the houses and roads in the eastern Dhaka were inundated by the floodwater (see Fig.2). The water supply system, the sewerage system, gas pipe network, underground electric and telephone cable system all remained submerged for the entire duration of the flood. About three million people in the affected areas of Dhaka got marooned. People had no other option but to use the manually driven small boats that replaced all the motorized and non-motorized vehicles on road.

The most serious problem encountered by the affected people was the quality of water, which deteriorated as a result of many factors. The floodwater, already fouled by the wash-aways from the upstream areas was further deteriorated by the complete submergence of the sewerage system, septic tanks and other sanitary facilities and by the direct disposal of human waste, kitchen waste and household refuse in the absence of sanitation and municipal facilities. The affected people were directly exposed to the polluted floodwater, which resulted in the outbreak of various skin diseases, in addition to serious diarrhoea and other waterborne diseases as a result of drinking contaminated water. This study was undertaken in order to assess the extent of deterioration of both the stagnant floodwater and the supplied drinking water quality.

Assessment of Water Quality


Figure 1: Comparison of the 1988 and 1998 Floods



(a)

(b)

Figure 2: (a) The eastern part of Dhaka city during the 1998 flood, (b) Small

boats plying on the roads in a flood-affected area


(c)

(d)

Figure 2: (c) Status of sanitation and solid waste disposal around residences in a flood-affected area, (d) Direct defecation in a flood-hit area




(e)

(f)

Figure 2: (e) Big bubbles coming from a manhole indicating mixing of

sewage and floodwater, (f) Solid waste floating on floodwater


STUDY PROGRAM Study Area

The affected area (see Fig. 3) of Dhaka city was confined within the eastern side of Airport road-Progati Sarani-Biswa road to the west and the Balu River on the east. The affected eastern part of Dhaka City was divided into four regions (see Table 1) for the study. These four regions covered almost the entire affected area. In order to assess the quality of water in the affected areas, two locations were then identified from each of these four regions for data and sample collection. Table 1: Sub-division of the study area

Location-1 Shahjadpur Region A Location-2 Nayabazar Location-1 Khilgaon Crossing Region B Location-2 Rampura Sluice Gate Location-1 Moddhya Bashabo Region C Location-2 Tilpa para Location-1 Waste Disposal Site Region D Location-2 Bibir Bagicha Gate

Methodologies

The study methodologies included field reconnaissance, questionnaire survey and water sample collection for subsequent laboratory analysis. Field situation was observed during and after the flood throughout the month of September. Field investigation included visual observations of the flood situation and its effects on the city life captured by still photography and questionnaire survey, that was designed to collect information on various impacts faced by the city dwellers, e.g., availability of food and drinking water, sanitation systems, solid waste disposal systems, communication, health and medical services. Samples of both floodwater and drinking water were collected from specified locations for laboratory analysis.

Collected water samples were analyzed in the Environmental Engineering Laboratory of Bangladesh University of Engineering and Technology (BUET). Floodwater samples were tested for Chemical Oxygen Demand (COD), Biochemical Oxygen Demand (BOD5), Turbidity and Faecal Coliform (FC) to assess the extent of pollution. Drinking water samples were tested for FC and Turbidity in order to determine their suitability for drinking.




Figure 3: Dhaka city map showing flood-affected areas

RESULTS AND DISCUSSION Quality of Floodwater

Floodwater samples were collected from eight locations and were analyzed for different water quality parameters. Figures 4 and 5 show BOD5 and COD of floodwater as a function of time.


0

10

20

30

40

50

60

70

80

5/9/98 8/9/98 14/9/98 17/9/98 19/9/98

Date

BOD

(mg/

L)

Shahjadpur Nayabazar Khilgaon CrossingRampura Sluice Gate Maddhya Bashabo Tilpa ParaWaste Disposal site Bibir Bagicha Gate

Figure 4: BOD5 at different locations as a function of time

0

50

100

150

200

250

300

5/9/98 8/9/98 11/9/98 14/9/98 17/9/98 19/9/98

Date

COD

(mg/

L)

Shahjadpur Nayabazar Khilgaon CrossingRampura Sluice Gate Maddhya Bashabo Tilpa ParaWaste Disposal site Bibir Bagicha Gate

Figure 5: COD at different locations as a function of time

The figures show that BOD5 and COD gradually increased gradually in the floodwater in all the areas and then declined after sometime. This trend is common for all the sampling locations. Generally very high concentrations of




BOD5 and COD were found between 5th and 14th September. Looking back to Fig. 1, it can be seen that during this period the flood water level gradually increased until it reached the peak level sometime between the 9th and 14th September. The onrush of floodwater from the upstream catchment occurred at a much higher rate than the passage of floodwater downstream leading to a gradual increase in the flood water level, which remained stagnant for a long period of time. It appears that during this prolonged stagnation period, pollutants from a variety of sources accumulated in the floodwater resulting in high concentrations of BOD5 and COD in floodwater. It may not be appropriate however, to suggest a strong correlation between high flood water level and high concentrations of BOD5 and COD, given that the number of measurements were not sufficient for establishing such a correlation. In addition, there are inherent uncertainties, at least to some degree, in the measurement procedures.

In the inundated areas, the submerged sewerage system failed and sewage became mixed with the floodwater. The flow of floodwater was very slow and was almost stagnant during the entire flood-period. As a result, people in these areas became waterlogged and had no option but to discharge different types of wastes, including human excreta, sullage, and the household waste, in the floodwater. This appears to be the primary reason for the high BOD5 and COD in floodwater.

Figures 6 and 7 clearly demonstrate a major disruption of the sanitation system in the affected areas. As evident from Fig. 7, about 60% of the people could not avail their normal sanitation services and had to resort to mostly unhygienic practices, e.g., open defecation (11%), direct defecation into flood water (19%) and defecation in polythene bags (10%) which were subsequently thrown into flood water. These practices contributed to the deterioration of the quality of the floodwater.

Figure 8 shows faecal coliform concentration in the floodwater samples. As shown in Fig. 8, faecal contamination of the floodwater was very high. As mentioned earlier, sewerage system in the inundated areas failed and became totally submerged under floodwater. As a result, there was mixing of sewage and floodwater. This caused the faecal coliform to rise in floodwater. Since the floodwater received all types of wash-aways including sewage and solid wastes, it was also found to be highly turbid.

Since the floodwater also received wastes from small and medium sized industries located within the city area, it is possible that the floodwater might have been contaminated by heavy metals like lead, cadmium, chromium and mercury. It was however beyond the scope of this study to assess the heavy metal contamination in floodwater.


Sanitation Practices: Normal Condition

39%

15% 4% 18%

24%

Sewerage Septic Tank Sanitary Pit LatrineHanging Latrine Community Facility

Figure 6: Percentage of population using different sanitation options under

no-flood condition

Abnormal Sanitation Practices: During Flood(Normal sanitation facilities not usable)

11%5%5%

10%

19%10%

Communal Facility Open Defecation Ploythene bagsDirect Defecation Distant Sanitation Others

Figure 7: Sanitation practices by the people in flood-affected areas during

flood




0

40000

80000

120000

160000

200000

240000

8/9/98 11/9/98 14/9/98

Sampling Date

FC n

o/10

0ml

A1 A2 B1 B2

C1 C2 D1 D2

Figure 8: Faecal Coliform (FC) concentration in floodwater Quality of Drinking Water

Drinking water samples collected from all the eight locations were tested for FC and the results are presented in Fig. 9. Presence of bacteriological contamination was observed in samples from all locations. Any water with presence of FC is recommended to be unsuitable for drinking. Presence of coliform in the drinking water supply appears to be the primary reason behind the outbreak of diarrhoea and other waterborne diseases during and after the flood.

0

1000

2000

3000

4000

5000

5/9/98 8/9/98 11/9/98 14/9/98 17/9/98 19/9/98

Sampling Date

FC n

o/10

0mi

A1 A2 B1 B2C1 C2 D1 D2

Figure 9: Faecal coliform concentration in drinking water


CONCLUSIONS Floodwater as well as drinking water was found to be grossly polluted in the all flood affected areas of Dhaka. The quality of floodwater deteriorated due to prolonged stagnation of floodwater and accumulation of pollutants from different sources. Complete submergence of the sewerage system and direct discharges of human wastes and other household solid wastes further aggravated the quality of floodwater. The drinking water supply was also found to be contaminated, primarily because of the submergence of the water supply pipe network. No correlation however, could be established between the extent of bacteriological contamination of the floodwater and that of the drinking water. The polluted water added to the sufferings of the people in the affected areas. People in the flood-affected areas suffered from a variety of diseases, e.g., skin diseases due to exposure to polluted floodwater, and diarrhoea and other waterborne diseases due to drinking of contaminated water.

The long-term solution to the flood problems in Dhaka city would require construction of embankments along the eastern boundary of Dhaka. The economic loss suffered by the eastern part of Dhaka city would well justify the substantial investments for embankment construction. However, until the funds for embankment construction are made available, several short-term measures should be considered in order to minimize the adverse effects of flood with respect to water quality problems. For example, the submerged segments of sewerage system and piped water supply system within the affected area should be sealed and made inoperative during the flood period. Alternative supply of drinking water must be ensured to the flood affected area. Alternate options must also be considered for disposal of human excreta and household waste in the inundated areas. Industrial activities generating and discharging wastes into the floodwater must be stopped during flood. REFERENCES

Islam, M. R. (1999) "Drinking water quality and sanitation condition of flood shelters in Dhaka city", M. Engineering Thesis, Department of Civil Engineering, Bangladesh University of Civil Engineering, Dhaka.

SWMC (1998) Personal Communication, Surface Water Modeling Center, Dhaka.

Nishat, A. (1993) “Sector Review: Water Resources Development”, Environmental Management Training Project, Instructors’ Manual, DanEduc Consulting, GOB and Asian Development Bank, Dhaka



Purification of Floodwater by Electrolysis

A. B. M. Badruzzaman Department of Civil Engineering

Bangladesh University of Engineering and Technology, Dhaka-1000, Bangladesh and

Md. Rezwan Khan Department of Electrical and Electronic Engineering

Bangladesh University of Engineering and Technology, Dhaka-1000, Bangladesh Abstract An adequate supply of pure water is absolutely essential to human existence. During floods, water is available in adequate quantity, but the quality is often not suitable for drinking or cooking purposes. The goal of the study was to develop an easier, cheaper and faster technique of water purification. Oxygen ions are well known for their high reactivity. An attempt has been made to use oxygen ion to disinfect water. If a container of water is subjected to electrolysis, hydroxyl and hydronium ions are produced which move through the liquid from one electrode to another. It is expected that these ions, particularly the hydroxyl ion, will come across bacteria and react with it chemically to eliminate it. The process is similar to the mechanism of disinfecting water using halogen tablets. Considerable reduction in indicator organisms was obtained through treatment of floodwater with alum followed by electrolysis. The results obtained are not conclusive but are encouraging. INTRODUCTION Halogen tablets are widely used to purify drinking water. When a halogen tablet is dissolved in water, chlorine ions are generated, which are strong oxidizing agent. Chlorine ions are of high toxicity and kill bacteria on contact. Based on this basic idea of oxidation of the bacteria, an attempt has been made to evaluate


A. B. M. Badruzzaman and Md. Rezwan Khan


the effect of hydroxyl ion, generated through electrolysis, on bacteria. The aim was to develop a simple water purification technique, based on electrolysis, which could be used in flood-affected areas in Bangladesh. Bangladesh experiences flood almost every year and its magnitude often reaches a catastrophic level. In most of the situations, halogen tablets are not available in the village shops. People suffer from shortage of drinking water and diarrheal diseases take an epidemic form. During floods, although water level rises significantly, electricity supply remains uninterrupted in many places and electrolysis may be adopted for purification of water. THEORY OF THE PROPOSED TECHNIQUE Pathogens

Pathogens are organisms capable of infecting, or of transmitting diseases to humans. These organisms are not native to aquatic systems and usually require an animal host for growth and reproduction. They can, however, be transported by natural water systems, thus becoming a temporary member of the aquatic community. Waterborne pathogens include species of bacteria, viruses, protozoa and helminths (parasitic worms). Pathogen Indicators

Analysis of water for all the known pathogens would be a very time consuming and expensive proposition. Purity of water, from bacteriological perspective, is usually checked through the indicator organisms. Most of the waterborne pathogens are introduced through fecal contamination of water. There are other coliform groups, which flourish outside the intestinal tract of animals. These organisms are native to the soil and decaying vegetation and are often found in water that was in recent contact with these materials. It is the usual practice to use:

• The total coliform (TC) group as indicators of the sanitary quality of the drinking water.

• The fecal coliform (FC) group as indicators of wastewater effluents. The test for total coliform organisms employs slightly different culture media and lower incubation temperature than those used to identify fecal coliform organisms.


The membrane-filter technique stated in the Standard Methods (Greenberg et al., 1999) was used to determine the degree of contamination of water. This technique is popular with environmental engineers and gives direct count of coliform bacteria. In this test, a portion of the sample is filtered through a membrane, the pores of which do not exceed 0.45 µm. Microorganisms are retained on the filter that is then placed on selective media to promote growth of coliform bacteria while inhibiting growth of other species. The membrane and media are incubated at the appropriate temperature for 24 hours, allowing coliform bacteria to grow into visible colonies that are then counted. The results are reported in numbers of organisms per 100 mL of water.

EXPERIMENTAL SETUP Preparation of the Electrodes

Since, the purification process was based on electrolysis, the success of the study greatly depended on having good and reliable electrodes. In this study, nickel-coated galvanized iron sheets were used as electrodes. The nickel coating was used to reduce corrosion of soft iron when electrolysis is performed. Approximately 3 µm thick nickel coating was provided on each iron sheet. The dimension of the electrodes and the beaker was interrelated. Measurements for the electrodes are given in the Fig. 1.

Twelve pieces of electrodes were prepared at a local galvanizing factory by applying appropriate nickel coating. Two important factors were considered during the fabrication of the electrodes. These were: (1) It is essential that the plates and the coatings should be as smooth as possible. It is likely that excessive corrosion would occur if there are bends or scratches on the plates (electrodes); and (2) A thick and uniform nickel coating is essential in reducing rapid degradation of electrodes. Preparation of the Reaction Vessel

A specially designed reaction vessel was constructed to perform the electrolysis of floodwater. A 600mL perspex box, fitted with guide slots on the shorter sides for electrodes, was constructed for the above purpose. The dimension of the beaker is shown in the Fig. 2. Experimental Setup and Procedure

The experimental setup included the following equipment:

• Auto-transformer for variable ac voltage




• Multi-meter to measure current • Voltmeter • Test tubes and beakers to perform chemical analysis • Incubators

Figure 1: Dimension of the Electrode TREATMENT OF FLOODWATER During the course of the experimental setup, the floodwater receded from around the Dhaka city. Thus, actual floodwater samples could not be collected, instead samples were collected from the pond at the Suhrawardhy Uddayan and the surface drains at the Dhaka Medical College Hospital (DMCH). Synthetic samples were prepared from these samples to substitute for the floodwater. It was suspected that the collected samples were extremely polluted with pathogens and indicator organisms. Thus, very small amount of this sample was added to a 17-liter bucket full of tap water, which was considered as the stock floodwater sample for the purpose of this experiment. From this stock each time 600mL was used at different stages of the experiment. Prior to each set of experiment, initial concentrations of the indicator organisms, TC and FC, were determined by the Membrane Filtration technique. However, since the initial concentrations of the


indicator organisms were suspected to be very high the synthetic sample was diluted prior to testing. Otherwise, the total number of the colony developed on the membrane filter would be so high that it would be impossible to count them.

Perspex Box

Electrode

Figure 2: Schematic Diagram of the Electrolytic Cell Used for Purification of Water

The electrodes were placed in the beaker and were clamped by the side

bands. Two types of power supplies (AC and DC) were used. For AC supply the Variac was used, whereas, for DC supply the regulated power supply was used. Ammeter and voltmeter were used to monitor current and voltage, respectively. A constant current was passed through the series circuit arrangement for a pre-designated period. Time span was recorded using a stopwatch. After the synthetic sample was treated with electricity for the pre-designated period a specific amount of sample was extracted from the reaction vessel for analysis of TC and FC using Membrane filtration. Figure 3 shows the schematics of the entire process.

The second step of the experiment involved the application of locally available household chemicals such as common salt (NaCl) and Alum [Al2(SO4)3, 2H2O] to the synthetic floodwater followed by electrolysis. Initially, 50 gm of common salt was added to the synthetic floodwater and was thoroughly mixed until it was completely dissolved in water. Then the sample was poured into the reaction vessel and treated with AC power as before. A small amount of the supernatant liquid was extracted from the vessel and processed as before for analysis.




Figure 3: Schematic diagram of the steps involved in treatment of floodwater by electrolysis

Following the above step another stock of the synthetic sample was dosed

with Alum (125 mg/L) and vigorously mixed. Then to hasten the flocculation process the mixed solution was stirred slowly. This allowed the smaller flocs to come into contact with each other and form larger flocs. The flocs were then allowed to settle for 30 minutes. The supernatant liquid was then poured into the reaction vessel and treated with AC supply. A small amount of treated water was then extracted as before for analysis of TC and FC. The entire procedure was repeated with DC power supply.

Floodwater samples were also collected from the Buriganga River during the 1998 flood and laboratory experiments were conducted using alum and


electrolysis as before. Water near the banks of the river is likely to be highly polluted and aesthetically unacceptable. Hence, floodwater sample was collected from the middle of the river.

RESULTS AND DISCUSSION The results of different treatment options of synthetic water samples are presented in Table 1. This table shows that electrolysis and combination of electrolysis and chemical addition did not prove to be very effective with extremely contaminated synthetic samples. Table 1 indicates that even after dilution the synthetic samples remained highly polluted with indicator organisms. Table 1: Results of Different Treatment Options of the Synthetic Sample

Power Supply

Condition Time Current Voltage Temp TC FC

(min) (amp) (volts) (°C) #/100mL #/100mL Direct No 0 -- -- 29 28,000 1,300

Current Chemicals 5 0.50 38 44 5,400 700 Before 0 -- -- 29 56,000 40,000

Alternating adding NaCl 5 1.01 117 43 9,600 30,000 Current After 0 -- -- 29 56,000 40,000

adding NaCl 5 1.00 60 34 7,000 1,100 Before 0 -- -- 29 72,000 32,000

Direct adding Alum 5 1.00 150 45 10,400 3,200 Current After 0 -- -- 29 4,400 1,200

adding Alum 5 0.50 81 32 800 400 Before 0 -- -- 24 TNTC* --

Alternating adding Alum 5 1.02 187 42 8,000 -- Current After 0 -- -- 24 TNTC* --

adding Alum 5 1.01 187 46 500 -- *TNTC = Too Numerous To Count

When the synthetic sample was treated with only the direct current about 80-

85% removal of Total Coliform was achieved, whereas, the same for the Fecal Coliform was about 50-90%. It was also observed that the removal rate of both the TC and FC increase with the increase of current flow. On the other hand, the removal rates for TC and FC using alternating current were 83% and 25%, respectively.

The second step of the experiment involved addition of locally available chemicals such as common salt and alum followed by electrolysis. No significant improvement in reduction of TC was observed (compared to the electrolysis




alone) when NaCl was added to the synthetic sample followed by electrolysis with AC supply. On the other hand, alum coagulation of the synthetic sample alone removed about 95% of both TC and FC. Further reduction of about 80% and 75% of TC and FC, respectively of the synthetic samples pretreated with alum were achieved through electrolysis using DC supply. Although reduction of TC and FC through the application of AC supply following alum coagulation was achieve, the actual rate of removal could not be established due to excessively high initial concentrations of the indicator organisms.

The above experiments on highly polluted synthetic samples indicated that the aforementioned approach has enormous potential for treating floodwater for domestic use. Electrolysis of the floodwater sample was performed using AC supply following alum coagulation. Table 2 shows the results of the laboratory analysis of the actual flood sample. It was observed that the initial concentrations of the TC and FC of the floodwater sample were considerably lower than the synthetic sample (Table 2). About 30% reduction in TC and almost complete removal of FC was obtained following alum coagulation. A complete removal of total coliform was observed when the floodwater sample pretreated with alum was electrolyzed using AC supply. However, it should be noted that the temperature of the sample increased considerably when AC supply at high voltage was applied. This increase in temperature may have contributed to the reduction of some of the total coliform of the pretreated sample.

Table 2: Results of Electrolysis Treatment of Floodwater Sample

Power Supply

Condition Time Current Voltage Temp TC FC

(min) (amp) (volts) (°C) # 100mL #/100mL Initial -- -- -- 250 50 Alternating 30 min. after -- -- -- 32 170 0

Current adding Alum After 0 -- -- 32 170 0 applying

current 5 1.02 203 53 0 0

CONCLUSIONS In this study a limited number of experiments were performed due to resource and time constraints. Although no concrete recommendations can be made based on such a limited study, the results obtained are encouraging. Thus, it is suggested that a more comprehensive study be conducted involving actual


floodwater samples. It is also recommended that application of high frequency power supply for treatment of floodwater be studied. ACKNOWLEDGEMENT The investigators are indebted to Mr. Gazi Naser Ali, Mr. Md. Emamul Hassan Bhuiyan and Ms. Samina Sultana, Lecturers of the Department of Electrical and Electronics Engineering, BUET for their assistance in conducting the laboratory experiments. Mr. Abdur Rahman and Mr. Abbasuddin of the Environmental Engineering Laboratory of the Department of Civil Engineering, BUET also deserve thanks from the investigators for their continuous support at the laboratory. REFERENCES Greenberg, A. E., Clesceri, L. S., and A. D. Eaton, A. D. eds (1999), Standard Methods for the Examination of Water and Wastewater, 20th ed., American Public Health Association, New York, 1999.



The Completeness and Vulnerability of Road

Network in Bangladesh

A. S. M. Abdul Quium and S. A. M. Aminul Hoque Department of Urban and Regional Planning

Bangladesh University of Engineering and Technology, Dhaka-1000, Bangladesh Abstract Road transportation has emerged as the most popular mode of transportation in Bangladesh. Despite being the most popular mode, it suffers heavily from network failures due to natural as well as man-made disruptions. This paper presents the major findings of a study relating to the completeness of the road network with respect to the pattern of traffic flow and its vulnerability with reference to the disruptions caused by the flood of 1998. The study considered development of road network by regions on the basis of degree of network completeness by connectivity indices namely, γ and α indices that apply the Graph Theory to measure the geometric pattern of a network. The analysis revealed that the road network development in Bangladesh at the national and regional levels is still in its early stage. The existing network has a very few circuits for movement at all the levels and, therefore, susceptible to high risk of disruption. In terms of circuitry, the network has a bare marginal grid configuration at the divisional and national levels. The present overwhelming dependence on road transportation system will continue in the future. In this context, completeness of the whole road network should be considered in all future road planning exercises. Assessment of vulnerability of links around Dhaka deserves special consideration as any link failure of the national highways around Dhaka affects almost the whole system. To improve the reliability of the network, selected links of the existing feeder roads can be gradually upgraded to regional road standards so that alternative routes around the most critical links can be created. Additional alternative paths for the major movements should also be considered to improve network reliability and thereby upgrade the network to a modest grid.


A. S. M. Abdul Quium and S. A. M. Aminul Hoque


INTRODUCTION AND BRIEF HISTORY OF ROAD TRANSPORT DEVELOPMENT The history of modern road development in Bangladesh is not very old. During the British period, water transport and railways served as the two major modes. At that time, road development was considered as a subject of local interest and therefore, the responsibility was given to the provincial governments. They in turn transferred the responsibility to the local bodies - the District Boards (Khan, 1982). The colonial British Government prepared a master plan for road development in 1938 for India, which included the then Bengal. However, road development according to that plan did not advance much due to outbreak of the World War II in 1939. Consequently, at the end of the British Rule in 1947, there were only few kilometers of inter-urban paved roads in the parts of present Bangladesh. Realizing the importance of road transportation, the Transport Advisory Council set up by the then Government of Pakistan recommended the preparation and implementation of a Six Year Plan. Implementation of this Plan also did not advance much due to non-availability of sufficient funds. After launching of the First Five Year Plan (FFYP) of Pakistan in 1955, the earlier road development plan had to be reviewed and revised. The FFYP was followed by two more successive Plans. Under these plans, about 2,500 miles (4,023 km) of mostly single lane paved roads and another 1,500 miles (2,414 km) of roads under various stages of construction were built by 1971 (Khan, 1982). Road building received a new impetus after the emergence of Bangladesh in 1971. Very soon road transportation became the most popular mode of transportation. Its shares of both passenger and freight traffic became higher than combined shares of rail and water transport. The system is, however, still being developed and there are various issues that deserve attention for its effective functioning. The system suffers heavily from network failures due to frequent natural as well as man-made disruptions. This is especially prevalent during annual floods and other local disorders. In view of these network failures, it is critical that the vulnerability of the present network should be examined with respect to the present pattern of traffic flow. The road network development is not still complete. As such, the element of reliability of the existing network can be considered in future road planning so that in the event of any link disruption, major or strategic traffic flows can be re-routed to keep the system effectively functioning. SCOPE OF THE STUDY Examining the completeness and reliability of the road network system of Bangladesh was the major concern of this study. Vulnerability of the network was examined with special reference to the network disruptions during the flood of

Completeness and Vulnerability of Road Network


1998. In addition, it considered the degree of completeness of the rail network as a supplementary mode of transportation. An attempt was also made to relate road network development to the overall transport demand and its spatial distribution in the country as revealed by the Bangladesh Transport Sector Study (BTSS) of 1993 and Bangladesh Integrated Transport System Study (BITSS) of 1998. OBJECTIVES OF THE STUDY AND OUTCOME The primary objectives of the study were to identify the nature of road network deficiencies at the national and regional levels and to assess network completeness in the event of emergency situations that may cause disruptions to the normal traffic flow pattern. The specific objectives of the study were: (1) Identification of disrupted road links during the flood of 1998; (2) Assessment of the degree of completeness of road network at national and regional levels; (3) Relating road network development to the major traffic flows at the strategic level and (4) Suggesting recommendations for network improvement.

The nature of the major deficiencies of the road network was identified. The degree of completeness and circuitry of the network were measured and suggestions have been made to make the road transportation system more reliable. METHODOLOGY Tools of Analysis and Their Theoretical Considerations

The study considered network development by regions on the basis of degree of network completeness, circuitry and other characteristics by connectivity indices namely, γ and α indices that apply the Graph Theory to measure the geometrical pattern of a network (Black, 1981; Taaffe and Gauthier, 1973). A brief discussion on these indices is presented next. Structural configuration and complexity of a road network can be analysed by measurements that describe the degree of its network connectivity. Graph theory provides a number of such discriminating measures. Two of the most commonly employed measures of connectivity are the gamma and alpha indices. These are explained hereafter. The Gamma Index The gamma index is the ratio of the number of edges in a network to the maximum number possible in that network: γ = actual edges / maximum edges = e / emax



The actual edges (e) or linkages can be readily ascertained by counting. For a planar graph, the maximum number of edges (emax) may be computed from the expression 3(v-2) as explained by Taaffe and Gauthier (1973). In this expression v represents the number of vertices in the network. Thus Eq. 1 can be rewritten as:

γ = e / emax = e / (3(v-2))

Network connectivity as measured by gamma index can involve a set of nodes having no interconnections (i.e., an unconnected graph) at one extreme; the other extreme being a set of nodes in which every node has an edge connecting to it to every other node in the graph (i.e., a maximum connected one). The connectivity of a network is evaluated in terms of the degree to which the network deviates from an unconnected graph and approximates a maximum connected one. The numerical range for the gamma index is thus between 0 and 1. The Alpha Index In a minimal configuration of a network, only a single and a unique path can be identified between all pairs of nodes. Additional edges between nodes create circuitry. A circuit is defined as a finite closed path in which the initial node of a linkage sequence coincides with the terminal node. In practical terms, the existence of circuitry implies availability of alternative paths between nodes in the network. The alpha index is a ratio measure of the number of actual circuits to the maximum number possible. The actual and maximum number of independent circuits in a network are functions of the number of nodes in the network and the number of linkages necessary for minimal connection between nodes. Taaffe and Gauthier (1973) explained that the actual and maximum number of circuits in a given network can be found from the following two expressions:

Actual circuits = e - v + 1

Maximum circuits = 2v - 5

where, e and v have the same meaning as before. Thus alpha index can be expressed as:

α = (e - v + 1) / (2v - 5)Data Collection and Analysis

The analyses were primarily based on network and other information available from secondary sources. Data collected from the RHD and LGED sources and findings of the Bangladesh Transport Sector Study of 1993 and Bangladesh Integrated Transport System Study of 1998 were used in this study. However, collection of additional data from primary sources and newspaper clippings was also necessary to supplement the collected data from secondary sources.



In this study road network comprising only the national and regional roads under the RHD was considered. No feeder or other local roads were considered. This was done for the fact that feeder and other local roads are of much lower standards than the national and regional roads. As such, these roads are not capable of carrying heavy flows of national and regional traffic. Consequently, most of these roads do not supplement the national and regional roads. Furthermore, reliable information on their actual condition, length, alignment, etc. was also not available. A GIS-based analysis of the road network using PC ARC/INFO and other computing resources was carried out to accomplish the objectives of the study. LIMITATIONS OF THE STUDY Information on link failures only during the flood of 1998 was available from the RHD headquarters. Information on past failures was not readily available and so it could not be ascertained which link failures were more frequent. Feeder or local roads had to be excluded at all the stages of network analysis as reliable information regarding their alignment and condition was not available. In the absence of these information the present study was not able to make any specific recommendation for upgrading of existing selected feeder roads around a vulnerable link of the national network to improve the reliability of road transportation. Reliable data on water transport links around the broken road sections were also not available from secondary sources. Consequently, it was also not possible to ascertain if water transport can effectively play a supplementary role to road transportation at these locations. ROAD NETWORK OF BANGLADESH Roads in the rural areas of Bangladesh can be classified into four major categories such as national, regional, feeder, and other unclassified local roads. The Roads and Highways Department (RHD) maintains the national (category N), regional (category R), and feeder roads of type A (category F). The Local Government Engineering Department (LGED) has the responsibility for the type B of F category feeder roads and some of the other local roads. The rest of the local roads belong to local government agencies. Although both RHD and LGED have adopted standards for the various categories of roads under their jurisdictions, at present not all roads meet these specified standards according to their designated categories. However, it is expected that roads that do not meet the specified standards will gradually be upgraded to their specified standards.



Figure 1: National and regional road network of Bangladesh



Density of national and regional highways (categories N and R) in Bangladesh and their lengths in seven RHD zones is presented in Table 1. The total lengths of national and regional highways in the country are 2,996 km and 1,710 km, respectively. The combined network of these two categories of roads is provided by Fig. 1. It may be mentioned here that RHD has adopted a numbering system for the roads under their jurisdiction. However, they have not yet finalised numbering of all the roads. It is also understood that some current road identification numbers may get changed in future. As shown in Table 1, there is some degree of variation in road network density by zones. Rangpur Zone has the highest density with 3.95 km of road per 100 sq. km of area, closely followed by Rajshahi Zone with 3.80 km per 100 sq. km of area. Chittagong Zone has the lowest density with 2.33 km per 100 sq. km. The coefficient of variation (COV) is 17.9%, which indicates a moderate variation of road density by zone. Table 1: National and Regional Road Network Density in Bangladesh by RHD Zones

RHD Zone Road Length (km.) Total Area

(sq.km.)

Network Density per 100

sq.km. National

Highway Regional Highway

Total Length

(N) (R) (T = N+R) (A) (T / A) Dhaka 562 263 825 24111 3.42 Comilla 558 229 787 25297 3.11 Chittagong 387 104 491 21070 2.33 Rangpur 556 243 799 20203 3.95 Rajshahi 336 208 544 14310 3.80 Khulna 338 376 714 22274 3.21 Barisal 259 287 546 20305 2.69 Total (km) 2996 1710 4706 147570 - Mean - - - - 3.22 Standard Deviation - - - 0.58 Coefficient of variation - - - 17.9%

Source: Based on data obtained from RHD, Dec 1998



SIGNIFICANCE OF ROAD TRANSPORTATION IN BANGLADESH As already discussed, road transportation has become the most popular mode of transport in the country. During the last twenty years the share of road transportation for the whole country has increased from about 40% for goods and 50% for passenger traffic to about 75% for both the sectors. However, when movement along the five major corridors (as identified by the BTSS of 1993) are only considered, share of road transportation is somewhat lower. Tables 2 and 3 provide movement of freight and passenger traffic by the three main modes along the five major corridors. Figures 2 and 3 portray these five major corridors, separately for passenger and freight movements. Relative significance in terms of volume of traffic of the three modes namely, road, rail and water transportation is also shown in these figures.

The BTSS forecasts showed that the share of road transportation for freight traffic would increase from 60.5% in 1993 to 75.5% in 2000. During the same period the share of passenger traffic is expected to increase from 41% to about 49%. It is important to mention here that during this period volume of freight and passenger movements were forecasted to increase by 84% and 80%, respectively from their levels in 1993. These findings clearly show that the strategic transportation system of the country will continue to remain very heavily dependent on road transportation in future. Similar conclusions were also reached by an earlier study concerning overall demand of road transportation (PCI et al., 1985). Table 2: Freight Movement Along Major Transport Corridors of Bangladesh

Corridors 1993 2000 Road Rail Water All Road Rail Water All (Million tons p.a.) (Million tons p.a.)

Dhaka-Chittagong 6.5 1.2 2.8 10.5 13.2 1.7 3.5 18.4 Dhaka-Northwest 3.8 0.7 2.6 7.1 15.3 0.6 2.2 18.1 Dhaka-Khulna 3.8 - 2.5 6.3 0.4 - 2.4 2.8 Dhaka-Sylhet 2.6 0.3 1.9 4.8 4.4 0.5 2.5 7.4 Khulna-Northwest 2.1 0.3 - 2.4 8.9 0.3 - 9.2 Total 18.8 2.5 9.8 31.1 42.2 3.1 10.6 55.9 (%) 60.5 8.0 31.5 100 75.5 5.5 19.0 100

Source: Calculated from the BTSS figures, (BTSS, 1994)



Figure 2: Strategic transport corridors of Bangladesh: Freight movement



Figure 3: Strategic transport corridors of Bangladesh: Passenger movement



Table 3: Passenger Movement Along Major Transport Corridors of Bangladesh

Corridors 1993 2000 Road Rail Water All Road Rail Water All (Million p.a.) (Million p.a.)

Dhaka-Chittagong 7.2 12.8 3.2 23.2 30.7 25.1 10.0 65.8 Dhaka-Northwest 4.6 4.5 1.6 10.7 16.0 7.2 2.9 26.1 Dhaka-Khulna 11.9 - 10.7 22.6 10.5 - 9.9 20.4 Dhaka-Sylhet 4.3 4.6 3.4 12.3 5.7 6.1 3.5 15.3 Khulna-Northwest 3.6 4.7 - 8.3 6.6 7.9 - 14.5 Total 31.6 26.6 18.9 77.1 69.5 46.3 26.3 142.1 (%) 41.0 34.5 24.5 100 48.9 32.6 18.5 100

Source: Calculated from the BTSS figures, (BTSS, 1994)

FINDINGS OF NETWORK ANALYSIS The structural configuration of the road network (comprising national and regional roads) was analysed at three spatial levels to evaluate its degree of completeness and availability of alternative paths for traffic flow. The three levels were: the former district level, the divisional level and the national level (to capture the five major BTSS corridors of traffic flow). However, some adjustments were necessary to represent the actual spatial nature of traffic movement at these levels. The values for the two indices namely, gamma and alpha were calculated for the road network at these three spatial levels and are presented in Table 4. The gamma index measures the connectivity or degree of completeness of a network. For an unconnected graph its value is 0 and for a maximum connected graph the value is 1. This index can also be used to compare relative network connectivity by regions. There are three basic network configurations: spinal, grid and delta. The spinal pattern or the tree structure is the characteristic of a minimally connected network whereas the delta represents a maximally connected one. The grid configuration is a transition between the spinal and delta types.

Boundary values for the gamma index can be logically established by considering the characteristics of these three types of network are as follows (Taaffe and Gauthier, 1973): Spinal: 1/3 ≤ γ ≤ 1/2 where, v ≥ 4 Grid: 1/2 ≤ γ ≤ 2/3 v ≥ 4 Delta 2/3 ≤ γ ≤ 1.0 v ≥ 3



Table 4: Degree of Completeness of Road and Rail Network at Different Levels

Level Name Road network Rail network Gamma Alpha Gamma Alpha Index Index Index Index

Old District Sylhet 0.42 0.07 Mymensingh 0.37 0.00 Jamalpur 0.44 0.00 Dhaka 0.39 0.06 Chittagong 0.37 0.00 Tangail 0.39 0.00 Comilla 0.36 0.00 Noakhali 0.40 0.00 Rangpur 0.36 0.00 Dinajpur 0.39 0.00 Bogra 0.39 0.00 Rajshahi 0.36 0.00 Pabna 0.37 0.00 Kushtia 0.38 0.00 Jessore 0.40 0.05 Khulna 0.36 0.00 Faridpur 0.35 0.00 Barisal 0.38 0.00 Patuakhali 0.38 0.00 Division Sylhet 0.42 0.07 Dhaka 0.38 0.06 Rajshahi 0.39 0.07 Chittagong 0.36 0.02 Khulna 0.38 0.05 Barisal* 0.34 0.00 National All Districts 0.38 0.06 0.35 0.01

(*) Faridpur region included

Findings of the analyses as presented in Table 4 show that at all the three

spatial levels the values of gamma index varies between 0.34 and 0.44. This is a characteristic of a spinal type of network - that is the network is connected only to a minimal degree. The variations of the index values either at the old district or divisional levels are not statistically significant. The index values indicate that the network can still be considered at its early stage of development. There are very few alternative links for movement from one node to another node. Consequently, the majority of the paths of traffic flow are unique.



The alpha index indicates availability of alternative paths or circuitry between nodes in a network. Thus it can give a measure of network reliability. The values of this index are also provided in Table 4. At the old district level of movement, only two districts have a single circuit or alternative path of movement. At the divisional level, except Barisal, the situation is marginally better. However, the values of the index are still lower than 0.1. It shows clearly, that the existing network has a very few circuits and therefore, susceptible to high risk of traffic flow interruptions. The situation is not different at the national level also where very few alternative paths are available for movements. The boundary values of this index for the three types of network configuration - spinal, grid, and delta are also provided in Taaffe and Gauthier (1973). In a spinal type the value is 0. For a grid the value is greater than 0 but less than 0.5; and for a delta it is higher than 0.5. Accordingly, it shows that in terms of circuitry the network has a bare marginal grid configuration at the divisional (except for Barisal) and national levels. All the five major corridors of traffic movement are served by a very few alternative routes. It can be seen from Table 4 that the network characteristics of the rail network are even worse. Furthermore, the present network was developed about a hundred years back to serve a different pattern of traffic movement. Unless the network is reoriented focussing the present movement pattern, it would be difficult to regain any part of rail's lost share of traffic. However, if network were improved to increase the accessibility of the major centres by railway at a level comparable to that by road transportation, it would make some difference only for the share of traffic at the strategic level. FLOOD AFFECTED ROADS The 1998 flood was the most severe in the last 100 years in terms of magnitude and duration, extent of damage to properties and physical infrastructure and human sufferings. It started in early July and did not recede until September. The flood affected about 75% of the whole country.

Many roads were under water for more than a month. Typical road damage was wearing of surface, shoulders and embankments due to heavy flow of water over the road surface. In the more severe cases, road embankments were breached and the backfills to bridges and culverts were washed away. However, the proportion of damage to feeder or F category roads was much higher than to national and regional (categories N and R) roads. According to the flood damage assessment report of RHD, 22 bridges on national highways were washed away and a total of another 1263 bridges were assessed as having suffered flood damage (RHD, 1998).



Figure 4: Flood disrupted and affected road sections of Bangladesh in 1998



The disrupted and affected road links are shown in Fig. 4. In Dhaka Zone four links of national highways (N1, N2, N4 and N5) were disrupted and traffic movement on one regional highway (R460) was affected. The Kanchpur to Daudkandi link of N1 was totally disrupted which obstructed the traffic movement along the Dhaka-Chittagong strategic corridor. The disruption of the Kanchpur-Narshingdi link on N2 affected another strategic corridor between Dhaka and Sylhet. The traffic movement along Dhaka-Northwest strategic corridor was also severely affected due to two link failures on N4 and N5. The Kaliakoir-Tangail link on N4 was totally disrupted due to a number of bridge failures. On N5, the Mirpur bridge to Nabinagar link was deeply flooded. This caused disruption of normal traffic along both the Dhaka-Khulna and Dhaka-Northwest corridors. Two links of the national highway N1 were affected in Chittagong Zone. The Feni-Chittagong link affected movement along the strategic Dhaka-Chittagong corridor. It may be mentioned here that there is no alternative route to this affected link. The second affected link was between Cox's Bazar and Teknaf. In Comilla Zone, regional highway R140 was disrupted in one location between Lalmai to Chandpur and was affected at two other locations between Chandpur to Raipur and Raipur to Lakshmipur. These links play a significant role in movement within this region. All the affected and disrupted road links in Rajshahi Zone were on regional highways. The Rajshahi to Nawabganj link of R680 was totally disrupted. Three links of another regional highway R545 were also affected. These were from Naogaon to Mohadevpur, Mohadevpur to Patnitola, and Patnitola to Dhamoirhat. Only one regional highway in Rangpur Zone was affected. The link between Dinajpur to Phulbari of the regional highway R585 was affected. Consequently, direct movement between Dinajpur and Gobindaganj was in despair and corridor movement to the North-West was also partially distressed. In Barisal zone, the link between Mawa Ferry to Bhanga of national highway N8 was disrupted and the link between Tekerhat to Mostafapur was affected. Because of this, the movement along the Dhaka-Khulna strategic corridor suffered seriously. In this zone two other links between Mostafapur to Madaripur and Madaripur to Shariatpur of R860 were also affected. Only one regional highway R745 in Khulna Zone was affected on two different links between Trimohoni to Meherpur and Meherpur to Chuadanga. From the discussion presented above, it is observed that although there were not many link-failures of the national and regional roads, movements along all the five strategic transport corridors were severely distressed. As there are very few alternative routes in the existing network, a single link failure can cause a total disruption along a strategic corridor. It is observed further that most of the link failures were around Dhaka. As Dhaka is the centre of the highway system, the effects of any link failure around Dhaka do not remain limited to a particular



corridor but severely affects most of the other corridors and thus the whole road transportation system is distressed. CONCLUSIONS The transport system of Bangladesh has an overwhelming dependence on road transportation. As the spatial spread of the country is limited, road transportation has an inherent advantage over the two other major modes - rail and water. This has been further helped by the country's consistent public policy of investment for transport development favouring the road sector. Due to limitations of rail and water transport, the present overwhelming dependence on the road transportation system will continue in the future. In this context, completeness of the whole road network should be considered in all future road planning exercises to make the system less vulnerable in the event of any link disruption. Assessment of vulnerability of links around Dhaka deserves special consideration as any link failure of the national highways around Dhaka affects almost the whole road transportation system. To improve the reliability of the network, selected links of the existing feeder roads can be gradually upgraded to regional road standards so that alternative routes around the most vulnerable links of the existing national and regional roads can be created Reliable information on alignment and condition of the existing feeder roads was not available and as such this study did not make an attempt to identify any such link for upgrading. However, initial examination of the feeder road links around Dhaka City revealed that identification of such links is possible. For example, upgrading of Tongi-Ashulia link has greatly increased the reliability of Both N4 and N5. There are other similar links (between Tongi-Gazipur area and Bhairab-Narshigdi area) which can be considered to increase the reliability of N2 and N4. The analysis presented in this paper revealed that the road network development in Bangladesh is still in its early stage. The existing network has a very few circuits for movement at all the levels and therefore, susceptible to high risk of interruption. In terms of circuitry, the network has a bare marginal grid configuration at the divisional and national levels. Additional alternative paths to cater to the major movements need to be considered to improve network reliability and thus upgrade the network at least to a modest grid. REFERENCES Bangladesh Railway (1997), Information Book 1997, BR, Dhaka. Black, J. (1981), Urban Transport Planning, Croom Helm, London.



Khan, H. A. (1982), "Planning of road communication and development of highway system in Bangladesh", Journal of the IEB, Vol. 10, No 3, pp. 27-33.

PCI, PADECO, DDC (1985), Intermodal Transport Study, The Planning Commission, Government of Bangladesh, Dhaka.

Planning Commission (1998), Bangladesh Integrated Transport System Study, Government of Bangladesh, Dhaka.

Planning Commission (1993), Bangladesh Transport Sector Study, Government of Bangladesh, Dhaka.

Roads and Highways Department (1998), RHD Road Network Database Annual Report for 1997/98, Volume 1: Main Report, Government of Bangladesh, Dhaka.

Taaffe, Edward J. and H. L. Gauthier (1973), Geography of Transportation, Prentice Hall, New Jersey, USA.


Effect of Submergence on Subgrade Strength

S. J. Md. Yasin, Md. Hossain Ali, Tahmeed M. Al-Hussaini, Eqramul Hoque, and Sadik Ahmed

Department of Civil Engineering Bangladesh University of Engineering and Technology, Dhaka-1000, Bangladesh Abstract Damages of roads by floods are common phenomena in Bangladesh and a huge expenditure is required almost after each flood for rehabilitation of the roads. Therefore, research aiming at finding the modes of damages to roads under flood has become necessary. Several factors may appear to be responsible for such damages, which need to be confirmed by experiments. This study aimed at determining the effects of depth of submergence and duration of submergence on the subgrade strength of soil samples collected from the Dhaka-Aricha highway which was badly damaged by the 1998 flood. CBR tests were performed with different heights of submergence after normal soaking period and also after prolonged submergence. Index and identification tests were performed for classification and for determination of the suitability of the studied soils as subgrade material. For the studied depth and duration of submergence, no significant effect of submergence on sub-grade CBR strength could be found for any of the three types of soils tested. However, it was observed that all the three types of soils tested are rated as poor materials for subgrade according to different soil classification systems. INTRODUCTION Floods are recurring phenomena in Bangladesh and after each occurrence, they leave behind huge scar on the national economy. However, the flood of 1998 surpassed the damage records of all the floods of the recent past in terms of devastation. Other than surpassing in the maximum flood level of all recent


S. J. Md. Yasin, Md. Hossain Ali, T. M. Al-Hussaini, E. Hoque and Sadik Ahmed


floods, this flood persisted for a relatively longer period compared to other floods. As a result, apart from loss of lives and properties, the country suffered immense damage to the road networks. Therefore, it was felt necessary to undertake research to identify the causes of damage, due to this flood, to a road near Dhaka with a view to minimize such damage during future floods. SCOPE OF THE RESEARCH When floodwater recedes, it is generally observed that ditches and holes are developed in the road pavement. Just after the flood, these are usually found to be smaller in size and depth. With continued vehicular movements, these ditches continue to increase in size and depth by loss or removal of aggregates. When the subgrade gets exposed, the extent of damage increases very fast, making the roadway completely unusable. It is assumed that submergence of the road during flood might have some link with the initiation of the damage process. There are several mechanisms, which may be responsible for the damage. Firstly, with the rise of water level to the embankment top level, the road subgrade might lose strength due to reduction in effective stress. Continued vehicular movements may result in local failure in the subgrade, which then causes secondary failure in the pavement. Secondly, it is found that even after the submergence of road surface by floodwater, vehicles (specially heavy vehicles viz. trucks and buses) continue to ply as long as the axles do not go under water. Movement of series of vehicles may impose cyclic loading on the roadway resulting in pore-pressure build up in the subgrade. In this case also, primary failure may occur in the sub-grade, which may then propagate in other areas. Thirdly, it may be possible that the pavement is damaged first due to failure of binder (as a result of loss of effective stress) resulting in loss of aggregates and finally creating holes on the roadway. Among the aforementioned failure modes, only the first one was investigated in this research program due to limited scope of the research. TEST PROGRAM Generally California Bearing Ratio (CBR) test is the most widely used test for evaluating the strength of sub-grade, sub-base and base course materials for use in the design of road and airfield pavements. To determine the effect of submergence on subgrade strength, it was planned to collect disturbed soil samples from Dhaka-Aricha highway embankment (which was severely damaged by the 1998 flood) and carry out CBR tests with different water height above the soil surface in the CBR mold. During the inception of this research work it was


also planned that field CBR tests would be performed on selected damaged sections of the Dhaka-Aricha Highway immediately after recession of floodwater. However, by the time this research proposal was accepted and funds allocated, the road embankment became dry, repair works were done and the roadway became busy with traffic. So soil samples could not be collected from the subgrade at damaged sections of the road. Also considering the practical difficulties of carrying field-tests on such a busy and narrow highway (as part of the road way would have to be blocked), programs of field-tests had to be abandoned. Apart from CBR tests, Grain size distribution (sieve and hydrometer), Atterberg limits and Standard Proctor compaction tests were performed for classification and characterization of the soils. COLLECTION AND IDENTIFICATION OF SOIL SAMPLES The construction works for the lateral expansion of the Dhaka-Aricha highway between Gabtali to Savar was underway during the time of sample collection and three types of samples were collected from embankment soil near Savar Bazar. Sufficient amounts of samples of each type were collected into gunny bags and carried to BUET laboratory for investigation. One type of soil collected was grey in colour, non-plastic in nature and the grains formed little or no lump upon drying. It also contained some mica. This soil may be termed as fine sand or silt and has been designated as Soil-1 in this paper. This type of soil is usually found in char areas or on low lands and is formed of sediment deposits during flood. The other two types of soil can be classified as clay; one is the typical Dhaka clay (red) and the other is yellow clay. Upon drying these soils formed lumps having very high dry strength. The yellow and red clays have been designated as Soil-2 and Soil-3, respectively. Probably the sources of these soils were roadside borrow pits. All these soils were inorganic in nature. APPARATUS The standard apparatus used for CBR testing had to be modified to allow specimens to be tested with a maximum of 3 ft of water above the specimen surface. This height was selected considering weight limitations of the container to submerge the CBR mold, which is again governed by the weight capacity of the existing CBR apparatus and difficulties of placing CBR mold in a long narrow container. The lengths of shafts of the CBR apparatus and the plunger were also increased (Yasin et al., 2000).




INDEX PROPERTIES Physical properties of the collected soils were determined for classification purpose. The results are described in the following sections.

Particle Size Distribution and Grain Characteristics

Since all the samples contained considerable amount of fines, both mechanical sieve analyses and hydrometer analyses were performed on it to determine their grain size distribution and for classification. The grain size distribution curves of these soils are shown in Fig.1. All the soils tested can be termed as uniformly graded soils. Soil-2 and Soil-3 have virtually identical particle size distribution curves, whereas Soil-1 is composed of relatively coarser particles. The grain diameters D10, D30 and D60 determined from the grain size distribution curves are shown in Table 1 along with Specific Gravity and Fineness modulus values. Table 1: Mechanical properties of the soil samples tested

Soil Type / Designation Description Specific

Gravity Fineness Modulus

D10mm

D30mm

D60mm

Soil-1 Clayey silt; Grey 2.624 0.173 0.001 0.018 0.052

Soil-2 Yellow clay 2.611 0.008 ---- ---- 0.015 Soil-3 Red clay 2.584 0.010 ---- 0.003 0.012

Atterberg Limits

The Atterberg limits serve as excellent basis for expressing the state of consistency of fine-grained soils. Moreover, several classification systems are based on these limits. As Soil-1 is non-plastic, Atterberg Limit tests were carried out on the other two soils i.e., Soil-2 and Soil-3. The results are summarized in Table 2. Compared to Soil-2, Soil-3 has slightly higher values of liquid limit, plastic limit, shrinkage limit, plasticity index and flow index. Fig.2 shows the ‘flow curves’ obtained from the liquid limit tests. Table 2: Atterberg limits of the soil samples tested

Soil type / Designation Description Liquid

Limit Plastic Limit

Shrinkage Limit

Plasticity Index

Flow Index

Soil-1 Clayey silt; Grey * ----- ----- ----- ------ -----

Soil-2 Yellow clay 45.7 17.9 14.2 27.8 15.9 Soil-3 Red clay 52.5 22.0 16.5 30.5 18.2

* Nonplastic soil


10 1 0.1 0.01 1E-3 1E-40

20

40

60

80

100

Red clay (Soil-3) Yellow clay (Soil-2) Clayey silt (Soil-1)

Perc

enta

ge fi

ner

Grain size, mm

#40 0.420mm

US sieve no.

#200

#100

#70

#50

#40

#30

#16

#8#4

Sieve Opening

#200 0.074mm#100 0.149mm#70 0.210mm#50 0.297mm

#30 0.590mm#16 1.19 mm#8 2.38 mm#4 4.76 mm

Figure 1: Grain size distribution of soils tested

55 6 7 8 9 10 20 30 40 5050

44

46

48

50

52

54

56

58

25

Flow curve

Wat

er C

onte

nt (%

)

Number of Blows (Log scale)

52.5 %

Yellow clayFlow index=15.9

(Soil-2)

45.7%

Red Clay

(Soil-3)

Flow index=18.2

Figure 2: Flow curves for Soil-2 and Soil-3 from liquid limit tests




CLASSIFICATION OF THE SOILS TESTED Textural Classification

In general, the texture of a soil refers to its surface appearance. However, textural classification systems are based on different size-groups of particles present in any soil. A number of textural classification systems were developed by different organizations to serve their own need. To classify a soil according to a particular textural system, the particle size distribution curve is usually plotted and the percentages by weight of the particles contained within each of the ranges of size specified in the system are calculated. Table 3 compares the amount of principal components (sand, silt, clay and gravel) in the soils studied by several classification systems (Bowels, 1984; Peck et al., 1974).

Table 3: Comparison of principal components of soils tested according to different textural classification systems

Classification system

Soil Designation

Sand (%)

Silt (%)

Clay (%)

Colloid (%)

Soil - 1 35.7 50.5 13.8 ----- Soil - 2 6.6 50.3 43.1 -----

US Bureau of soils

Soil - 3 4.5 55.0 40.5 ----- Soil - 1 24.7 63.6 11.7 ----- Soil - 2 1.9 59.8 38.3 ----- Unified Soil - 3 4.1 95.9 ---- ----- Soil - 1 24.7 63.6 1.7 10.0 Soil - 2 1.9 59.8 34.5 3.8 AASHTO Soil - 3 4.1 95.9 ---- ----- Soil - 1 24.7 61.5 3.8 10.0 Soil - 2 1.9 55.0 39.3 3.8 ASTM Soil - 3 4.1 55.4 40.5 ----- Soil - 1 24.7 61.5 13.8 ----- Soil - 2 1.9 55.0 43.1 ----- FAA Soil - 3 4.1 55.4 40.5 ----- Soil - 1 40.8 47.5 11.7 ----- Soil - 2 10.8 50.9 38.3 ----- USDA Soil - 3 6.2 93.8 ---- ----- Soil - 1 35.7 52.6 11.7 ----- Soil - 2 6.6 55.1 38.3 ----- MIT Soil - 3 4.5 95.5 ---- -----


Classification Based on Use

Airfield Classification System / Unified Classification

Casagrande originally proposed this classification system in 1948 for use in the airfield construction works undertaken by the Army Corps of Engineers during World War II. This system was revised in 1952 in co-operation with United States Bureau of Reclamation as Unified system. In 1969, the Unified system was adopted by the American Society for Testing and Materials (ASTM) as a standard method for classification of soils for engineering purposes (ASTM D-2487). Fig.3 shows the Plasticity chart used in the Unified soil classification system. The values of Plasticity Index (PI) and Liquid Limit (LL) for Soil-2 and Soil-3 are plotted in Fig.3 to assign their classification symbol. Soil-1, which has 75.3% (more than 50%) material passing No.200 sieve and which is non-plastic, can be classified as ML.

Soil-2 (yellow clay) having 98.1% (more than 50%) material passing No.200 sieve and having liquid limit of 45.7 and plasticity index 27.8 can be classified as CL. Soil-3 (red clay) having 95.9% (more than 50%) material passing No.200 sieve and having liquid limit 52.5 and plastic limit of 30.5 falls into group CH. Unified System of soil classification grades ML and CL soil as “Fair to poor” and CH soil as “Poor to very poor” (Horonjeff and Mckelvy, 1994).

0 10 20 30 40 50 60 70 80 90 1000

10

20

30

40

50

60

(Soil-2)

(Soil-3)

PI=0.9(LL-8)Upper or U-line

74

CL-ML

MH or OH

ML or OL

CL or OL

CH or OH

Yellow clay

Red clay

PI=0.73(LL-20)A line

Silt

Clay

HighMediumLowPlasticity of clays

Plas

ticity

Inde

x, P

I

Liquid Limit, LL

Figure 3: Plasticity chart used by UNIFIED classification system




AASHTO Classification

The AASHTO classification system was developed in 1929 by the US Bureau of Public Roads. Since then it has undergone several revisions. The classification of considered here is based on system reported by Das (1985) (Yasin et al., 2000). Soil-1 (clayey-silt) having 75.3% (more than 35%) of total sample passing No.200 sieve and being non-plastic falls into the category A-4. Soil-2 (yellow clay) having 98.1% (more than 35%) material passing No.200 sieve and having liquid limit of 45.7 and plasticity index 27.8 can be classified as A-7-6. Soil-3 (red clay) having 95.9% (more than 35%) material passing No.200 sieve and having liquid limit 52.5 and plasticity index of 30.5 also falls into group A-7-6. As rated by AASHTO classification, all of Soil-1, Soil-2 and Soil-3 fall into the category “fair to poor” as subgrade material. The data for Soil-1 and Soil-2 are plotted on a plasticity chart along with the group symbols by AASHTO in Fig.4.

To evaluate the quality of a soil as a highway subgrade material, a number called the group index (GI) is also used in the AASHTO classification system along with the group or subgroup designation. In general the quality of performance of a soil as a subgrade material is inversely proportional to the group index. The group index is always reported to the nearest whole number unless its calculated value is negative whereupon it is reported as zero.

0 10 20 30 40 50 60 70 80 90 1000

10

20

30

40

50

60

70

(Soil-2)

(Soil-3)

A-2-4 & A-4 A-2-5 & A-5

A-2-7 & A-7-5

A-7-6

A-2-6 & A-6

Yellow clayRed clay

Plas

ticity

Inde

x, P

I

Liquid Limit, LL

Figure 4: Range of liquid limit and plasticity index for soil groups by

AASHTO


The group index is appended to the group and subgroup classification in parenthesis e.g., A-2-6(3). The group index is calculated as follows:

GI=(F-35)[0.2+0.005(LL-40)]+0.01(F-15)(PI-10) where, F = Percent passing No. 200 sieve

LL = Liquid limit PI = Plasticity index

This classification rates a soil as follows: (1) Poorer for use in road construction as one moves from left to right in the AASHTO classification chart (Das, 1985) i.e., A-6 soil is less suitable than A-5 soil, (2) Poorer for road construction as the group index increases for a particular subgroup, i.e., an A-6(3) is less satisfactory than A-6(1).

For Soil-1, GI=0, for Soil-2, GI= 29 and for Soil-3, GI=33. The values of GI for Soil-2 and Soil-3 are quite high indicating their unsuitability for subgrade construction. Since GI for Soil-1 is zero it may be considered as a relatively better material than the other two soils, but not a good quality material for subgrade, as indicated by the group designation.

COMPACTION AND CBR TEST

Compaction Characteristics

To determine the compaction characteristics, Standard Proctor compaction tests (ASTM D698; Test methods for moisture-density relations of soil and soil aggregate mixture using 5.5-lb rammer and 12 in drop) were carried out on the collected samples. Fig.8 compares the dry density versus water content curves from these tests. The Maximum Dry Density (MDD) and Optimum Moisture Content (OMC) values are presented in Table 4. As seen from Fig.5, all the three soils show quite different compaction characteristics. Among the three types of soils tested, Soil-2 has the highest MDD (108.3 pcf) with OMC of 17.7% and Soil-3 has the lowest MDD (102.6 pcf) with OMC of 21.3%. Soil-2 has an MDD of 106.8 pcf at an OMC of 16.8%. The zero air void curves in Fig.5 are plotted using the relationship

)/(1 SGw

G wd +=

γγ

to show the theoretical limits of density of these soils at different water contents. Here, G denotes specific gravity of soil particles, γw denotes wet unit weight at a moisture content of w and degree of saturation S. Specific Gravity values used for the computation of γd are given in Table 1.




10 12 14 16 18 20 22 24 26 28 3090

95

100

105

110at 100 % saturationZero air void lines

102.6 pcf

21.3

%

17.7

%

108.3 pcf

106.8 pcf

16.8

%

COMPACTIONCHARACTERISTICS

Clayey Silt Yellow Clay Red Clay D

ry d

ensit

y, p

cf

Water content, %

Figure 5: Dry density versus water content curves from standard Proctor

Compaction Test Table 4: Compaction characteristics of soil

Soil type / Description

Maximum Dry Density γd (MDD), pcf

Optimum Moisture Content (OMC), (%)

Clayey silt; Grey 106.8 16.8

Yellow clay 108.3 17.7

Red clay 102.6 21.3

CBR TEST California Bearing Ratio (CBR) tests were performed on all the three types of soils according to ASTM D1883-87. Tests were performed on soaked specimens without submergence and also with submergence of 3 ft of water. Specimens were prepared at optimum moisture content (Table 4) and with Standard Proctor compacting energy (ASTM D698) i.e., 5.5 lbs hammer, height of fall 12 in, number of layers 3 and 25 blows per layer. An automatic compactor was used for


preparation of CBR molds. During soaking, a surcharge of 10 lbs was applied to the specimens. The tests performed can be classified into three groups. In the first group, after a soaking period of 96 hours the specimens were drained for 15 minutes and then the CBR test was performed without any submergence. There are three tests in this group – one for each type of soil. In the second group, after a soaking period of 96 hours the specimens were placed in an empty cylindrical container placed on the base of the CBR apparatus. Then water was gently poured into the container until water level rose to 3ft above the soil surface in the mold. Then CBR test was carried out with submergence of 3ft of water. Three tests were conducted – one from each type of soil. Only one test was done for the third group, in which the specimen was soaked for 15 days and then tested with 3 ft of submergence. For all the samples, soaking was done with 6” of water above the specimen surface. Fig.6 compares the load penetration curves obtained in ‘group one’ and ‘group two’ tests on each types of soil. The CBR values corresponding to 0.1 inch and 0.2 inch penetration are compared in Table 5. No significant difference in the load – penetration response or CBR values could be observed between ‘group one’ and ‘group two’ tests.

These tests showed that there should not be any detrimental change in the subgrade strength due to submergence of a road by 3ft of water. Also the single test performed to see the effect of prolonged submergence did not reveal any change in the CBR characteristics of Soil-2 (Fig.7). Therefore, it can be concluded that the damage of the road in this case was not initiated in the sub-grade by reduction of effective stress due to submergence. However, other mechanisms such as hydrodynamic load and pore pressure build up due to passage of heavy vehicles during submergence might have played a key role in the failure of the road sub-grade and the pavement material. Although, initially it was planned to perform CBR tests with depths of submergence of 1 ft and 2 ft of water, these were later aborted due to time limitations. T

able 5: Comparison of CBR values

Soil type

CBR values at penetrations of

Without submergence

With 3 ft submergence

Tested after 15 days of submergence with

3 ft of water 0.1 inch 5.6 6.1 -- Soil - 1 0.2 inch 7.2 7.8 -- 0.1 inch 4.4 4.1 3.7 Soil - 2 0.2 inch 4.3 4.0 3.8 0.1 inch 4.7 5.1 -- Soil - 3 0.2 inch 4.6 4.6 --




0.0 0.1 0.2 0.3 0.4 0.5 0.60

50

100

150

200

error correction

Without sittingcorrectionsitting errorPlotted after

CBR test

With 3ft submergence Without submergence

Clayey siltSoil type - 1

Pene

tratio

n re

sist

ance

, psi

Penetration, inches

0.0 0.1 0.2 0.3 0.4 0.5 0.60

25

50

75

100

125

not requiredSitting error correction

CBR test


Yellow claySoil type - 2

Pene

tratio

n re

sista

nce,

psi

Penetration, inches

0.0 0.1 0.2 0.3 0.4 0.5 0.60

20

40

60

80

100

not requiredSitting error correction

CBR test


Red ClaySoil type - 3

Pene

tratio

n re

sista

nce,

psi

Penetration, inches

Figure 6: Comparison of penetration resistance among samples without submergence and with 3 ft of submergence


0.0 0.1 0.2 0.3 0.4 0.5 0.60

20

40

60

80

100

tested with 3 ft submergence water for 15 days & then sample kept under 3 ft with 3ft submergence

Yellow claySoil Type - 2CBR Test

without submergence

Pene

tratio

n re

sist

ance

, psi

Penetration, inches

Figure 7: Comparison of penetration resistance among samples with different conditions of submergence

CONCLUSIONS This research work was aimed at investigating whether the depth of submergence and duration of submergence during flood affects the sub-grade strength causing damage to a roadway. For the studied depth and duration of submergence, no effect of submergence on sub-grade CBR strength could be found for any of the three types of soils tested. However, this study points out that future research should be directed to other possible failure mechanisms such as failure caused by pore pressure build up in the sub-grade due to dynamic loading from the vehicles and change of properties of the pavement material itself due to submergence. Also it needs to be assessed whether subgrade soil stabilisation (during initial construction) will reduce the long-term cost (i.e., maintenance) of a roadway when poor quality soils are to be used in the subgrade.




REFERENCES Ali, M.H., Yasin, S.J.M., Al-Hussaini, T.M., Hoque, E. & Ahmed, S. (1999)

Damage of a road subgrade during 1998 flood, "BUET studies on 1998 floods in Bangladesh"-report., Bangladesh University of Engineering & Technology, Dhaka, December, 1999.

ASTM D698; Test methods for Moisture-density relations of soil and soil aggregate Mixture using 5.5-lb rammer and 12 in (305mm) drop.

ASTM D1883-87, Standard test method for CBR (California Bearing Ratio) of laboratory compacted soils.

ASTM D2487-85, Standard Test method for classification of soils for engineering purposes

Bowels, Joseph E. (1984) Physical and Geotechnical properties of soils, McGraw-Hill Intl. Eds.

Das, B.M. (1985) Principles of Geotechnical Engineering, PWS-KENT publishing company.

Horonjeff, R & Mckelvy, F. X. (1994) Planning and Design of Airports, Forth Edition, McGraw Hill Inc.

Peck, R.B., Hanson, W.E. and Thornburn, T.H. (1974) Foundation Engineering, Wiley Intl. Edns.

Yasin, S. J. M., Ali, M. H., Al-Hussaini, T. M., Hoque, E. and Ahmed, S. (2000), “Effect of submergence on subgrade strength”, report submitted to DAERS, BUET.


Design and Construction of Roads in Flood Affected Areas

M. J. B. Alam and M. Zakaria Department of Civil Engineering

Bangladesh University of Engineering and Technology, Dhaka-1000, Bangladesh Abstract Bangladesh is a flood prone country. Floodwater causes enormous damage to the road infrastructure of the country. Every year it is observed that the roads of the flood affected areas become unusable after flood. The process of rehabilitating the roads consumes a lot of valuable resources of the country. In this context, this paper investigates the vulnerability of the roads under prolonged submergence and presents a parametric study on the road construction materials that are affected by the long-term flooding condition. INTRODUCTION Flood is a perennial problem for Bangladesh. Almost every year it is observed that the roads of the flood affected area become unusable after flood. Lots of resources are required to rehabilitate the roads. Also the rehabilitation process needs valuable time during which the damaged roads remain unusable or only partially usable. During the reconstruction time congestion, vehicle operating cost and inconvenience of the users increases greatly. This is particularly important for major highways of the country. As roadway carries the largest share of passenger and freight movement of the country, any disruption in road network would impose significant loss to the economy of the country. In this context it is essential to design and construct roads, at least the major ones, that can withstand long-term flooding condition.


M. J. B. Alam and M. Zakaria


At present the roads of the country are designed and constructed using standard manuals such as Road Note 31 and AASHTO Design Guide. These manuals provide guidelines for adverse drainage condition and other environmental variables. Considering the environmental variables in the USA and UK, detailed studies have been performed on effects of these factors on road pavements in Transport Research Laboratory, UK and AASHTO Laboratory in USA. The major concerns in the design and construction of road pavement are stability of fills and slopes, drainage, capillarity and frost heave, permafrost, elasticity, rutting etc. Numerous research works have been carried out on these aspects in the laboratories mentioned above. As long-term inundation by floodwater is not a very common phenomenon in these countries, research in this regard is almost non-existent. Even in Bangladesh where flood is very common, comprehensive research on the effect of inundation by floodwater on road pavement has not been done yet.

This paper presents the results of the study on the effects of long term inundation on different layers of road pavement structure. The objective of the study is to identify the parameters of pavement structure, which are affected by flood. The results of the study is expected to help in modifying the present road pavement design and construction practice to build roads that are capable of withstanding stresses caused by long term inundation. FLOODS IN BANGLADESH: ITS EFFECT ON ROADWAYS Flood is a recurring problem in this extremely flat and riverine country. The earliest flood recorded in the history of Bengal is that of 1584-85 in the Meghna Basin, which caused the death of about 200,000 people and widespread damage to cattle, food-grain and crop (Hassan, 1998). In the current century and after independence in 1971, flood has hit this country time and again. Analyzing the data of flood incidences it is observed that the frequency of major floods has increased in recent years. Studies reveal that in the country moderate floods have occurred once in every two years while severe floods have occurred with an interval of 6-7 years. The scars of the devastating flood of 1988 were still alive in the memories of the people of this country when the flood of 1998 struck. In fact the country has just recovered from the devastation and finished the reconstruction and rehabilitation of the infrastructures damaged by the flood of 1988 when the flood of 1998 struck.

Flood of 1998 affected about 76 percent areas of the country. The floodwater had started its onslaught by the middle of June 1998 and stayed until the end of September, ’98. Considering the duration and extent of the flood of 1998 it is recognized as one of the most devastating floods in the history of this country.



The extent of damages of the roads of the country caused by the flood is summarized in Table 1. It shows that the total cost of the damage stands at about Taka 1641 crore which is about 11.4 percent of annual GDP of the country (Annual Flood Report, 1998).

After flood it is observed that most of the roads in flood-affected areas become badly damaged. The damages to the roads can be divided into two categories. One category of damage involves the failure of the embankment on which the roads are constructed. Usually this type of collapse is caused by the slope-failure of the embankment, which is the subject matter of geotechnical engineers. The other category of damage involves failure of the pavement structure. In this case, one or more layers of pavement structure may collapse resulting in the damage of the road. This also includes damage to the surface layer. The results of the study presented in this paper deals with the second type of damage only. The study investigates the effect of inundation by floodwater and its duration on strengths of different layers of pavement structure. Table 1: Assessment of Damages to Highways Caused by Flood of 1998

Damaged Road

(km) Damaged Bridges &

Culverts (No)

Rehabilitation Cost ( mil.

Taka)

RHD Roads

Length Sub-

merged(km) Emban

-kment Pave-ment

Badly Partial

No. of Dama-

ged Ferry Ghats

Imm. Repair

Cost ( mil. Taka)

Short Term

Long Term

Total Cost (mil.

Taka)

National Highway

1381.2 623.6 599.3 88 232 13 678.4 490.9 918.1 2087.8

Regional Highway

783.9 329.8 329.2 59 126 7 374.3 268.4 502.0 1144.8

Feeder Road

7457.5 3376.1 3315.9 229 470 34 3761.9 2720.5 5087.6 11570.0

Mecha. Equipment

- 30.0 70.0 0 100.0

Total 9622.6 4329.5 4244.4 376 828 54 4845.1 3549.9 6507.7 14902.6

Source: Roads and Highways Department.

METHODOLOGY AND RESULTS The main objective of the research project is to investigate the extent of damage to road pavement structure caused by inundation from floodwater. For this purpose, detailed laboratory tests have been performed on surface and subsurface layers of flexible pavement structure. As almost all of the major roads of the country are constructed as flexible pavement, only this type of pavement has been examined in the study. California Bearing Ratio (CBR) tests have been performed on samples of sub-grade layer and Marshal Stability and Flow tests



have been performed on the samples of surface layer. To simulate the effect of inundation by floodwater, the samples were kept in water for 4, 7, 30 and 45 days. To simulate the weathering action the samples were passed through alternate drying and wetting cycles. Sub-grade layer samples have been collected from Katchpur area along Dhaka-Chittagong highway and from Aminbazar area along Dhaka-Aricha highway. A total of 12 sets of samples were prepared for the purpose of testing.

Table 2 and Fig. 1 show the effect of inundation on the density of the test samples. Table 3 and Fig. 2 presents the effect of the same on strength of sub-grade material measured in terms of CBR value. Table 2: Effect of Inundation by Flood on Unit Weight of Sub-grade Material

Average Unit Weight (pcf) Compaction

(No of Blows) 4-day

Soaking 7-day

Soaking 30-day

Soaking 45-day

Soaking 56 108 106 104 103 35 103 100 98 97 10 94 89 86 84

0 10 20 30 40 50

50

60

70

80

90

100

110

120

C om paction w ith 56 B low s C om paction w ith 35 B low s C om paction w ith 10 B low s

Uni

t Wei

ght

N um ber o f D ays o f Inunda tion

Figure 1: Effect of Inundation on Density of Sub-grade



Table 3: Effect of Inundation by Flood on CBR of Sub-grade Material

Average California Bearing Ratio (CBR) Value Compaction (No of Blows)

4-day Soaking

7-day Soaking

30-day Soaking

45-day Soaking

56 3.5 3.4 3.1 3.0 35 2.7 2.5 2.2 1.9 10 1.6 1.2 1.1 1.0

0 10 20 30 40 50

0

1

2

3

4

5

Com paction w ith 56 B lows Com paction w ith 35 B lows Com paction w ith 10 B lows

CBR

(%)

Num ber of Days of Inundation

Figure 2: Effect of Inundation on CBR of Sub-grade

From these results, it is evident that both unit weight and CBR reduces with the number of days of inundation by water. As the road pavement is designed with the CBR value of 4-day soaking, relative changes are calculated on the basis of this value. In the case of inundation for 45-days, the unit weight reduces by 4.6, 5.8 and 10.6 percent for compaction efforts of 56, 35 and 10 blows, respectively. In the case of CBR the reductions are 16.7, 29.6 and 37.5 percent, respectively. This implies that the more compact the material, the lower will be the loss of unit weight and strength caused by inundation. It is expected that the



sub-grade soil become saturated within four days. The subsequent loss of strength and unit weight may be caused by the loss of fine particles from the sample.

For surface layer, the samples were prepared in the laboratory and the optimum bitumen content was estimated. It was found that the optimum bitumen content is 4.75 percent. Marshall test samples have been prepared for surface layer used for medium traffic using the same aggregate and bitumen content mentioned above. Four sets of samples have been prepared to test the initial strength and flow as well as the effects of inundation for 4, 7 and 30 days with alternate drying and wetting cycles. The results of Marshall Stability and Flow tests on the samples are presented in Table 4 and Fig. 3.

0 5 10 15 20 25 30

0

10

20

30

40

50 Flow Values

Mar

shal

Flo

w V

alue

Number of Days of Inundation

1000

1500

2000

2500

3000

Stability Value, lbs

Stability Value

Figure 3: Effect of Inundation on Flow and Stability of Bituminous Surface

Layer

From these results, it is evident that stability and flow of flexible pavement is affected by the duration of inundation by water. In the case of inundation for 30-days, the flow value increases by about 93 percent and stability reduces by 26 percent. The figures also imply that the longer the period of inundation, the more severe will be the deterioration although the rate of destruction may decrease.



Table 4: Effect of Inundation by Flood on Unit Weight of Sub-grade Material

Number of Days of Inundation

Flow Stability (lb)

Initial Values 14 2476 4-Day Inundation 19 2145 7-Day Inundation 21 2002 30-Day Inundation 27 1826

0 5 10 15 20 25 30

50

60

70

80

90

100

Inde

x of

Ret

aine

d St

reng

th (%

)

N um be r o f days o f Inunda tion

Figure 4: Effect of Inundation on Stripping and Swelling of Bituminous Mixture

The results of Marshall Stability tests can also be interpreted in the form of Immersion-Compression Test (AASHTO Designation T165) which is widely used as an indirect measure of the tendency of aggregates to strip or swell under the effect of moisture (Oglesby and Hicks, 1982). The test result is presented as a



numerical index of resistance of bituminous mixture to the detrimental effect of water as shown below.

100(%)Strength Retained ofIndex 1

2 ×=SS

where, S1 = Strength of dry specimen, and S2 = Strength of immersed specimen.

The results of the analysis are presented in Fig. 4. It shows that the strength of the sample reduces by 26 percent for 30 days of immersion. This loss can be attributed to stripping or swelling of bituminous mixture caused by water.

CONCLUSIONS From the results of the study it can be concluded that the period of inundation by floodwater affects the strength of pavement layers significantly. In the case of inundation of road surface for 45 days the CBR of sub-grade material may reduce by about 30 percent for medium compaction. Also in the case of surface layers the strength reduces substantially.

The results of the study imply that long-term inundation should be considered as a design parameter for designing roads in Bangladesh. As flood is a perennial problem for Bangladesh and measures to prevent flood is not economically feasible, strategies to cope with flood may prove to be economically more justified. Considering the cost of rehabilitation of roads after the flood of 1998, it can be concluded that even the increased initial cost to construct roads capable of withstanding inundation by flood would prove to be much better in economic terms.

In this study only two layers of pavement structure and some specific parameters of the materials have been examined. In order to develop working guidelines, detail testing and analysis is required. ACKNOWLEDGEMENT The authors acknowledge the contributions of Mr. A.K. Bhuiyan and Mr. S. Kabir as research assistants in the project.



REFERENCE Hassan, K. M. (1998), Flood 1998: The Condition of the Highways in

Bangladesh, Paper Presented in a Seminar on “Impact of the 1998 Flood on the Economy of Bangladesh”, Dhaka, September, 1998.

Annual Flood Report (1998), Flood Forecasting & Warning Centre, Processing & Flood Forecasting Circle, Bangladesh Water Development Board.

Oglesby, C. H. and Hicks, R.G. (1982), Highway Engineering. Fourth Edition, John Wiley and Sons, USA.


Effect of Flood on Earth Structures: A Case Study

M. A. Ansary and A. M. M. Safiullah

Department of Civil Engineering Bangladesh University of Engineering and Technology, Dhaka-1000, Bangladesh Abstract Flood is a recurring problem in Bangladesh. In the middle of 1998, Bangladesh experienced the most devastating and prolonged flood in its history that caused serious disruption of the economy of the country. All the major rivers of the country namely the Padma, the Meghna, the Brahmaputra and the Jamuna flowed above the danger level during the 1998 flood. During this flood, earth structures such as flood, rail and road embankments, bridge abutments and piers were threatened. The road links between the capital city and other major cities were disrupted. About two-third of the country went under water. In order to develop strategies to mitigate the miseries of people in the event of a future flood, it is necessary to record and analyze the damages that had occurred due to the 1998 flood. In this study, two damaged earth structures have been analyzed and measures to protect earth structures have been suggested. INTRODUCTION In 1998 Bangladesh experienced one of the worst floods in terms of severity, destructiveness and duration. The rivers went above danger level in July 1998 and remained above the danger mark up to the second half of September 1998. Never in the history of flood in Bangladesh, it was so protracted and the sufferings of the people so great. Incessant rainfall for days together added to the misery of people. All the major rivers of the country namely the Padma, the Meghna and the Brahmaputra and the Jamuna flowed over the danger level.




People felt helpless in the hand of nature. Earth structures such as flood, rail and road embankments, bridge abutments and piers were threatened. The road link with the capital city was disrupted. The two-thirds of the country went under water. Many people left their houses for a shelter. Some lived on embankments and on raised lands under the open sky. Those who did not move from their home, lived on the roofs of their houses along with their livestock. Men, women and children had to swam through floodwater to fetch drinking water, food, fuel and medicine.

In order to develop a strategy to mitigate the miseries of people in the event of a future flood, it is necessary to record and analyze the damages that had occurred due to this flood. To rehabilitate damaged infrastructures, it is essential to note the nature of damages that had occurred during the flood. Extent of damage and its nature and cost for repair in terms of money and capital need to be assessed to arrive at a rehabilitation and repair strategy and also in planning and design of future infrastructure facilities. This study is a compilation of information on damages of earth-structures during the 1998 flood. The study also looks into the nature of damages and aims at classifying them and suggests possible remedial measures. CHARACTERISTICS OF THE 1998 FLOOD In the monsoon of 1998, due to excessive rainfall in the upper catchment areas from July to September, and intermittent rainfall within Bangladesh, all the rivers of the country experienced significant increase in flow far above the danger level. The flood situation started to become alarmingly worse from the middle of July. By this time the low-lying areas of the country had already gone under water. At that time, about 45,000 square kilometer areas of 37 districts of the country were affected by flood. Although flood situation started improving in early August, the flow of the two main rivers Padma and Brahmaputra-Jamuna increased significantly in the middle of August due to heavy rainfall in the upper catchment areas. By the end of August flood situation became worse and about 60,000 sq.km area of 42 districts were affected. During the early September the flow of the major rivers increased abruptly worsening the situation. The flood situation became worst in the second week of September and about 75,000 sq.km area of 52 districts were affected during that time. The flood prevailed from early July to the last week of September, for more than three months at different places in different magnitudes. Thus the flood of 1998 became the most prolonged flood in the history of Bangladesh.

Figures 1 shows hydrographs of 1988 and 1998 floods for river Buriganga at Dhaka and Brahmaputra-Jamuna at Serajgonj. In 1998, the river

Effect of Flood on Earth Structures

stage at the Hardinge Bridge point exceeded all its previous records. It can be observed from the hydrographs that although flood stage for 1988 flood was higher at most places, duration of water level above danger mark was much longer during the 1998 flood.

1988

1988

1998 1998

1988

(a) (b)

Figure 1: Flood hydrographs of (a) Buriganga River at Dhaka and (b) Brahmaputra River at Sirajganj, during the 1988 and 1998 floods

FLOOD DAMAGE The 1998 flood broke all previous records for its duration and devastation. Large-scale damage occurred all over the country. According to The Annual Flood Report (1998), this flood affected about 68 per cent area of Bangladesh. Table 1 shows statistics on damages due to 1998 flood, which was provided by the Emergency Operations Center (EOC, 1998) of the Ministry of Disaster Management and Relief.

According to Bangladesh Institute of Development Studies (BIDS, 1998), the 1998 flood may have caused a damage worth Tk.. 10,228 crore to different sectors of the national economy including infrastructure, industry and agriculture. The agricultural sector topped the list of the worst affected sectors, with estimated damages to the extent of Tk. 5052 crore, of which crop sector alone accounted for Tk. 4377 crore and non-crop sector Tk. 675 crore. Infrastructure suffered damages worth Tk. 3949 crore, including Tk. 1087 crore in roads and bridges. Tk. 153 crore in the railways, Tk. 313 crore in embankment and irrigation canals, Tk. 273 crore in educational institutions, Tk. 33 crore in health




centres and facilities and Tk. 2090 crore in the residential sector. The damage to industries was estimated to be Tk. 1227 crore of which large industries lost Tk. 222 crore, small and medium industries lost Tk. 904 crore and cottage industries lost Tk. 100 crore. Table 1: 1998 Flood damage statistics (EOC, 1998)

Number of affected districts 52 Number of affected thanas 366 Number of affected unions 3323 Number of affected people 3,09,16,351 Crop damage (acre) 14,23,320 Number of damaged houses 9,80,571 Number of deaths 918 Number of dead livestock 26,654 Damaged road in km 15,927 Damaged bandh (embankment) in km 4,528 Number of damaged bridge/culverts 6,890 Number of damaged educational institutions 1,718 Number of damaged shelters 2,716 Number of sheltered people 10,49,525

The severe flood of 1998 damaged most of the infrastructures including the National, Regional and Feeder roads under RHD. At present RHD owns about 20,285 km road network (2,862 km National Highway, 1,565 km Regional Highway and 15,860 km Feeder roads) and during the 1998 flood 1381 km National Highway, 784 km Regional Highway and 7458 km Feeder roads were submerged. Figure 2 presents percentage cost of road damages according to RHD zones and Fig. 3 presents comparison between some damaged and undamaged roads. Figure 4 shows that the damage to road pavements have been extensive in most greater districts. EFFECT OF FLOOD ON DIFFERENT STRUCTURES The structures under this study are from infrastructure projects such as road-communication network, flood protection embankments and irrigation projects. The patterns of damage observed in different civil structures due to the 1998 flood are briefly summarized below; details are available in Safiullah and Ansary (2000).


RAJSHAHI7%

RANGPUR5%

BARISAL18%

KHULNA11%

CHITTAGONG11%

COMILLA 26%

DHAKA 22%

Figure 2: Percentage of total cost resulting from road damages according to

RHD zones

Length of road (km)

Figure 3: Comparison between damaged and undamaged roads

Damage to Road and Railway Embankments

Overtopping

This is very common in road embankments where the road top level is set below high flood level. In such situation, when adequate allowance for flood water level or ground settlement is not considered at design and construction stage, water rises above the top of the embankment. The usual practice to prevent overtopping during emergency is to create a floodwall on the two sides of the road usually by




dumping bags filled with soils. There is considerable amount of seepage through this temporary barrier and provision is needed for pumping out seeped water. Pavement Damage

This is a common occurrence when flood level reaches the base of the pavement. Due to running of vehicles high pore water pressure develops below the edge of the pavement. High pore water pressure during wheel traction may pump out soil from underneath the pavement edge thereby progressively damaging the pavement structure. A variety of damages such as development of pot holes, ruts, etc may develop as a result. Erosion of Slopes and Scour

Erosion of slopes may develop due to various reasons but the most common mechanism is soil removal at water level due to wave action on the slopes. Removal of soil from wave level creates a undercut steep slope which later collapses due to instability and as a result waterline regresses inwards. Sometimes scour may result at the toe of the slope inducing slope failure. Damage to Flood Embankments

The difference between flood and road embankment is that in the former case a considerable difference in water level exists on two sides of the embankment. This results in well-established seepage lines. Various conditions that may deteriorate the stability of a flood embankment include: (a) Seepage, (b) Leakage, (c) Overtopping, (d) Piping, (e) Settlement, (f) Scour, (g) Attack by Rodent, and (h) Slope Failure. Damage to Bridges

Common types of failures of bridges due to high flood usually result from development of scour around piers or abutments and removal of soil support near the toe. Instability to bridge pier or support system and slope stability may result from: (a) Scour of pier support/abutment, (b) Removal of soil providing passive resistance, (c) Progressive erosion/scour and bypassing through road embankment, and (d) Failure of bridge support system. Damage to Houses and Buildings

Floodwater may submerge buildings and induce damages of various degrees from peeling of plaster to total collapse depending on the nature and condition of flood and of the building. Scouring of soil cover over foundation may reduce bearing capacity of shallow foundations and thereby foundation instability may result.


Damage to River Bank Protection Work

River bank protection works include riprap and falling apron over a considerable length of the riverbank. High velocity currents during flood may remove the protective covers resulting in bank failures. The mechanisms that may interplay to destabilize riverbank protection works include: (a) Scour, (b) Slope failure, (c) Removal of protection layer, and (d) Collapse. TWO CASE STUDIES OF FLOOD DAMAGES In this paper, case studies of an affected railway embankment and a damaged riverbank protection work have been presented. Permanent Protective Work for Railway Embankment along Ishurdi-Sirajganj Route

The Railway Embankment requiring protective work is located within the Chalan Beel area. The rail line within this area runs across a very low lying plain and the height of embankment varies between 13 and 20 ft. The higher elevation relates to sections near bridge crossings. The highest flood level, average normal flood level and, variation in ground level and formation level of the embankment for the route is shown in Table 2. The datum shown is that used by Bangladesh Railway and not related to PWD datum. Table 2: Ground, formation and flood levels for the railway embankment

Reduced Level (Railway Datum)

Ishurdi-Sirajganj Route

Highest Flood Level in 1998 (ft) 38.50 Average Normal Flood Level (ft) 34.50 Ground Level (ft) 25.50 – 29.25 Formation Level (ft) 39.00 – 45.00

The railway embankment is very old and most of the ground settlement has

already taken place. There is no problem with settlement in the embankment. The embankment lies above water level during the dry period of the year. But during the monsoon, the water level rises very close to the embankment top. The highest flood level is reported to be at RL 38.5 ft. (railway datum). It was learnt that during the 1998 flood, the embankment nearly overtopped at some locations and rail communication had to be stopped for a few days. Vast water surface of the beel is subjected to wind pressure and wind generated waves were responsible for




major damage of the embankment slope in the form of soil erosion and slips on the slope.

Figure 4 shows typical cross-section of railway embankment in Ishurdi-Sirajgonj route after the 1998 flood. The section shows considerable erosion of the existing slope. From an analysis of the available cross-section, it was noted that most of the damage occurred above RL. 30 ft (railway datum). There was no evidence of scour at the toe of the embankment suggesting that the instability of the embankment was mainly associated with high water level and wave action. Protection measures to protect slopes against wave erosion may be taken to slopes lying above RL.30 ft.

The general strategy for preventing damage to embankment during flood is to provide appropriately designed revetment to cover the embankment slope within the portion vulnerable to wave attack. In addition to wind-generated waves, waves may also be generated from boats navigating close to the embankment and this should be considered in designing the revetment work. Within the Railway embankment areas visited, no evidence of deep sliding was observed. At some locations longitudinal cracks at the top of the embankment were noted which appeared to have resulted from instability due to loss of soil from toe and mid portion of the embankment slope. There was evidence of very little scour near the embankment borrow pit areas. From Fig. 4, it is apparent that originally the embankment had a uniform side slope of 1 (Vertical): 2 (Horizontal), which was considerably modified by the process of removal of soil by erosion, particularly within the zone of waves action. The present slope shown in Fig. 4 reflects typical S-shaped profile usually generated by wave action on a slope.

Figure 4: Typical cross-sectional profile of railway embankments in

Ishurdi-Sirajganj route


At one location revetment system with aggregate filter and cement concrete (CC) blocks was installed by the Railway. The protective system appeared to be quite effective against wave attack during the 1998 flood. But at some points damage to revetment system was noticed due to loss of soil underneath aggregate filter resulting in subsidence of the CC blocks. Therefore adoption of a design similar to that already used may be considered adequate, provided provision is made to prevent migration of soil particles from underneath the CC blocks. Use of a geotextile layer may serve this purpose (Safiullah and Ansary, 2000). Serajgonj Town Protection

The construction of the 2.5 km long revetment works at Sirajganj forms part of the Phase I Priority Works of the River Bank Protection Project (RBPP) on the right bank of Jamuna. One of the major components of the Works is the construction of a bank stabilization structure comprising a precast concrete block armoured revetment laid on a geotextile filter, with a falling apron. Following some slides during construction, the revetment slope was modified for parts of the Works and the apron level was also raised at some locations. On 15th August, 1998, when the Works were nearing completion, a part of the revetment (around 60 m long extending from A260 to about A320) collapsed. Further slides occurred upto 18th September, the total length of damaged sections being about 300 meters. The Construction Supervision Consultants (CSC), in co-operation with the Contractor, initiated some mitigatory measures, which included dumping of CC blocks. A team of specialists mobilised by CSC reviewed the damage and a report was submitted to BWDB on 20th October 1998.

The water levels and near bank velocity assumed in design are based on a 100-year return period using simulation over a 25-year period (1965-89). However, during 1993-95 considerable additional data were collected under FAP24 (River Survey Project), which do not appear to have been used by the Consultants while reviewing/updating the design. A comparison of scour depths obtained by different methods shows that none of the predicted scour depths were less than 40 m with respect to 100-year flood level. The formulae yielded an average value of 48 m from 100-year flood level. The Consultant’s design of falling apron, however, considered 33 m scour depth for the upstream face and 29 m for the straight part, although the maximum scour depth obtained in the physical model tests was 43.5 m. The observed scour depth during the flood of 1998 matched very closely with the scour depth predicted in the Physical Model Study, while the design scour depth was much less than the observed scour depth. The setting level of the falling apron (–4.2m PWD) appears to be too high.

One of the surprising features of the whole project is the lack of field information on geotechnical parameters. Most of the geotechnical parameters used in the design and reported in the analysis of slope stability, relate to general soil conditions without being site specific. The only boring was performed after




the 1998 flood at one location. An analysis of the data shows that most of the soil down to –2m PWD is in a very loose state with N-values between 4 and 7. From this depth downward, the soil density increases slightly, down to –14m PWD, below which the soil is very dense. The percentage of fine material is very high, particularly when compared with the soil in the West Guide Bund (WGB) of JMBP. Whereas the fine content of the soil in WGB is limited to 12% only, it is as high as 80% in the upper 25 m of soil in Sirajganj. After the flow slides during construction of WGB in late 1995 and early 1996, the Consultants for the RBPP project were asked to review the design at Sirajganj. It appears that they failed to recognize the differences in the soil parameters in the two locations. The relative density of the soil at Sirajganj between +11m PWD and –2m PWD is less than 30%. Sand size mica content has been found to be around 5% to 10%, but it is the resultant low value of relative density, which determines the slope stability. No effort appears to have been made by the Consultant to assess the relative density during original design or review of design after WGB, JMBP failures. Calculations based on the slope failure records during construction suggest that a mobilized angle of friction φ between 17o and 18o would produce failure of the dredged slopes. Above –10m PWD, the RBPP site has a much lower relative density (<30%) than the WGB, JMBP site. However, below –10m PWD, soil at WGB site has a lower relative density (< 40%) compared to Sirajganj site. As the soil above –12m PWD at Sirajganj RBPP site has a relative density less than 40%, it is likely to develop flow slides.

The maximum water level during the 1998 flood at Sirajganj was +14.76m PWD that was approximately 1 m below the design high flood level (HFL). The peak discharge of Jamuna was around 90,000 m3/s, which has a return period of nearly 20 years. Velocity measurement at structure B2, Sirajganj during 14-16 September 1998 gave a maximum velocity of 3 m/s while the design velocities for revetment along straight section, upstream termination and head of groyne are 3.7, 4.4 and 4.8 m/s, respectively. Thus the flood of 1998 in the Jamuna at Sirajganj was not unusual in terms of peak stage, peak discharge and velocity. The 1998 flood was extreme in terms of total duration of flood flow, but took a serious turn with the third failure of structure at Sirajganj on the 27th August 1998.

The flow slide that occurred in the Sirajganj structure in the 1998 flood was due mainly to the presence of loose micaceous sandy soils present in the banks of the Jamuna. Studies on the Jamuna soil indicate that it exhibits wholly contractant behavior at relative density of 55%. These soils may show contractant behavior even at a relative density value of 60% for simple shear and triaxial extension shear stress paths. Due to contractant behavior of sand, pore pressure is induced in these soils during undrained deformation making them prone to failure by flow sliding. A number of failures of similar nature during construction


of the West Guide Bund of the Jamuna Multipurpose Bridge (late 1995 to early 1996) testify to the susceptibility of the Jamuna river bank soil to flow sliding.

Available information indicate that between depth +13.5m PWD and –16.0m PWD the soil in the right bank of the Jamuna river has a very low relative density with a mean value of only 32% and standard deviation of 7%. This soil has virtually no reserve shear strength as established in various previous investigations. The first flow slide occurred on the 15th of August when the bottom of scour hole was at depth of –17.0m PWD, which is the level of scour assumed in design. At this depth of scour, had the slope been properly designed, the protection works should have prevented the slope from collapse. This failure, which occurred at this depth of scour, indicates that some important parameter had not been taken into consideration in the design. A study of the available information points to the fact that very low value of relative density of soil is the main factor responsible for failure. As low value of relative density is a common phenomenon in the Jamuna bank soils, the Consultants should have taken into consideration its effect in their design.

According to the BRTC-BUET report (1998), bed levels around the failure sections are much lower than –17m PWD. Thus it appears that the river has already played most of the hydro-morphological activities by scouring at the vicinity of the upstream termination of RBPP. In the event of future flood with magnitude and duration similar to the 1998 flood, the scour depth likely to occur will be comparable to the 1998 event. It is also likely that the scour depth at the upstream termination would not be exceeded significantly in the future flood events. Hence if any remedial measures are taken using the existing geometry of scoured area (which is to some extent trapped), it is likely to be stable against future floods and is expected to provide long-term solution. The idea behind this solution is to stabilize the existing slope profile from the top of the scar to the deepest part of the scour hole. The main work may be divided into two parts: (a) filling up of the depression over the scar face to form a slope not steeper than 1:3.5, and (b) placement of armour such as CC blocks or stone (Madhyapara hard rock) of appropriate thickness over the filled up slope.

The filling of the depression over the scar face may be done by using four alternative materials as indicated below in order of decreasing cost: (i) Sand filled jumbo-size geotextile bags: This is the most costly alternative but has the advantage that geotextile bags will act as a filter and separation layer in retaining soil; (ii) Madhyapara hard rock (stone): Assorted sizes of rocks may be used but use of graded material is preferable as grading may be designed to enhance filtration and separation capability. Also use of rock materials will permit a slope steeper than 1:3.5 for the filled section; (iii) Sand filled polypropylene bags: These bags will serve the purpose of filtration and separation, but may not be as effective as the geotextile bags. Such sand filled polypropylene bags have been used by the BWDB in the Sailabari groyne; (iv) Sand filled gunny bags: This




may serve a function similar to sand filled polypropylene bags but has the disadvantage that the bags are likely to tear during handling and deteriorate with time.

Developing a slope covered by filler materials and armour layer from the deepest part is likely to stabilize the existing slope and arrest probable future soil flow towards the deep trough. The work should be carried out in a careful manner so as to ensure correct thickness of armour as well as correct aerial coverage. CONCLUSIONS The major floods, like the one of 1998 have significant impact on the earth structures of Bangladesh. Even though the 1998 flood exceeded their respective design flood levels, different flood protection projects were able to withstand the flood because of the preventive measures taken by the concerned Government Agencies. In this study two damaged earth structures were studied in detailed and development of strategy to protect earth structures during the floods were discussed.

Two case studies reported in this study show different types of attack during high flood of 1998. In the first case, it is seen that a significant damage can result from wave action during high flood. Most embankment structures in Bangladesh are built with silty clay with low cohesion and is likely to be eroded unless wave protection works are undertaken. The second case shows the effect of bed scour due to high flow velocity at bed level. In this case local scouring at the bed level of the earth structures make the slope steeper and subsequently collapse at some parts of the structures take place. So some type of preventive measures have to be undertaken to protect the structures. ACKNOWLEDGEMENT The authors would like to thank the Bangladesh Railway and Bangladesh Water Development Board (BWDB), for their assistance and co-operation during the field visits. REFERENCES Annual Flood Report 1998, Flood Forecasting and Warning Centre, Bangladesh

Water Development Board.


EOC (1998), Report of Emergency Operations Centre 1998, Ministry of Disaster Management and Relief.

BIDS (1998), Report on 1998 Flood Damage, Bangladesh Institute of Development Studies.

BWDB (1998), Report on damage to infrastructure during 1998 flood and the probable costs of rehabilitation. Monitoring and Evaluation Office, Bangladesh Water Development Board.

Safiullah, A.M.M. and Ansary, M.A (2000), Assessment of Damaged Structures during the 1998 Flood, Report on Flood 1998, CASR, BUET.

BRTC-BUET (1998), Investigation into the damage to the works of contract B2 at Sirajganj during monsoon 1998, Submitted to Bangladesh Water Development Board.



Evaluation of Performance of Rajshahi Town Protection Embankment

Abu Siddique, Md. Hossain Ali, Md. Shariful Islam

and Md. Monwarul Islam Department of Civil Engineering

Bangladesh University of Engineering and Technology, Dhaka-1000, Bangladesh Abstract The 18 km Rajshahi Town Protection Embankment was threatened during 1998 floods. Seepage, piping, sliding and rain cuts occurred at a number of sections of the embankment. Considerable portions of both the T-Groyne and Shashanghat Closure (I-Groyne) were damaged. The left bank of the Padma eroded severely. A number of protective measures were undertaken to prevent seepage and erosion of the embankment, groynes, spurs and closures which included bamboo pilling, dumping of rock boulders, sand bags, brick crates and brick bats. Polythene cover was used to prevent erosion due to rain cut. Re-sectioning and thorough repairing of the flood embankment had to be done to raise the embankment section up to the design level. In December 1998, Bureau of Research, Testing and Consultation, Bangladesh University of Engineering and Technology, Dhaka carried out field and laboratory investigations. The detailed results of the field and laboratory investigations have been presented in this paper. Based on results obtained from the present investigation, the embankment soils have been found to be in a state of fairly good degree of compaction. Results of stability analyses indicated that despite the significant damages that occurred during the floods, the embankment retained adequate factor of safety against bearing capacity failure. Moreover, timely implementation of protective measures and good in-situ soil parameters played key roles in the overall successful performance of the embankment during the floods.


Abu Siddique, Md. Hossain Ali, Md. Shariful Islam and Md. Monwarul Islam


INTRODUCTION During 1998 floods significant damages occurred to the Rajshahi Town Protection Embankment. The embankment and its components were threatened during the floods. Seepage, piping, sliding and rain cuts occurred at a number of sections of the embankment. During the occurrence of floods, a number of protective measures were undertaken to prevent seepage and erosion of the embankment. In December 1998, after the floodwater had receded, the Bureau of Research, Testing and Consultation (BRTC), Bangladesh University of Engineering and Technology, Dhaka carried out field and laboratory investigations. The major objectives of the investigations were to assess the performance of the embankment during 1998 floods and to examine the existing condition of the embankment.

This paper presents the extent of damage occurred in the embankment during the floods and the preventive measures taken during the flood to prevent erosion of the embankment and its components. Future plan of protective works has been addressed in this paper. The results of the field and laboratory investigations have been presented in this paper. Finally, using the results of field and laboratory investigations, an attempt has been made to assess the present condition of the embankment with particular emphasis on the existing compaction state and overall stability of the embankment. RAJSHAHI TOWN PROTECTION EMBANKMENT Rajshahi Town protection embankment was constructed in 1932 to protect the town from flood. Initially it was 16.9 km long and later increased to a length of about 18 km having 20 single sluice gates and brick mattress in some places. The height and constructed slope of the embankment were 21.3 m (above MSL) and 1: 2 (vertical: horizontal), respectively. Only about 2 km on the upstream side is paved and brick mattress was provided on the slope of riverside. With the development of river training work on the Indian side, the main stream of the Padma started flowing vigorously towards Rajshahi town and the city was threatened. In view of that, 3 groynes and 6 spurs were constructed including brick mattress of 1676 m in 1979-80. A closure was constructed near Shashanghat during 1987-88. The embankment is provided with a number of groynes (T-Groyne and I-Grorne), spurs and closures. T- Groyne is an earthen groyne made of silty clay. Head of the T- Groyne is covered by brick mattressing and boulders are placed mainly on the upstream side. Height of the crest of T-Groyne is 21.3 m (MSL). An I-Groyne is placed about 1.3 km down of dam head, which is also an earthern groyne having a length of approximately 500 m.

Performance of Rajshahi Town Protection Embankment

Head portion of this groyne is covered by brick mattress. Another I-Groyne was constructed 2 km down of the above mentioned I-Groyne, having a length of 1 km. Construction details of this groyne are the same as the other I- Groyne. The 2B-Spur situated 800 m from the second I-Groyne was constructed with brick box. A total of 220 bricks weighing 750 kg is netted by wire and placed one upon another. The width of the spur is staged. On the top the width is about 10 ft. From here up to the T-Groyne, the riverbank is provided with brick mattress. Other 5 spurs were constructed at the down of T-Groyne in several places having lengths of 150 m to 200 m. These are also earthen spurs with brick mattress at the nose. A closure was constructed at Shashanghat having a length of 500 m with 200 ft brick mattress at the nose. DAMAGES OCCURRED DURING THE 1998 FLOOD The embankment was threatened during the 1998 flood with seepage, piping, sliding, and rain cuts occurring at different sections of the embankment. The velocity of water current during the 1998 floods was recorded as approximately 18 m/s. The highest water level recorded during the floods was 19.68 m, the highest level ever recorded and this water level was more than 1 m above the danger level. The major thrust took place at the T-Groyne. Initially on 18th August 1998, piping on the nose took place and the nose settled. Later, on 30th August 1998, downstream nose and shank were attacked and damaged badly. About 200 m2 of downstream shank was washed away. The Shashanghat Closure (I-Groyne) near Panchabate was also threatened during the 1998 flood and considerable portion from the head was damaged. The other groynes and spurs were not attacked by the flood significantly, except the 2B-Spur. The netting of brick blocks was tore off which needed major repairing. The current took away a 150-ft section of the nose of closure. The left bank of the Padma eroded severely, specially, at a location 2 km downstream of the T-Groyne and in some places in upstream. This area shall be protected using brick mattress or other suitable type of bank revetment measures. Re-sectioning and repairing of the flood embankment and its components has to be done thoroughly to revive the section up to the design level. PROTECTIVE WORKS UNDERTAKEN Several types of protective works were undertaken for the protection of the embankment from erosion due to flood. Photographs of various protective measures are shown in Figs. 1 to 3. The following protective works were undertaken: (i) Bamboo Pilling: Extensive bamboo piling was carried out for the




protection of embankment against erosion; (ii) Dumping of Boulders: Boulders were dumped on the upstream and downstream of groynes and closure for their protection; (iii) Dumping of Sand Bags: Gunny bags filled with local sand were dumped in order to prevent seepage, erosion of groynes, spurs and closure. Approximately 1,35,000 sand bags were dumped, of which 85000 bags were dumped at the T-Groyne; (iv) Percupine: This is a method in which bricks were encased in bamboo boxes. These boxed were placed at the locations where boulders and sand were found to be washed away; (v) Brick Crates: Bricks were placed in a netted wire and placed on embankment slopes for their protection; (vi) Dumping of Brick Bats: Gunny bags filled with brick bats were dumped at several places of the embankment for erosion protection; (vii) Polythene Cover: This was used to prevent erosion due to rain cut. The whole embankment (from upstream toe to the downstream toe) was covered with polythene when heavy rainfall took place. A total of 1200 m polythene cover having width of 30 ft was used.

With the implementation of the various protective measures during the flood, it was possible to protect the embankment from catastrophic failure. The Surface Water Modelling Center (SWMC), Dhaka has been requested to conduct a mathematical model study on the 1998 floods. After the completion of mathematical and physical model studies, the type of appropriate protective measures could be ascertained and permanent protective measures could be undertaken. Meanwhile overall repair of the damaged T-Groyne and other infrastructure including the flood embankment should be undertaken as soon as possible. For the 1998-99 financial year, Bangladesh Water Development Board has taken up the Rajshahi Town Flood Protection Embankment (5th Term) Project which includes: (i) Repair works of groynes: Taka 290 lac; (ii) Repair and strengthening of foot of 2B-Spur: Taka 50 lac; (iii) Brick mattress in upstream and downstream of groynes and cement concrete blocks at the foot (1100 m): Taka 580 lac; (iv) Brick mattress work at Char Kazla (250 m): Taka 50 lac; (v) Resectioning of embankment (5 km): Taka 50 lac; and (vi) Repair of brick mattress (3 km) : Taka 100 lac. POST FLOOD GEOTECHNICAL INVESTIGATION A detailed field investigation was carried out at the Rajshahi Flood Protection Embankment after the 1998 flood. Laboratory tests were also conducted on disturbed and undisturtbed soil samples collected from the embankment. The major objectives of the soil investigation program were as follows: (i) Determination of stratigraphic sequences of sub-soil strata by drilling of boreholes; (ii) Evaluation of consistency and relative density of the sub-soil by carrying out Standard Penetration Test (SPT); (iii) Identification


Figure 1: Photograph showing bamboo piling for erosion protection

Figure 2: Photograph showing dumped rock boulders on groyne of the embankment




Figure 3: Photograph showing dumped sand bags and brick crates on the embankment slope

and classification of the embankment soil and the foundation soil by carrying out index property tests in the laboratory; (iv) Evaluation of the existing compaction state of the embankment soils by carrying out in-situ density test and Standard Compaction Test in the laboratory; and (v) Assessment of the strength, compressibility and permeability properties of the soil by carrying out consolidated undrained direct shear test, unconfined compression test and one-dimensional incremental loading consolidation test. FIELD INVESTIGATIONS AT THE EMBANKMENT Five boreholes were drilled vertically at this site using wash boring technique. Wash borings of small diameter (approximately 100 mm) were drilled by water flush aided by chiselling. The depth of boreholes below the surface varied from 30 ft to 50 ft. The density and stiffness characteristics of the sub-soil layers in the boreholes were measured by performing Standard Penetration Test (SPT) at 1.5 m (5 ft) intervals by means of standard 50.8 mm outside diameter split-spoon sampler. Disturbed and undisturbed samples were collected from the boreholes. A split-spoon sampler was used to obtain the disturbed samples. Undisturbed


samples were also retrieved from cohesive layers of the boreholes by pushing conventional 76 mm external diameter thin-walled Shelby tubes following the procedure outlined in ASTM D1587 (ASTM, 1989). Siddique et al. (1999) reported summary of the sub-soil stratification for these boreholes.

Density of embankment soils in place was determined at ten locations by the Sand Cone Method as outlined in ASTM D1556 (ASTM, 1989). In order to assess the degree of field compaction, maximum dry density of four categories of the embankment soils collected from four different locations of the embankment was determined in the laboratory. LABORATORY INVESTIGATIONS Index Properties of Soil Samples

The results obtained from index property tests on different samples are presented in Table 1. The natural moisture contents for the samples obtained from boreholes BH-1, BH-2, BH-3, BH-4 and BH-5 varied from 21.9 % to 42.4 %, 23.9 % to 32.2 %, 19.5 % to 28.1 %, 21.5 % to 27.9 % and 18.1 % to 36.4 %, respectively. The values of liquid limit, plastic limit and plasticity index of the cohesive samples varied from 35 to 46, 20 to 28 and 8 to 26, respectively. From the particle size distribution curves, percent clay (< 0.002 mm), percent silt (0.002 mm to 0.06 mm) and percent sand (0.06 mm to 2 mm) were determined using MIT Texural Classification System. Percent finer than number 200 sieve and fractions of sand, silt and clay of the samples tested are presented in Table 1. Using the results of index property tests, soil samples obtained from the embankment have been classified according to Unified Soil Classification System (USCS) as outlined in ASTM D2487 (ASTM, 1989). Classifications of the cohesive and non-cohesive soil samples are also shown in Table 1. The cohesive samples obtained from the boreholes are typically inorganic clays and silts of low to medium plasticity (USCS Symbols are CL and ML) while the sandy samples are either SM or SP-SM. Moisture-Density Relations and Assessment of In-situ Compaction of the Embankment

Moisture-density relationships were determined for four samples obtained from four selected locations of the embankment, namely, slope of T-groyne, Buattala, Noboganga and Shashanghat. Each test was carried out following the procedure outlined in ASTM D698 (ASTM, 1989). The moisture-density relations of the four samples are presented in Fig. 4. From the moisture-density curves, optimum moisture content (wopt) and the corresponding maximum dry density for the soil samples were estimated. The values of the wopt and maximum dry density of the




samples varied from 15.1 % to 19.5 % and 100.5 lb/ft3 to 110.4 lb/ft3, respectively. In-place density of the embankment soil was determined at ten spots of the above-mentioned four locations of the embankment, following Sand Cone Method. A summary of the field density test results is shown in Table 2. It can be seen from Table 2 that the in-situ dry density and compaction of the soil samples at four locations of the embankment varied from 84.3 lb/ft3 to 105.1 lb/ft3 and 79.5 % to 95.2 %, respectively. The average value of the degree of field compaction at the ten spots was found to be 85.3 %. Although the embankment suffered significant erosion and that the embankment soils were not compacted after the floodwater had receded, an average compaction of 85.3 % indicates that the embankment soils are still in a state of fairly good degree of compaction. Table 1: Summary of index properties of soil samples

Grain Size Distribution Borehole

No./ Sample No.

Depth

(ft)

LL

PL

PI

USCS

Symbol % Finer No. 200 Sieve

Sand (%)

Silt (%)

Clay (%)

BH-1 /UD-1 8.0 to 9.5

36 28 8 ML 91.5 12 76 12

BH-1 / D-1 5.0 46 26 20 CL 99.0 3 70 23 BH-2 / D-2 and D-3

10.0 & 15.0

40 28 12 ML 98.0 2 85 13

BH-2 / D-7 and D-8

35.0 & 40.0

- - - - 41.2 64 32 4

BH-3 / UD-1 8.0 to 9.5

37 20 17 CL 87.4 14 72 14

BH-3 / D-3 and D-4

15.0 & 20.0

35 21 14 CL 88.0 13 75 12

BH-3 / D-6 30.0 36 22 14 CL 73.7 26 60 14 BH-4 / UD-1 8.0 to

9.5 38 24 14 CL 93.2 7 80 13

BH-4 / UD-2 13.0 to 14.5

- - - SP-SM 7.5 - - -

BH-4 / D-4 and D-5

20.0 & 25.0

46 20 26 CL 92.7 10 61 29

BH-5 / UD-1 8.0 to 9.5

- - - SM 16.8 - - -

BH-5 / UD-2 13.0 to 14.5

- - - SP-SM 5.9 - - -

BH-5 / D-7 35.0 - - - - 42.5 62 35 3 Direct Shear Test Results

Five consolidated undrained direct shear tests were carried out on the undisturbed samples obtained from the boreholes. In each test, cylindrical samples of 63.5


mm diameter by 25 mm height were initially consolidated using three different normal loads and subsequently sheared under undrained condition. From the failure envelopes of the samples, the values of cohesion (c) and angle of internal friction (φ) of the samples have been determined. A summary of the direct shear test results is presented in Table 3. Comparing the values of angle of internal friction obtained for the samples with those reported by Terzaghi and Peck (1967), it can be concluded that the relative density of samples tested from boreholes BH-2, BH-3 and BH-4 are loose while that for the two samples tested from borehole BH-5 are dense.

Figure 4: Moisture-density relations of selected embankment soil samples

Unconfined Compressive Strength of Soil Samples Unconfined compressive strength tests were carried out on two undisturbed samples, one from borehole BH-1 and the other from borehole BH-3. From the stress-strain data, unconfined compressive strength (qu), and axial strain at failure (εf) were determined. A summary of the unconfined compression test results is presented in Table 4. On the basis of the value of undrained shear strength, which is half of the unconfined compressive strength for clays, the samples UD-1 of BH-1 and UD-1 of BH-3 are soft and firm, respectively.




Table 2: In-place dry density and % compaction of soil samples

Sample Location

Optimum Moisture Content

(%)

Maximum γd in

Laboratory (lb/ft3)

Sample No.

In Place Water

Content (%)

In Place Dry

Density (lb/ft3)

In-situ Compaction

(%)

1 35.6 87.6 87.2 2 36.4 86.6 86.2

Slope of T-Groyne

19.5 100.5

3 31.3 90.8 90.3 1 17.1 90.1 81.6 Buattala 15.1 110.4

2 23.7 94.0 85.1 1 18.7 95.7 86.7 2 16.3 105.1 95.2

Noboganga 19.4 110.4

3 17.1 89.6 81.2 1 21.3 84.3 79.5 Shashanghat 15.4 106.1

2 20.2 85.0 80.1 Table 3: Summary of direct shear test results

Borehole No.

Sample No.

Depth (ft)

Average Water

Content (%)

Average Dry

Density (kN/m3)

Cohesion (kN/m2)

Angle of Internal Friction (Degree)

BH-2 UD-1 8.0 to 9.5 22.1 16.15 20 27 BH-3 UD-1 8.0 to 9.5 18.6 15.89 22.5 27 BH-4 UD-1 8.0 to 9.5 25.2 15.28 20 29 BH-5 UD-1 8.0 to 9.5 27.5 14.45 0 41 BH-5 UD-2 13.0 to 14.5 28.9 14.66 0 39

Table 4: Summary of unconfined compression test results

Borehole No.

Sample No.

Depth (m)

Water Content

(%)

Dry Density

(kN/m3)

Value of qu

(kPa)

Value of εf(%)

BH-1 UD-1 8.0 to 9.5 ft 21.2 15.24 72.6 6 BH-3 UD-1 8.0 to 9.5 ft 18.1 16.88 146.9 11

Compressibility and Permeability Properties

Compressibility and permeability properties of one sample obtained from borehole BH-1 were determined from incremental loading one-dimensional consolidation tests. Void ratio versus logarithm of effective vertical stress plots and coefficient of consolidation versus logarithm of effective vertical stress plots for the sample are presented in Fig 5. Compression index (Cc) of the sample, determined from the slopes of the loading portion of the void ratio versus


logarithmic of pressure curve shown in Fig. 5, was found to be 0.17. The initial void ratio (e0) of the sample was found to be 0.69. Low values of Cc and e0 indicate that the compressibility of the sample is low. Depending on the stress range, the values of coefficient of consolidation and coefficient of volume compressibility of the soil sample have been found to be in the range of 1.49 x 10-3 to 3.71 x 10-3 cm2/sec and 0.88 x 10-4 to 9.8 x 10-4 m2/kN, respectively.

Figure 5: Compressibility plots of a sample: (a) void ratio versus log of effective vertical pressure, (b) cv versus log of effective vertical pressure Coefficient of permeability of the samples was determined indirectly from

one-dimensional consolidation tests. Depending on the void ratio, the values of coefficient of permeability of the soil samples varied from 1.29 x 10-10 m/sec to 3.29 x 10-9 m/sec. Void ratio versus logarithm of coefficient of permeability plot has been presented in Fig. 6. It can be seen from Fig. 6 that, the relationship between void ratio and permeability is approximately linear. The average slope of the relationship, which is termed the permeability change index (Tavenas et. al., 1983), Ck for this sample is 0.13. The ratio of compression index to permeability change index, i.e., Cc/Ck for the sample is 1.3. Berry and Wilkinson (1969) reported




that for many soils Cc/Ck often lies within the limits of 0.5 and 2.0, while Mesri and Rokhshar (1974) observed that the experimental values of Cc/Ck were found to vary between 0.5 and 5.0.

Figure 6: Void ratio versus coefficient of permeability plot for a sample STABILITY ANALYSIS OF THE EMBANKMENT The overall stability failure mechanism is development of slip circles resulting in a deep sliding surface. This is a conventional soil mechanics stability problem. Pre-existing slip planes within the soil, or lenses and bends of cracker material can have a significant effect on slope stability. Stability analyses were carried out to evaluate the factor of safety against bearing capacity failure of the embankment for a number of conditions. The following four conditions have been considered: (i) deep stability of the embankment on the river side during high water period; (ii) deep stability of the embankment on the river side during low water period; (iii) deep stability of the embankment on the country side during high water period; and (iv) deep stability of the embankment on the country side during low water period


XSTABL program has been used for the stability analysis. XSTABL performs a two-dimensional limit equilibrium analysis to evaluate the factor of safety for a layered slope using the Simplified Bishop Method or the Janbu Method. Based on the data of the present investigation, stability analyses of the embankment were performed for two different sections of the embankment with heights of 20 ft and 10 ft on the riverside. The height of the embankment on the countryside for both the sections was taken as 8 ft with berm. Soil properties were also varied for the embankment sections. Two types of soil properties for the embankment soil were considered while the properties of the foundation soil were kept the same for all the analyses. Table 5: Summary of the results of stability analyses

Soil Properties

Foundation Embankment

Analyses

No.

EmbankmentHeight

on River Side

Height of Water on

River Side

Position of

Sliding Surface

Factor

of Safety

1 20 ft 15 ft R/S 1.88

2 20 ft 0 ft R/S 2.10

3 20 ft 15 ft C/S 4.98

4 20 ft 0 ft C/S 4.97

5 10 ft 7.5 ft R/S 2.51

6 10 ft 0 ft R/S 2.77

7 10 ft 7.5 ft C/S 5.26

c = 20 kN/m2

φ = 28° γm=17.0kN/m3

γs=18.5 kN/m3

8 10 ft 0 ft C/S 5.26

9 20 ft 15 ft R/S 1.56

10 20 ft 0 ft R/S 1.72

11 20 ft 15 ft C/S 2.19

12 20 ft 0 ft C/S 2.19

13 10 ft 7.5 ft R/S 1.68

14 10 ft 0 ft R/S 1.73

15 10 ft 7.5 ft C/S 2.19

c = 0 φ = 25° γm=15.5kN/m3

γs=17 kN/m3

c = 0 φ = 40° γm=16.0kN/m3

γs=17.5 kN/m3

16 10 ft 0 ft C/S 2.19 Note : γm = Moist unit weight of soil; γs = saturated unit weight of soil; R/S = River side; C/S = Country side

A two-layer soil model has been used. The stability of the embankment was

evaluated for a degree of consolidation equal to zero. Circular failure surfaces were assumed for all the eight cases of analyses as mentioned above. All together




sixteen analyses were carried out. Table 5 shows the values of factor of safety of the embankment for the analyses performed. It can be seen from Table 5 that the factor of safety of the embankment on the riverside for heights of 20 ft and 10 ft varied from 1.56 to 2.10 and 1.68 to 2.77, respectively. However, the factor of safety of the embankment on the countryside has been found to be higher than those obtained for the riverside. This has been attributed to low height of the embankment on the countryside and also due to the presence of berm on the countryside. The factor of safety of the embankment on the countryside has been found to vary from 2.19 to 5.26. The results of stability analyses, therefore, indicate that that the embankment has adequate factor of safety against bearing capacity failure and that the embankment has adequate overall stability. CONCLUSIONS During 1998 floods considerable damages occurred to the Rajshahi Town Protection Embankment. The 18 km embankment and its components were threatened during the 1998 floods. Seepage, piping, sliding and rain cuts occurred at a number of sections of the embankment during the occurrence of flood. Major damages took place at the T-Groyne and the Shashanghat Closure (I-Groyne). About 200 m2 of downstream shank of the T-Groyne and a considerable portion from the head of the I-Groyne were damaged. The netting of brick blocks was teared off which needed major repairs. The current took about 150 ft section of the nose of closure away. The left bank of the Padma eroded severely, specially, at 2 km downstream of the T-Groyne and in some places in the upstream.

A number of protective measures were undertaken to prevent seepage and erosion of the embankment, groynes, spurs and closures which included bamboo pilling, dumping of rock boulders, sand bags, brick crates and brick bats. Polythene cover was used to prevent erosion due to rain cut. With the implementation of the various protective measures, it was possible to protect the embankment from catastrophic failure. Due to considerable damage of the embankment and its components, re-sectioning and thorough repairing of the flood embankment had to be done to raise the embankment section up to the design level.

In December 1998, Bureau of Research, Testing and Consultation (BRTC), Bangladesh University of Engineering and Technology, Dhaka carried out field and laboratory investigations. The field investigations at the embankment consisted of drilling of boreholes, performing Standard Penetration Test (SPT), collection of sufficient numbers of disturbed and undisturbed tube samples, and performance of in place density tests. A total of five boreholes were drilled. A detailed laboratory investigation was carried out on soil samples collected from the boreholes. The cohesive samples obtained from the boreholes are typically


inorganic clays and silts of low to medium plasticity (USCS Symbols are CL and ML) while the sandy samples are either SM or SP-SM. The values of the optimum moisture content and maximum dry density of the samples collected from four selected locations of the embankment varied from 15.1 % to 19.5 % and 100.5 lb/ft3 to 110.4 lb/ft3, respectively. In-place density of the embankment soil was determined at ten spots of the above-mentioned four locations of the embankment. The in-situ dry density and compaction of the soil samples at four locations of the embankment varied from 84.3 lb/ft3 to 105.1 lb/ft3 and 79.5 % to 95.2 %, respectively. The average value of the degree of field compaction at the ten spots was found to be 85.3 %, indicating that the embankment soils are still in a state of fairly good degree of compaction. Direct shear tests conducted on five undisturbed samples indicated that the angle of internal friction varied from 27° to 41°. On the basis of the value of undrained shear strength, two samples tested from two boreholes were found to be soft and firm, respectively. The values of compression index (Cc) and initial void ratio (e0) of a sample has been found to be 0.17 and 0.69, respectively. Low values of Cc and e0 indicate that the compressibility of the sample is low. Depending on the stress range, the values of coefficient of consolidation and coefficient of volume compressibility of the soil sample have been found to be in the range of 1.49 x 10-3 to 3.71 x 10-3 cm2/sec and 0.88 x 10-4 to 9.8 x 10-4 m2/kN, respectively. Depending on the void ratio, the values of coefficient of permeability of the soil sample varied from 1.29 x 10-10 m/sec to 3.29 x 10-9 m/sec. The relationship between void ratio and permeability has been found to be approximately linear. The average slope of the relationship, termed as permeability change index, Ck is 0.13 for this sample and the ratio of compression index to permeability change index (Cc/Ck) for the sample is 1.3.

Stability analyses were carried out to evaluate the factor of safety against bearing capacity failure of the embankment for a number of cases. Deep stability of the embankment on the river side and country side during both high water and low water periods have been investigated using XSTABL slope stability program. A total of sixteen analyses were carried out. The factor of safety of the embankment on the riverside for heights of 20 ft and 10 ft varied from 1.56 to 2.10 and 1.68 to 2.77, respectively. The factor of safety of the embankment on the countryside has been found to vary from 2.19 to 5.26, which are substantially higher than those on the riverside. This is due to low height of the embankment and also due to presence of berm on the countryside. The results of stability analyses indicate that although the embankment suffered considerable damages during the 1998 floods, the existing factor of safety against bearing capacity failure of the embankment is adequate.

The present investigations clearly demonstrate that although the embankment showed distresses, e.g., local erosion, rain cuts, sloughing, etc. at various sections, the overall performance of the embankment has been satisfactory. The satisfactory values of soil parameters, adequate in-situ density




and timely undertaking of protective measures played key roles in the overall successful performance of the embankment during the 1998 floods.

REFERENCES

ASTM (1989), "Annual Book of ASTM Standards", Volume 04.08, Soil and Rock, Building Stones; Geotextiles.

Berry, P.L. and Wilkinson, W.B. (1969), "Radial Consolidation of Clay Soils", Geotechnique, Vol. 19, No. 2, pp. 253-284.

Mesri, G. and Rokhshar, A. (1974), "Theory of Consolidation for Clays", Journal of the Geotech. Engng. Div., ASCE, Vol. 100, No. GT8, pp. 889-904.

Siddique, A., Ali, M.M. and Islam, M.S. (1999), "Performance Evaluation of Rajshahi Town Protection Embankment During 1998 Floods", Department of Civil Engineering, Bureau of Research Testing and Consultation, Bangladesh University of Engineering and Technology, Dhaka.

Tavenas, F., Jean, P., Leblond, P. and Leroueil, S. (1983), "The Permeability of Natural Soft Clays. Part II: Permeability Characteristics", Canadian Geotechnical Journal, Vol. 20, No. 4, pp. 645-660.

Terzaghi, K and Peck, R. B. (1967), "Soil Mechanics in Engineering Practice", Modern Asia Editions, Tokyo, Japan.

NOTATION c cohesion cv coefficient of consolidation Cc compression index Ck permeability change index e void ratio LL liquid limit mv coefficient of volume compressibility PL plastic limit PI plasticity index qu unconfined compressive strength wopt optimum moisture content γd dry density of soil γm moist unit weight of soil γs saturated unit weight of soil φ angle of internal friction εf axial strain at failure


Grain Size Distribution of Materials Deposited Over Floodplain Due to Embankment Failure

S.J. Md. Yasin, Md. Hossain Ali, Eqramul Hoque, Sadik Ahmed

and Rezaul Hoq Department of Civil Engineering

Bangladesh University of Engineering and Technology, Dhaka-1000, Bangladesh Abstract Rivers in Bangladesh carry huge amount of sediment load and as a result flood embankment failure causes sedimentation over the floodplains. Nature of the sediments e.g., grain size and distribution are important factors to be considered in a numerical model for estimating the risk and extent of damage by sedimentation. Samples collected at random from the sediments deposited due to failure of the Dharala and Dudkumar river embankments by the 1998 flood show considerable variation in grain size and gradation. It has been argued that numerical models need to be calibrated for such wide range of gradations in deposits. INTRODUCTION Failure of river embankments during flood is a frequent natural phenomenon in Bangladesh and often it results in sedimentation over the floodplain. Apart from untimely damage to crops, this sedimentation entails potential long term damage to crop cultivation in the surrounding areas. The long-term damage potential depends on many factors, of which the characteristics of sediments carried by the river is most important. The present study is aimed at determining the grain size distribution of sediments deposited at selected locations of the floodplains of Dharala and Dudkumar rivers at Kurigram, due to breach of embankments during the 1998 flood.


S. J. Md. Yasin, Md. Hossain Ali, E. Hoque, Sadik Ahmed and Rezaul Hoq


EMBANKMENT FAILURE AND CONSEQUENT SEDIMENTATION Bangladesh is a riverine country, which is criss-crossed by numerous rivers. The landmass of Bangladesh was created from the sediments carried by the major rivers. Most of the rivers have not yet attained stability from morphological point of view and carries huge amount of sediments to the sea. These sediments originate either from materials washed from the upstream or from the breaking of the riverbanks due to scouring along the river channel. Islam (1994) has shown that breaches in river embankment result in huge amount of sediment deposition over the natural levees. It was reported that during 1985-1991 ten breaches in eight embankments resulted in more than 16 million m3 of sediment deposits over the natural levees causing damage to agricultural crops of 66,700 hectres of land (Islam, 1994). A number of studies have been carried out on floodplain sedimentation due to flood inundation (e.g., Muramoto et al., 1989; Islam, 1994; Islam and Salehin, 1997). In fact, failure of river flood embankment causes greater volume of damage because of the increased economic activities in the protected area after the construction of the embankment (Chowdhury, 1996). Breach of embankment causes harm to agricultural land by depositing coarse sand whereas natural flooding of floodplains by riverbank overflow improves the fertility of land due to deposition of silt and clay. BACKGROUND OF THE PRESENT STUDY Since failures of flood embankments are usually accompanied by huge damages, it is necessary to make risk assessment in such a project. For this purpose numerical models are currently being used. The numerical models are based on information on the mechanism of floodplain sedimentation and physics of the phenomena. Towards this goal, “physical model laboratory tests” are necessary. Islam and Salehin (1998) reported results of few such tests. Grain size of soils used in such tests is an important parameter. In this respect, the findings of this study will help as a guideline in calibrating future experimental and numerical models and for using them in the actual analysis. SCOPE OF THE PRESENT STUDY The Kurigram Irrigation and Flood Project area is one of the severely affected regions of the country during the 1998 flood. The project area is bounded by two rivers, the Dudkumar river and the Dharala river. Flood control embankments, built on the right bank of Dudkumar River and on the left bank of Dharala river encompasses this area. The Dudkumar right embankment is 41.34 km long

Grain Size Distribution of Sediments

starting from Pateshswari to proposed pumping station at Begomgonj. The Dharala left embankment is 42.30 km long starting from the Bangladesh-India border to the proposed pumping station at Begomgonj. During the flood of 1998 the Dudkumar right embankment failed at 3 (three) locations. Near Char Berubari the river eroded and engulfed about one km of the embankment. The Dharala left embankment also failed at one location; about 1.3 km of the embankment was washed away from Sitaijhar in Panchgachi union to Krishnapur in Mogalbasha union. At this location the Dharala river shifted its course meandering into the project area. Figures 1 and 2 show schematic diagram of the breached embankment and riverbank erosion near Char Berubari of Dudkumar river and from Sitaijhar to Krishnapur along Dharala river, respectively. Sedimentation occurred over huge area adjacent to these breached sections of embankment and caused immediate and long term damage to plant cultivation. This study is aimed at determining the grain size distribution of sediments from these areas.

Figure 1: Schematic diagram showing breach of Dudkumar right embankment during the 1998 flood




Figure 2: Schematic diagram showing the breach of Dharala left embankment during the 1998 flood

COLLECTION OF SOIL SAMPLES AND LABORATORY TESTS

Just after the recession of floodwater, Mr. Sadek and Mr. Rezaul Hoq who worked as Research Associates in this project, visited the damaged sections of the Dudkumar and Dharala embankments and collected soil samples from the area. A total of 40 (forty) samples were collected from the area of which 9 (nine) were from the floodplains near failed section of Dudkumar embankment at Berubari and 31 (thirty one) from the floodplain near failed section of Dharala embankment at Panchgachia Union. These included 3 (three) samples from the original soil near the breach of Dharala embankment and 4 (four) samples from the original soil near the breach of Dudkumar embankment. The rest were from the recent deposits due to flood. These samples were collected from locations selected at random along the riverbank. No attempt was made to measure the aerial expansion, distribution or pattern of the sediments deposited.


Mechanical sieving and hydrometer analysis were performed on the collected soil samples to determine their gradation. US standard sieves #4, #8, #16, #30, #40, #50, #80, #100 and #200 and a power driven sieve shaker were used for mechanical sieving. Hydrometer analysis was performed using a 151H ASTM hydrometer. ASTM-D422-63 test procedure was followed. Sodium Hexameta Phosphate was used as dispersing agent and necessary corrections, e.g., temperature, dispersing agent, meniscus and immersion corrections were applied in the calculations. Specific gravity determination was carried out for each sample that required hydrometer analysis and the respective specific gravity was used in the calculation of grain diameter and percent finer.

RESULTS AND DISCUSSION

The specific gravity values of the collected soils are summarized in Table 1. The average specific gravity of the sediments of Dharala river are 2.703 and that of the Dudkumar river sediments are 2.766. As seen from Table 1, there is no significant variation in the specific gravity of samples collected from different aerial locations near the breached sections of embankments. Table 1: Variation in specific gravity of the collected soil samples

Dharala Dudkumar Original soil

from the bank Soil deposited in 1998 flood

Original soil from the bank

Soil deposited in 1998 flood

No. of samples* 3 21 2 6

Maximum 2.739 2.769 2.695 3.021

Minimum 2.680 2.652 2.624 2.709

Average 2.714 2.703 2.660 2.766 Standard deviation 0.031 0.032 0.050 0.125

* Specific gravity determinations were done for samples that needed hydrometer analysis.

Figure 3 compares the grain size distribution of samples collected from near the breached sections on the left bank of Dharala River. A wide range of grain sizes was found to be deposited at different locations within a small area near the failed sections of the embankments. The average curve is also shown on the same figure. The average curve was obtained by taking mean of the percentage finer values corresponding to any diameter for all the samples. Figure 4 shows similar curves for samples from Dudkumar riverbank. However, the gradation of samples from Dudkumar riverbank is less varied compared to those from the Dharala riverbank.




1 0.1 0.01 1E-30

20

40

60

80

100

Average

Coarse Medium Fine Coarse Medium

MIT soil classificationSand Silt

Perc

enta

ge fi

ner

Grain size, mm

Figure 3: Grain size distribution of soil samples collected from deposits

along Dharala riverbank

1 0.1 0.01 1E-30

20

40

60

80

100

Average



Perc

enta

ge fi

ner

Grain size, mm

Figure 4: Grain size distribution of soil samples collected from deposits along Dudkumar riverbank


The gradation curves for samples from original soil from the banks are also compared with those from the flood deposits in Figs.5 and 6 for samples from Dharala and Dudkumar riverbanks, respectively. Figure 5 shows that the 1998 flood deposits were relatively coarser than the Dharala bank soil and relatively finer for the Dudkumar riverbank soil. These figures suggest that the deposited sediments originated not from the erosion of the bank near the breach but possibly from upstream locations.

In Figs. 7 and 8, selected diameters d10, d30, d50 and d60 obtained from individual grain size distribution curves are plotted to show their range of variation. In these figures horizontal scale do not represent any quantity. Here d10 means the diameter such that ten percent of the grains are finer than that size and so on. The d10, d30, d50 and d60 values at different locations are spaced apart along the horizontal axis for clarity. Table 2 shows the mean and standard deviation of these diameters for all samples. Table 2: Variation in grain diameters of the samples collected from Dharala

and Dudkumar riverbanks.

Dharala Dudkumar Diameter

Mean Standard deviation

Values from avg.

curve

Mean Standard deviation

Values from avg.

curve d10, mm 0.047 0.0314 0.020 -- -- --

d30, mm 0.084 0.0366 0.073 0.004 2.957x10-4 0.004

d50, mm 0.113 0.04156 0.106 0.007 4.588x10-4 0.009

d60, mm 0.128 0.04481 0.125 0.009 6.276x10-4 0.012

CONCLUSIONS Soil samples collected from different aerial locations near the breached sections of Dharala and Dudkumar river embankments during the 1998 flood showed that there are considerable variations in grain size distributions of the sediments deposited at different location on the natural levees. Also the grain size distribution of the soil samples collected from just beneath the 1998 deposition, which probably resulted from depositions prior to embankment construction, were found to be very much different from the recent deposits. So, it is suggested that laboratory model studies on flood plane sedimentation and calibration of relevant models be carried out taking into consideration this wide range of distributions of grain sizes.




1 0.1 0.01 1E-30

20

40

60

80

100

Average (All samples)

(1998 deposit)Average

Average (Original soil)



1998 floodSoil deposited inOriginal soil

Perc

enta

ge fi

ner

Grain size, mm

Figure 5: Comparison of grain size distribution of deposited soil due to

embankment failure with original soil along Dharala riverbank

1 0.1 0.01 1E-30

20

40

60

80

100

(1998 deposit)Average

Average (original soil)



Average (all)

Sediment deposited during 1998 floodOriginal soil

Perc

enta

ge fi

ner

Grain size, mm

Figure 6: Comparison of grain size distribution of deposited soil due to

embankment failure with original soil along Dudkumar riverbank


-0.01

0.00

0.01

0.02

0.03

0.04

0.05

samples are spaced alongd10, d30, d50 and d60 for different

Ave

rage

(All)

Ave

rage

(orig

inal

soil)

Ave

rage

(dep

osit)

represent any quantity.Note : Horizontal scale does not d10

d30 d50 d60

Dia

met

er (d

10, d

30, d

50, d

60),

mm

Figure 7: Variation of d10, d30, d50 and d60 of different sediment samples deposited along Dharala riverbank

0.00

0.04

0.08

0.12

0.16

0.20

0.24

the horizontal scale.

Ave

rage

(All)

Ave

rage

(orig

inal

soil)

Ave

rage

(dep

osit)

Orig

inal

soil

samples are spaced alongd10, d30, d50 and d60 for differentrepresent any quantity.

Note : Horizontal scale does not d10 d30 d50 d60

Dia

met

er (d

10, d

30, d

50, d

60),

mm

Figure 8: Variation of d10, d30, d50 and d60 of different sediment samples deposited along Dudkumar riverbank




REFERENCES Chowdhury, J. U. (1996) Flood control activity in wetlands: An environmental

concern, Tri-annual journal ‘Earth’, No.05, Dhaka. Islam, M. Z. (1994) River embankment failure resulting in sedimentation over

the floodplain in Bangladesh and their model experiments, Disaster prevention research institute, Kyoto University, Kyoto, Japan, pp1-76.

Islam, M. Z., and Salehin, M. (1997) Embankment failure resulting in sedimentation over the floodplain: physical model experiments, Technical report 01, RO2/97, Institute of Flood Control and Drainage Research, Bangladesh University of Engineering and Technology, Dhaka.

Islam, M. Z. Salehin, M. (1998) “River Embankment failure resulting in

sedimentation over the floodplain: physical model experiments, Technical report 02, R042/98, Institute of Flood Control and Drainage Research, Bangladesh University of Engineering and Technology, Dhaka.

Muramoto, Y., Fujita, Y. and Okubo, K. (1989) Floodplain erosion and sedimentation due to overbank flood flow, Proceedings of the Japan-China (Taipei) joint seminar on natural hazard mitigation, Kyoto, pp. 429-438.

ASTM-D422-63, Standard method for Particle size analysis of soils.


Assessment of Flood Damage to Roads in and around Dhaka City and Remedial Measures

Moazzem Hossain, Alamgir M. Hoque and M. Zakaria

Department of Civil Engineering Bangladesh University of Engineering and Technology, Dhaka-1000, Bangladesh Abstract The devastating flood of 1998 has caused considerable damages to the road pavements in and around Dhaka City. After the recession of floodwater, a major investment is usually necessary for rehabilitating the road network. An assessment of the volume and extent of damage is needed for framing the corresponding fiscal policy. In this study, assessment of the damages to the roads in and around Dhaka city has been made and investigations into the road material characteristics have carried out. A field survey has been undertaken which includes manual measurement, field interview and photographic evidence. Material samples have also been collected from the damaged road sites. Photographs have been analyzed in the studio and materials have been tested for important engineering characteristics. From the study, it has been found that flow of large vehicles during flooding period has a significant influence on the extent of damage. Also, the pavements that have been in good condition before the flooding have experienced less or almost no damage. It has also been found that for a number of pavements, design parameters and proper specifications have not been followed. INTRODUCTION

The devastating 1998 Flood inundated a significant portion of road network in and around Dhaka City. It has generally been observed that considerable damages have been incurred in all the inundated roads. Right after the recession of the floodwater, a major investment would be necessary for rehabilitating the




road network. An assessment of the volume of damage is urgently needed for framing the corresponding fiscal policy. It is again necessary to investigate the reasons for such wide scale damage to the road pavements. In the long-term basis, it is also important to investigate whether any changes in the pavement design criteria and material specifications should be made in view of the frequent inundation of roads by flood.

The major objectives of the present research study were: (i) To assess the flood damage to roadway pavements; (ii) To record the inundation characteristics of major roadway corridors of Dhaka City; (iii) To measure characteristics of in-situ road materials, and to compare the characteristics with those of the standard specifications; and (iv) To suggest remedial measures for future construction and design processes. METHODOLOGY For this study, major roads of Dhaka City were identified. Physical visits to those roads were made and questionnaire surveys were carried out at the road sites to record the inundation conditions. Right after the recession of the flood, photographs of selected sections of each major road were taken for assessing the scale of damage. Simultaneously, material specimens were collected from each section of the road. The photographs were analyzed in the studio and the materials were tested in BUET laboratories. The extent of damage was assessed by analyzing the photographs. Efforts were made to identify the reasons of wide scale damages to roads by the flood. Also, results of material tests were compared with the standard specifications to see whether those were followed during construction. FIELD SURVEY Field survey is an effective way of gathering information regarding flood damage to roads. When this research work started, the floodwater from Dhaka City roads had already receded. Therefore, alongside photography and visual observation interview of roadside residents and retailers were taken regarding the inundation characteristics of the roads (Hossain et al., 1999). In-situ materials from road surface layers were also collected for laboratory tests. Twelve major roads/corridors of Dhaka City and its suburbs were identified for carrying out the filed survey. These were Shantinagar Road, Khilgaon-Taltala Road, Shahid Suhrawardy Avenue, Pragati Sarani, Dhaka-Aricha Highway, Sayadabad–Jatrabari Road, Demra Road, Basabo Corridor, Mirpur Darussalam Road,

Assessment of Flood Damage to Roads

Satmasjid Road, Jatrabari-Postogola Road and Dhaka–Mawa Road. Apart from visual observation, photographs and interview, road materials samples were also collected from surface as well as base or subbase layers, where surface layer was washed out. Visual Observation After arriving at the road site, general observation of the road surface was made. As the floodwater had already receded by that time, efforts were made to find the mark of the maximum flood water level on the roadside structures i.e., curb, footpath, electric/telephone pole, walls, etc. Where found, the approximate floodwater height from the road surface was measured with scale/tape. Also, the general condition of the pavement was observed and recorded. The maximum pond depth created by flood damage was also measured in each 100m section of the road. Interviews Roadside residents and retailers are the best witnesses of the flood and hence the best source of information regarding the duration and extent of flooding. After confirming their knowledge the road under consideration, such people were asked about the flood duration in days. If at least three estimates of the flood duration from three different persons matched, then it was recorded as the flood duration for that portion of the road. Photographs Photographs can be used as the hard evidence of flood damage. In the present study, photographs of the flood-damaged roads were taken for use in assessing the damage. One snap is taken in each 100m portion of the road. In order to be able to determine the approximate scale of the photo, two bright markers (bright colored cloth in this case) were included in each photo. Sample collection The final task of field survey at each site was to collect material sample from roads. From each road site more than one sample consisting of approximately 5-7 kg of asphalt concrete from the wearing course were collected. In cases where wearing coarse was fully washed away, material samples from base/sub-base coarse were collected. A mild steel pavement cutter along with a heavy-duty hammer was used for this purpose.




LABORATORY TESTS Laboratory tests were performed on the samples collected from the road site. The objective of the laboratory tests was to know the field material characteristics relevant to the design for the asphalt concrete pavement. The main material characteristics used in the design process were: percentage of bitumen, aggregate types, grading and strength. Asphalt content of the wearing course was found by extracting the asphalt with tri-chloroethylyne in accordance with ASTM 2172 (ASTM, 1992). The aggregate type (whether stone chips or shingles) was noted by observing the asphalt free aggregates. Then sieve analysis was performed on the aggregate samples to know the gradation. Aggregate Impact Value (AIV) test was performed (BS, 1985) to determine the strength of the aggregate used in base/sub-base coarse. ANALYSIS OF PHOTOGRAPHS In order to be able to observe the photographs in a magnified scale, slides were developed. All the required linear measurements regarding the extent of damage were taken from the projected views of the slides. Thus, it was possible to minimize the magnitude or error in distance measurements. Most of the photographs represented little over one thousand square foot of pavement surface area. Therefore, the extent of pavement damage was estimated as a percentage of damaged pavements per one thousand sq.ft. of pavement area. A typical damage assessment and flooding characteristics for Pragati Sarani Road has been presented in Table 1. The same in summary form for all other roads under the present study has been presented in Table 2. From Table 1 it can be observed that even a low height flooding (3 inch to 6 inches) for duration of five to ten days caused damages upto 95% of the pavement area in case of Progati Sarani. The damage assessment listed in Table 2 also reveals that flood duration and flooding height have no definite correlation with the extent and severity of damage. Again, from Table 2 it can be observed that DIT road to Khilgaon Chowdhurypaara road section was damaged only about twenty five percent with no severe pond/ditch type of damage, although flood duration here was twenty days with a maximum flood height of one foot (30.48 cm). For similar types of flood duration and flooding height, the road section from Chowdhurypaara to Taltala market have not experienced any visible damage. However, it has been reported by the roadside residents that the road is resurfaced just before the flooding season. Similarly, Dhaka-Mawa section of the Dhaka-Khulna highway incurred insignificant damage although flooding duration was fifteen days with a maximum height of two to three feet (61 cm). From field interviews, it has been


revealed that the road was in good condition before flooding and was resurfaced only recently. Table 1: Assessment of flood-damage and flooding condition of Progati Sarani Section

No. Location Damage condition Flooding information

1 In front of U.S. Embassy

• Severe damage • Up to 5.0 inches deep

ditches • 20 % in 1000 sq. ft • Uniform crack

• Water logging • Flood height: 6.0 inches • Duration 10 days

2 Progati Smarani (Shahjadpur)

• Ditch like a pond • Up to 15.0 inches deep • 80 % in 1000 sq. ft


3 400 ft south from Progati Smarani

• 10 inches deep ditches • Great damage • Uniform interval • Large extend of ditches • 85 % in 1000 sq. ft


4 North Badda (in front of Hossain Market)

• Great damage • 6.0 inches deep ditches • Crack • 80 % in 1000 sq. ft


5 Middle Badda • 15 inches deep ditches • Pond type ditch in large

extent • 95 % in 1000 sq. ft


6 Middle Badda (Bus Stand)

• 15 inches deep ditches • Pond type ditch in large

extent • 95 % in 1000 sq. ft


7 Middle Badda (Bus Stand)

• Pond in road ! • 15.0 inches deep ditch in

1000 sq. ft area • 95 % in 1000 sq. ft


8 Middle Badda • Pond in road • 15.0 inches deep ditch in

1000 sq. ft area • 95 % in 1000 sq. ft


9 Merul Badda • 15 inches deep ditches • Very bad condition • 85 % damage in 1000 sq. ft


10 North of Rampura Bridge

• 15 inches deep ditches • Very bad condition • 85 % damage in 1000 sq. ft





Table 2: Summary of assessment of flood damage to roads

The critical loading condition on the pavement is resulted from heavy-duty

trucks and large buses (TRL, 1993; BRRL 1987). Therefore, it is expected that during flooding the flow of these large vehicles would have multiplying damaging effect on the pavement structure. From a recent study (Hossain et al., 1999a) it has been revealed that among the national highways Dhaka-Mawa corridor of Dhaka-Khulna highway carries lowest percentage of large vehicles.

Road/Corridor

% Pavement area damaged per 1000 sq. ft.

Depth of ditch/ pond,

if any

Inundation characteristics

Shantinagar Road 25-50 4"-6" Flood water high 1'-1.5' flood duration 25-40 days

DIT Road to Khilgaon Chowdhury Para

25-30 - Flood water height = 1 ft Flood duration 20 days

Chowdhury para to Taltala Market

No damage, Good condition, Road repaired just before flood

- Flood water height = 1 ft Flood duration 20 days

Shahid Suhrawardy Avenue

10-20 3"-5" Flooding height = 3"-7" Flood duration = 10-20 days

Pragati Sarani The pavement is almost fully damaged

6"-15" Flooding height = 3"-4" Flood duration = 5-10 days

Dhaka – Aricha Highways

Technical to Amin Bazar portion is fully damaged

6"-8" Flooding height = 0.5'-1.0' Flood duration = 10-20 days

Amin Bazar to Hemayetput Portion

30-60 2"-3" Flood height 1'-4' Flood duration 15-20 days

Sayadabad to Jatrabari

30-70 4"-8" Flood height = 3"-4" Flood duration = 5-15 days

Basabo corridor 20-40 4"-6" Flood height = 1'-3' Flood duration 20-25 days

Mirpur Darussalam Road

40-90 3"-4" Flood height = 1 ft Flood duration = 15-20 days

Satmasjid Road Mohammadpur Bus Stand to Physical College Portion

40 2" Flood height = 2"-3" Flood duration = 5 days

Jatrabari – Postogola 50-100 4"-6" Flood height = 1'-3' Flood duration = 15 days

Dhaka – Mawa Insignificant damage

Flood height = 2'-3' Flood duration = 15 days


While the percentage of heavy vehicles on this road is in the range of 5%-6%, the same on the other national highways varies in the range of 12%-25% with the maximum percentage of trucks on Dhaka-Aricha highway. Also, within the city area the less damaged (Table 2) roads like DIT road to Chowdhurypara and Chowdhurypara to Taltala do not include the three major truck routes as identified by a recent study (Chowdhury et al., 1991). So, it appears that flow of large vehicles during flooding period has a significant influence on the extent of damage. Generally, the flooded road with higher percentage of truck traffic experienced greater damage.

ANALYSIS OF RESULTS Material characteristics of Dhaka–Aricha highway have been investigated through laboratory tests. Tests have been performed on four different pavement samples collected from four different sections of the highway. The percentage of bitumen content found in the wearing course mix is given in Table 3. While optimum bitumen content for most of the asphalt concrete mixes lies between 5.5% to 6.5% (RHD, 1994), from Table 3 it can be observed that three out of four samples contained either very low or very high percentage of bitumen. This lower or higher amount of bitumen content might be a major reason for lower stability of the mixes. Table 3: Percentage bitumen content in the paving mixture sample

Road Sample No.

Sample layer Inundation period (Days)

% bitumen content

1 Surface layer 5 9.7 2 Surface layer 20 8.2 3 Surface layer 10 5.4

Dhaka - Aricha highway (Gabtali-Aminbazar section)

4 Surface layer 15 3.7

From bitumen extraction it has been found that two out of four aggregates samples are shingles and rest two are stone chips. The gradation curves of all the four aggregate samples along with the suggested (RHD, 1994) gradation range have been plotted in Figure 1. It is suggested (RHD, 1994) that aggregates used should consist of chipping with a maximum size 8 to 14 mm. Thus, from Fig. 1 it can be observed that none of the samples, especially the coarser portion of the aggregate, are within the specified range. The results of AIV tests on base/sub-base aggregates are presented in Table 4. Aggregate gradation has a significant influence on the overall stability and strength of the asphalt concrete mix. From




Table 4 it can be observed that AIV values for three out of four samples are significantly higher than the specified maximum value of 25 (TRL, 1993).

Figure 1: Comparison of aggregate sample gradation with the suggested (RHD, 1994) gradation range

Table 4: Aggregates impact values of aggregates samples collected from base/subbase layer

CONCLUSIONS Assessment of flood damage reveals that even a low height flooding (3 inch to 6 inches) for duration of five to ten days caused wide scale damages to pavements. It proves that flood duration and flooding height have no definite correlation with the extent and severity of damage. However, it has been revealed that flow of

Gradation Curve

0

20

40

60

80

100

0.01 0.1 1 10 100Sieve size (mm)

Perc

ent f

iner

Sample 1

Sample 2

Sample 3

Sample 4

Upper Sp.

Lower sp.

Road Sample No

Sample layer Inundation period (Days)

AIV value

1 Base/Subbase 20 34 2 Base/Subbase 5 23 3 Base/Subbase 10 34

Dhaka – Aricha highway (Gabtali-Aminbazar section)

4 Base/Subbase 15 32


large vehicles during flooding period has a significant influence on the extent of damage. Generally, the flooded road with higher percentage of truck traffic experienced greater damage. Also, the pavements that were in good shape before the flooding experienced less or almost no damage.

While optimum bitumen content for most of the asphalt concrete mixes should be between 5.5% and 6.5%, it has been observed that most of the wearing course samples contained either too low or too high percentage of bitumen. This lower or higher amount of bitumen content might be a major reason for lower stability of the mixes. It has also been revealed that in selecting the gradation and type of coarse aggregate used in the wearing course, proper specifications have not been followed. REFERENCES ASTM (1992) American Standards for Testing Materials, Vol. 4.03, Road and

Paving Materials Part-I, USA BRRL (1987) A guide to the design and construction of bitumen-surfaced roads

in Bangladesh, Roads and Highways Department, GOB. BS (1985), British Standard 812, Part 3. Chowdhury, K. A., Karim, S. R., Masudunnabi, M. and Fazal, M. A. (1991) A

study of freight movement in metropolitan Dhaka, B.Sc. Engg. Thesis, Dept. of Civil Engg., BUET.

Hossain, M,, Hoque, A. M. and Zakaria, M. (1999) Assessment of flood damage to the roads in and around Dhaka city and remedial measures, Final report, DAERS, BUET, Dhaka.

Hossain, M., Rahman, M. A. and Iqbal, G. A. (1999a) Vehicular speed study on two-lane two-way national highways of Bangladesh, Proceedings of Civil & Environmental Engineering conference, 1999, Bangkok, Thailand.

RHD (1994), Road materials and standards study of Bangladesh, Roads and Highways Department, Ministry of Communications, GOB, Dhaka, Bangladesh.

TRL (1993), Road Note 31: A guide to the structural design of bitumen surfaced roads in tropical and subtropical countries, London, UK.



Assessment of Economic Loss Caused by Flood Damaged Transportation Network

M. J. B. Alam, Alamgir M. Hoque and Md. Mazharul Hoque

Department of Civil Engineering Bangladesh University of Engineering and Technology, Dhaka-1000, Bangladesh Abstract Flood causes direct as well as indirect loss to the economy of a country. The direct losses include damage to the infrastructure, agricultural and industrial products, human being and livestock. The indirect losses include the damage to the economic activities that cannot remain operative due to lack of communication facilities during and immediately after flood. The indirect losses, particularly losses due to damage of transportation facilities may be substantial and warrant special attention. This paper presents an estimate of the indirect loss caused by flood damaged road network. The paper reveals that such indirect loss is much higher than the direct loss. It suggests that some economically important transportation corridors should be constructed in such a way that they remain workable even during flood. INTRODUCTION Transportation provides the essential linkages for economic activities of a country. With the increase of specialization of different types of industries in different areas, transportation acts as arteries and veins for the operation of the whole system. Damage to transportation system makes a country lame and hampers economic development severely.

Natural disasters like flood can affect economy of a country in two ways. The first one is the direct effect where it destroys national assets like human



lives, crops, livestock, industries, houses and infrastructures. The other is the indirect effect where the economic activities, which are not destroyed by the disaster, cannot remain operative due to the lack of transportation infrastructure. The monetary value of the first type of damage can be perceived and analyzed relatively easily. For this reason it is usually included in the estimation of the economic losses and in the formulation of rehabilitation policies. On the contrary, it is very difficult to estimate the damages caused by the second type of effect mentioned above. But the economic value of the second type of damage may also be important. This type of loss can be avoided provided the major transportation linkages are protected during the disaster.

Flood is a perennial problem for Bangladesh. If the indirect effect of flood, as mentioned above, proves to be substantial, it may influence future investment policies in transportation significantly. In this case the major transportation linkages should be built in a way that they are not affected by flood. These transportation corridors should interconnect the major industrial areas of the country.

The complete structural measure to prevent flood from occurring is not economically feasible for Bangladesh. By protecting the areas of higher economic potential, e.g., major industrial areas, and by maintaining transportation facilities among these areas and ports, economic losses may be significantly alleviated. This study deals with the economic loss caused by unavailability of transportation facility due to flood-damaged road network. Here the monetary value of this kind of damage is estimated for the devastating flood of 1998. FLOOD OF 1998 AND ITS EFFECT ON TRANSPORTATION SYSTEM Flood is a recurring problem in Bangladesh. About 60 percent of the country is flood prone while 25 percent is inundated during monsoon in normal years (Hasan, 1998). The flood of 1998 is one of the most devastating in the living memory. The flood started in early July and remained until late September. Due to the flood, transportation infrastructure in general and highways, railways, bridges and culverts in particular have been severely damaged.

The flood of 1998 has affected most of the road and railway infrastructures of the country. A preliminary assessment of the damages made by the Ministry of Communications and Roads and Highways Department (RHD) is presented in Table 1.


Loss Caused by Flood-Damaged Transportation Network


Table 1: Assessment of Damages to Highways and Railways Caused by Flood of 1998

Damaged Road (km)

Damaged Bridges &

Culverts (No)

Rehabilitation Cost ( mil.

Taka)

RHD Roads

Length Sub-

merged(Km) Emban-

kmentPave-ment

Badly Partial

No. of Dama-

ged Ferry Ghats

Imme. Repair Cost

( mil. Taka)

Short Term

Long Term

Total Cost (mil. Taka

National Highway

1381.2 623.6 599.3 88 232 13 678.4 490.9 918.1 2087.8

Regional Highway

783.9 329.8 329.2 59 126 7 374.3 268.4 502.0 1144.8

Feeder Road

7457.5 3376.1 3315.9 229 470 34 3761.9 2720.5 5087.6 11570.0

Mec.l Equipment

- 30.0 70.0 0 100.0

Total 9622.6 4329.5 4244.4 376 828 54 4845.1 3549.9 6507.7 14902.6 Railway Tracks and Signaling Bridges Total Cost Railway 496 km 117 1524.9 Source: Roads and Highways Department.

A survey conducted by Bangladesh Garments Manufacturers and Exporters Association (BGMEA) reports that about 250 garments factories of the country were submerged during the flood, over 300 thousand workers were affected and failed to attend to work and as the major highways were snapped by flood transportation cost went up five times. The total amount of production loss in country’s top foreign exchange earning sector amounted to US$173 million (Quddus, 1998).

ASSESSMENT OF SECONDARY DAMAGES CAUSED BY FLOOD Due to the flood of 1998 about 68 percent of areas of Bangladesh had been inundated (Annual Flood Report, 1998). Although most of the inundated areas were rural agricultural lands, many urban areas were also flooded. But among the industrial areas of the country only 25 percent had gone under water. The flood did not directly affect many industries in Dhaka, Khulna, Chittagong and Bogura. But a lot of industries in these areas could not be kept operative due to the fact that either the raw materials or the finished products could not be transported to and from these industries. This was particularly true for the garments industries, one of the most important industries of the country. During flood it was reported that some major garments industries of the country used airplanes to transport their goods to the Chittagong port. Although this increased the production cost quite significantly, the manufacturers opted for it because of lack of any other alternative. The flood in 1998 continued for more than three months. Had the air transport alternative not been used, all the produced goods would have been


wasted and the future opportunities would have been lost. Many small industries were bankrupt due to this reason.

This study is aimed at assessing the monetary value of the secondary type of damage. The analysis for all kinds of industries would have involved large amount of time and resources. As an initial step in this regard, the present study is concentrated on the garments industry considering its importance. For the purpose of the assessment, a survey was conducted among garments manufacturers after the flood of 1998. The main objective of the questionnaire survey was to quantify the significance of unavailability of transportation facilities, due to damage of transportation system caused by flood. It was designed to extract information about the characteristics of the damages caused by flood and their relative importance. A total of 39 randomly selected garments industries around Dhaka were surveyed. The questionnaire included questions on the types of damages caused by flood, the level of production before and during flood and the losses attributed to various factors, including those due to unavailability of transportation.

In the analysis it was observed that about 97 percent of the garments industries were affected by flood (see Fig. 1). Among these industries, only ten percent were submerged and discontinued production. These industries remained closed during the flood. About 87 percent of the garments industries were partially affected by the flood. Although these industries remained operative during flood, their production had to be reduced. Productions were hampered by three major factors as shown in Fig. 2. In 68 percent of cases, transportation of either raw materials or finished products and in 27 percent of cases, transportation of the workers was hampered. So, it is evident that for the partially operative garments industries unavailability of transportation facilities were the major issue and prime cause of losses. The duration for which the industries remained affected by flood varied from 20 to 60 days with an average of 30 days as shown in Fig. 3.

For all the industries that remained partially operative, the production capabilities reduced by about 55 percent on an average, as shown in Fig. 4. For the 39 garments industries surveyed in the study, the total loss amounted to about Taka 75 million. As shown in Fig. 5a, transportation related issues were directly responsible for 26 percent of the loss, which amounts to Taka 19.5 million. About 49 percent of the losses were attributed to reduced production, 15 percent were attributed to additional transportation cost and 11 percent were attributed to the fact that produced goods could not be sent to the market or ports. The rest (i.e., 25 percent) of the total loss were attributed to the fact that the flood caused the garments manufacturers to fail in satisfying the commitments and engaging in new contracts. If this kind of loss is excluded, damage caused by inadequate transportation facilities will increase to 34 percent (Fig. 5b).




Extent of Damage to the Garments Industries Caused by Flood of 1998

10%

87%

3%

Submerged and stopped production

Partially affected but continuedproduction at reduced level Not affected at all

Figure 1: Extent of Damages to Garments Industries Caused by the Flood

Types of Damages Caused by Flood

5%

39%

27%

29%

Production stopped due to innundationby floodRaw materials could not be imported

Produced Goods could not be marketedor exportedWorkers could not come

Figure 2: Types of Damages Caused by the 1998 Flood


Duration of Flood Related Problems

16%

28%11%

21%

24%

20 or Less Days

20-30 Days

30-40 Days

40-50 Days

Above 50 Days

Figure 3: Duration of the Flood-related Problems

Effect of Flood on the Production of Garment Industries

0

10000

20000

30000

40000

50000

60000

70000

80000

90000

Num

ber o

f Pie

ces

Prod

uced

(Per

Mon

th)

Production Before FloodProduction During Flood

Figure 4: Effect of the 1998 Flood on the Production of Garments Industries




Sources of Loss Caused by Flood

(As the Percentage of the Total Loss)

15

49

11

25Loss due to aditional transportation costcaused by flood

Loss due to reduction or stoppage ofproduction

Loss due to the fact that the producedgoods could not be Marketed or Exported

Loss due to long term contract reduction

Figure 5a: Proportions of Losses Attributed to Different Factors

Sources of Loss Caused by Flood(As the Percentage of the Loss Excluding Contact Reduction)

1915

66

Loss due to aditional transportation costcaused by flood

Loss due to reduction or stoppage ofproduction

Loss due to the fact that the producedgoods could not be Marketed or Exported

Figure 5b: Proportions of Losses Attributed to Different Factors


As mentioned earlier, unavailability of transportation for the materials was

the principal reason for reduced production. So, a portion of the losses attributed to reduced production can be indirectly related with transportation. In a nutshell it can be concluded that inadequate transportation is responsible for about 67 percent (half of 66 percent attributed to production loss plus 34 percent mentioned above) of the total losses caused by flood. The production loss caused mentioned above) of the total losses caused by flood. The production loss caused by the flood of 1998 amounted to US$ 173 million (Taka 865 crore) in the garments sector. So the amount of loss that can be attributed to transportation is Taka 878 crore (865*0.67/0.66). Had transportation facilities been maintained, this large sum of money could have been saved from garments sector alone. The enormity of the aggregate indirect loss caused by damaged transportation facility during flood can be imagined from this figure. CONCLUSIONS Flood is a perennial problem for Bangladesh. The people of this country must live with flood. The major objective of the study is to investigate the significance of transportation in the economic loss caused by flood related damages. In the study, it was observed that the 1998 flood did not directly affect most of the garments industries. Rather the industries suffered from inadequate transportation facilities caused by damaged transportation infrastructure due to flood. Factors related to the transportation of either raw materials or workers were responsible for loss of production in 95 percent of cases. Transportation was directly responsible for 26 percent and indirectly responsible for 67 percent of the secondary type of monetary losses caused by flood. Such losses may be avoided by building some major transportation corridors to withstand most severe flood so that connections among the industrial zones, sources of raw materials and ports can be maintained during that time. Attempts should be made to keep the industries of the country operative even during flood. This study dealt only with garments industries of the country. It should be extended to include other products and to reformulate national policies in this regard. ACKNOWLEDGEMENT The authors acknowledge the contributions of Mr. M. F. K. Pasha and Mr. M. A. Uddin as research assistants in the project.




REFEREMCES Hassan, K. M. (1998), Flood 1998: The Condition of the Highways in

Bangladesh, Paper Presented in the Seminar on “Impact of the 1998 Flood on the Economy of Bangladesh.”, Dhaka, September, 1998.

Quddus, M. G. (1998), Overview of the Impacts of Flood on Garments Sector, Paper Presented in the Seminar on “Impact of the 1998 Flood on the Economy of Bangladesh.”, Dhaka, September, 1998.

Annual Flood Report (1998), Flood Forecasting & Warning Centre, Processing and Flood Forecasting Circle, Bangladesh Water Development Board.


Assessment of Flood Damages to Inland Water

Transport Sector of Bangladesh

Abdur Rahim and M. Reaz H. Khondoker Department of Naval Architecture and Marine Engineering

Bangladesh University of Engineering and Technology, Dhaka-1000, Bangladesh

Abstract This paper contains results of survey carried out to assess the damages caused to the inland water transport sector of Bangladesh during the colossal flood of 1998. The results of the survey show that the flood caused extensive damages to this sector. A large number of accidents involving mechanized country boats took place during the flood. A number of factors including high current of water flow, excessive load on the boats, poor technical standard of the boats, poor maintenance and unskilled operation contributed to the accidents. A large number of installations of the Bangladesh Inland Water Transport Authority (BIWTA) suffered extensive damage due to submergence or the water level rise above the design limit. Many pontoons and jetties were either damaged or dislocated. Siltation of rivers during and immediately after the recession of the floodwaters also caused closure of waterways. This forced the crafts to use alternative routes, which were much longer. Dislocations of ghats have also caused inconvenience to the passengers and disruption in cargo movement, increasing the cost of the transportation. It is recommended that the damaged structures be immediately restored and funds be made available for dredging the entire water transport system. Immediate measures should be taken for regulation of the mechanized country boats so as to ensure their safer operation even in times of natural calamities like flood. INTRODUCTION


Bangladesh, a land of 1,48,393 km2, provides drainage to 1,660,000 km2 of the combined catchment of the Ganges-Brahmaputra-Meghna System i.e., to an area

Abdur Rahim and M. Reaz H. Khondoker


of 11 times the area of Bangladesh. The catchment area comprises the northern slope of the Himalayas in Tibet of China, Northern India, Northeastern India, Bhutan and Bangladesh. Out of the whole basin area only 7.5% is within Bangladesh, while 92.5% of the basin area lies outside the territory of Bangladesh i.e. in India, Nepal, Bhutan and China. Table 1 provides some statistics on river basin and drainage area. Table 1: Statistics on river basin and drainage area

Area of Bangladesh 1,47,570 sq. km. Total rivers 230 nos. Cross boundary rivers 57 nos. Major river basin area (India, Nepal, Bhutan, China and partly Bangladesh)

(a) Brahmaputra basin 5,83,000 sq. km. (b) Ganges basin 9,07,000 sq. km. (c) Meghna basin 65,000 sq. km. (d) Southeastern hill basin 45,000 sq. km.

Total 16,00,000 sq. km. Drainage area

(a) Drainage basin inside country 7.50 % (b) Drainage basin outside country 92.5 %

The Ganges rises from the Gongotri glacier on the southern slope of the

Himalayas at an elevation of over 7,000 m west of Nanda Devi range in Himachal Pradesh and northernmost Uttar Pradesh, west of Nepal. The river comes out of the Himalyan and Siwalik range near Dehradun and enters the plains at Hardwar.

The Brahmaputra rises in the northern slopes of the Himalayas in the Kailash range and flows 1127 km straight to the east parallel to the Himalayan range. In Tibet the river is called Tsanpo. The Tsanpo is a sluggish river in the southeastern part of Lasha where the river is possibly highest navigable river in the world. The river makes a hairpin bend in the eastern edge of the Himalayan range where the Himalayan also makes a right angle bend to form the Arakan – Yoma of Burma. The Tsanpo cuts a number of deep gorges here at Namcha Barwa (7755m) and enters the Assam valley at Sadiya (135 m) in Northeast Assam. The Brahmaputra in Assam is called Dihang. Enriched by a number of large tributaries in the Assam valley the river enters Bangladesh some 12 km upstream of Noonkhawa in Kurigram district.

The Meghna River drains an area of 77,000 km2 of which about 46,500 km2 (60%) lies in Bangladesh. The Barak is the principal headquarters of the Meghna.

Assessment of Damages to Inland Water Transport Sector


The Barak rises at an elevation of 2900 m on the south side of Mount Javpo on the Navaland-Manipur border. The Barak in India has a catchment of 25,265 km2 and on entering in Bangladesh at Amalshid bifurcates into the Kushiyara and Surma. At the point of bifurcation the larger portion of flow enter the Kushiyara while the smaller portion of the flow enters the Surma.

The Surma in its westerly course carries the flow from the Meghalaya and the Kushiyara in its southwesterly course carries the flow from the Tripura hills and meet at Markuli and the combined flow is known as upper Meghna. The Boulai system draining Garo hills in the north Mymensingh confluence with the upper Meghna at Dilalpur not far from Bhairab Bazar.

The major rivers mentioned above and their innumerable tributaries like the Rupsha, the Lakhya, the Dhaleswari, the Bhaguakul, the Pasur, etc. give a good access to most areas of the country when flowing at low level. Moreover, these rivers also give almost complete access to all parts of the country during flood. The geographical features have made Bangladesh one of the most difficult areas of the world with respect to developing a modern surface transport system suitable for guaranteed communication round the year. In almost all parts of the country, the highways and railways require embankment, sometimes as high as 6 meters, so that they are usable during floods. However, sometimes these floods become severe and cause devastating damages to human lives and properties in addition to causing breach and damages to the embankments.

In 1998, Bangladesh experienced one of the most severe flood in its history. The two-month long deluged caused damage to every sector of the national economy. For a riverine country like Bangladesh, inland water transport is vital for the transport system of the country. The water transport system and especially the mechanized country boats have been of great help to the people for movement and distribution of relief goods during the floods. However, soon after the deluge was over, people started to discover the damages caused to the water transport sector by the flood. One of the major reasons for the damage was the massive siltation and sediment transportation in the river system of the country during the flood. The floodwaters is estimated to have carried more than 3 billion tons of silts with it and a large part of the same have settled in the rivers and floodplain. Also, the onrush of water has caused massive movement of sands, silts and clays in the floodplain. This has resulted in loss of navigability in many navigation canals, damage to ports, ghats and other transport infrastructures as well as loss of forestry, which is likely to seriously affect the country boats sector.

The overall objective of the present study was to assess the extent of damage to the water transport sector of the country, with specific reference to the rural waterways mainly used by the mechanized country boats. The specific objectives were: (i) to investigate accidents of marine vehicles during flood, (ii) to assess



damages of the water transport infrastructures, and (iii) to assess disruption in navigation channels due to siltation during and immediately after the flood.

In order to achieve the objectives, extensive field visits were made for data collection. A questionnaire was prepared and distributed amongst the inland water transport operators, traders, passengers and other related persons. The data were collected mainly from Bhairab, Kishoreganj, Pabna, Khulna, Bagerhat, Barisal, Kurigram, Sunamganj, Brahmanbaria, Patuakhali and Pirojpur. Data were also collected from organizations like Bangladesh Inland Water Transport Authority (BIWTA) and Local Government Engineering Department (LGED). FLOOD OF 1998 The 1998 flood has been termed as the most severe flood of the century. The fury of the flood in terms of magnitude and duration, and its devastation and human suffering was unparallel. Starting from 8th July when the Brahmaputra-Jamuna crossed danger level for the first time, the country was in the grip of flood for about 79 days until the Meghna at Bhairab Bazar dropped below danger level on 25 September of 1998.

During the 1998 flood four flood waves in succession passed through the Brahmaputra. These waves arrived before the river level could drop down sufficiently (from the effect of the previous wave) to accommodate for the next wave. These flood waves were recorded at Bahadurabad (IEB, 1999) with respect to danger level as shown in Table 2.

Table 2: Flood waves during the 1998 flood

Peak ID Start date

End date

Peak date

Peak level cm Above DL

1st 9/7/98 19/7/98 15/7/98 19.95 45 2nd 19/7/98 2/8/98 26/7/98 20.05 55 3rd 12/8/98 31/8/98 20/8/98 20.17 67 4th 31/8/98 12/9/98 8/9/98 20.37 87

It may be noticed that the peaks of the four flood waves were increasing and

the river at this point was above danger level for 57 days and very close to danger level for another 9 days (from 3rd to 11th August 1998). What is more important is that the peak flood of the Ganges at Hardinge bridge occurred on 9th and 10th September at a record breaking level of 15.19 and coincided with the Brahmaputra-Jamuna peak at Aricha.



The effect of these flood levels and propagation was obvious. The left bank distributaries of the Brahmaputra from the old Brahmaputra down to Chital, Jhenai, Fatikjani, Lohajang and Dhaleswari flooded the greater districts of Mymensingh, Tangail and Dhaka. Backwater effect originating at the Hurasagar-Jumuna confluence traveled into the Chalan beel depression and flooded the vast area in the greater Pabna, Rajshahi and Bogra districts.

Lower reach of Mohananda was seriously affected from the backwater effect of the Ganges and flooded areas in Chapai-Nawabganj district including district headquarters. Flows through the Gorai and the Arial Khan and right bank distributaries of lower Meghna flooded the Gangetic delta in Bangladesh. Water level hydrograph of Panka, Chapai-Nawabganj (of Mohananda) and Rajshahi testifies this. A quick assessment inferred that global warming, glacier melt, sea level rise or the Tsunamis were not the causes of the 1998 flood in Bangladesh. Very high rainfall in all the three major river basins is the immediate cause of the 1998 flood.

Government of Bangladesh has already published the damages caused by the 1998 flood. Bangladesh water Development Board (BWDB) has also published a flood inundation map showing the flood-affected areas. The summary of losses and damages upto September 30, 1998 caused by flood is shown in Table 3.

Table 3: Summary of losses and damages caused by the 1998 flood (upto September 30, 1998)

Total area affected by flood About 1,00,000 sq. km. Total shortfall in production About 2.2 million MT Number of Districts 52 Number of Police Stations 366 Number of Affected Union Parishad 3,323 Number of Affected People 3,09,16,351 Affected Standing Crops in Acre 14,23,320 Number of Affected Homesteads 9,80,571 Number of Deaths 918 Cattle heads killed 26,564 Road Damaged (km) 15,927 Embankment Damaged (km) 4,528 Number of Damaged Bridge/Culverts 6,890 Number of Educational Institutions 1,718 Number of flood Shelter 27 Number of People taking refuge 10,49,525



The water transports, especially the mechanized country boats, were the only means of transportation in the flood-affected areas of the country. In one way, the floods were a blessing to the inland water transport sector of the country. However, the benefits appeared to be short-lived and were followed by harmful effects soon after the water receded. In this study an attempt has been made to identify how the flood affected certain navigation routes and how siltation forced boatmen to take detour. Such phenomenon caused losses to the vessel operators and traders. The flood also damaged a large number of ghats and pontoons and this resulted in extra difficulties and costs to be borne by the people and traders for using the same. There will be a cost involved for relocating the ghats at an appropriate place. The results of the study will also assist in preparing strategy and program for repair and rehabilitation of the identified damages. ANALYSIS OF DATA The data available from the survey was compiled and analyzed with the objective of estimating the extent of the damages caused to the inland water transport system in the regions investigated. Some of the damages could be easily quantified in monetary terms, while certain types of damages could not be quantified easily. The passengers had to pay directly or indirectly due to damage to the jetties/pontoons or shifting of the same. Evaluation of such damages was very difficult. Within the limited scope of the present works, it has not been possible to quantify such damages. Such damages have only been recorded and presented in the paper. Accidents

Table 4 shows the data and information on accidents with the mechanized country boats that took place in the areas covered by the study. The data and information contained in the table shows that quite a good number of accidents (capsize/damage) of mechanized country boats took place during the flood throughout the country. However, the number of accidents has been reported to be highest at Nabinagar, Brahmanbaria. The total number (i.e., 20) of capsize was reported for Meghna. This is apparently due to the abnormally high level of water in this river during the flood. There was continuous flow of large volume of water for more than two months (Chowdhury and Islam, 1999). This flow was obstructed by the very large spring tides from the opposite direction that resulted in building up of a very large water depth.

The action of the strong current and wave has been identified to be the main cause of the accidents. There were accidents due to stoppage of the engine and collision with launch. Discussions were made with the boat operators regarding



the reasons for such accidents during the flood. The following major reasons were identified: (i) The number of mechanized country boats in Bangladesh is estimated to be

more than 600,000. During the floods, these boats were the main and practically the only mode of transportation. Consequently, the boats were operated beyond their capacities and often were operated by inexperienced crew.

(ii) The hull of the boats is basically meant for non-mechanized propulsion,. Moreover, the quality of the mechanization is very poor with very poor quality components being used. The workmanship is also extremely poor by marine standard (Rahim et al., 1993).

(iii) During the flood, boats were being built hurriedly and as a result of the quality of construction and fitting of the engines as well as the propulsion system were poor, even by the indigenous standard.

(iv) Due to the excessive demand during the flood, maintenance of these boats was poor.

(v) The direct drive and absence of reverse/reduction gearbox made the boats extremely poor in terms of maneuverability and control.

(vi) The currents were generally very strong during the flood, which made maneuvering of the vessels very difficult.

Some data on monetary losses arising out of the accidents has been reported

as presented in Table 4. There was no opportunity to verify the accuracy of the figures. Moreover, the survey could not cover the whole of the country. Thus practically no inference could be drawn from the reported data. However, it can be emphasized that considerable monetary losses were incurred due to the accidents. Siltation of Rivers Causing Loss of Navigability in the Dry Season

The biggest problem facing the inland waterway transport after the recession of flood was the unprecedented siltation in the navigational channels and routes, caused by the unprecedented flood. Hydrographic surveys by BIWTA and reports from pilots indicate the shoal formation in Meghna River (Dhaka-Chittagong and Dhaka-Barisal route), Kirtonkhola River (Barisal inland port basin area), Jamuna River (approach and basin of Notakhola Ferryghat), Padma River (approach and basin of Daulatdia Ferryghat), etc. The quantum of dredging requirement has been estimated to be very high. Various infrastructure development works have to be undertaken by the BIWTA to mitigate the damages. However, no hydrographic survey or data collected have been undertaken in the thousands of canals and khals throughout the country, which



are mainly used by the country boats for navigation. Some useful data and information have been obtained by the present survey on the type and extent of such damages to the waterways as a result of the flood. These are furnished in Table 5. The main effect has been found to be the shifting of the routes causing loss and disruption to business activities and movement of passengers and cargoes. Many routes have been closed altogether. The worst case reported is the closure of a 30-km waterway in Nabinagar, Brahmanbaria. The route is the one connecting Bhairab with Belabo. As a result, the vessels, mainly country boats have to take a much longer route with consequent increase in the transportation cost. This increase in cost is not fully compensated by the increase in the freight. Some routes were closed as early as October 1998. Closures of waterways have necessitated the use of alternative routes that are much longer. For example, the Nasirnagar-Kuliar char route at Nasirnagar, Brahmanbaria had a length of 25 km. However, this closure has also created an alternative route, which is 55 km long. Damage or Shifting of Ghats Due to Erosion or Accretion Causing Inconvenience to the Passengers, Crew and Boat Operators

Most of such damages were experienced by the formal inland water transport sector. Various establishments of the Bangladesh Inland Water Transport Authority (BIWTA) were extensively damaged by the flood. Being a formal sector and under one organization, survey were taken up immediately after the recession of flood to assess the extent and nature of damage. The following excerpts are reproduced from the BIWTA flood damage assessment report (IEB, 1999).

The installations related to the inland waterways were constructed/reconstructed or rehabilitated by the BIWTA based on the experiences of the 1998 flood. The levels of the jetties and bank protection works, approach roads, etc. of the waterfront structure were raised above the 1988 flood levels wherever possible. Some of the problems faced during the 1998 floods in water transport sector are as follows:

(i) Many pontoons in the rural areas were set adrift by strong currents. (ii) All wooden jetties were damaged. In some cases the river and bank erosion

caused the wooden piles to dislocate. In other cases, the country boats, engine boats etc. plied over the decking severely damaged them.

(iii) Small-scale bank protection works between the jetties and the approach roads were damaged mainly by vessels ramming against them in their attempt to discharge passengers and cargo to flood-free points.

(iv) The approach roads to jetties, the internal roads, the parking yards within the ports remained underwater for considerable time causing soil subsidence and potholes.



(v) The floors and walls of the inundated port terminal buildings were severely damaged by country boats plying and dragging over them.

Table 4: Accident occurring with mechanized boats during the 1998 Flood

Place Name of river

No. of boats affected

Property loss (Tk.)

Loss of life

Reason of accident

Cap-sized

Dama-ged

(see below)

Sunamganj, Sylhet 1 70,000 - - Bajitpur, Kishoreganj Khora Utra 5 400,000 - - Rjabari Sadar - 13 230,000 - - Brahmanbaria Sadar Titas 1 10,000 - - Birgaon, Nabinagar Meghna 1 20,000 - a Nasirnagar, B. baria Meghna 20 920,000 - - Kalapara, Patuakhali - 4 370,000 12 - Mahipur, Patuakhali - 1 375,000 60 b Kalapara, Patuakhali Shapurer

Khal 2 600,000 -

Indurkani, Pirozpur Panguti River

1 380,000 1 b

Bhairab Ferryghat Meghna 1 10,000 - c Sujanagar, Pabna - 14 220,000 - - Chilmari, Kurigram Brahmaputr

a 1 30,000 - d

Kotwali, Khulna Kaji Bacha 1 80,000 e Dakop, Khulna Shibsa 4 400,000 4 f Paikgacha, Khulna Korulia

river 2 200,000 - f

Paikgacha, Khulna Rupsha river

1 75,000 1 g

Kotwali, Barisal Kirtinasha 19 1,600,000 - g Tahirpur, Sunamganj - 25 1,200,000 - h Swarankhola, Bagerhat Jamtalar

Khal 1 125,000 - g

Morolganj, Bagerhat Payra River 2 200,000 - h Swarankhola, Bagerhat Bishkhali 1 180,000 - h Morolganj, Bagerhat Sundarban

Khal 1 125,000 g

Swarankhola, Bagerhat Bhola river 1 200,000 1 g Swarankhola, Bagerhat Boleshar

river 1 180,000 - g

Morolganj, Bagerhat Kochar 1 200,000 - i Morolganj, Bagerhat Pangachi 1 210,000 - h

a: collision with launch in strong current, b: strom, c: collision with ferry, d: Engine stopped and boat drifted by strong wind, e: Bottom damaged by strong current and wave, f: Drifted by strong current, g: Drifted by strong current and wave, h: Slamming against wave, i: Drifted by current and hit by shoal.



Table 5: Siltation of rivers causing loss of navigability in the dry season

Location River/route Damages occurred Shifting of river route

Kotwali, Barisal Kalijira River Silted at places minor shift Kotwali, Barisal Kumarkhali River Almost fully silted -do- Kotwali, Barisal Jagua-Gutia Massively silted -do- Kotwali, Barisal Dapdapia-Ranirhat Massively silted -do- Kotwali, Barisal Char Jagua-

Kumarkahali Silted in many places -do-

Shujanagar, Pabna Dhawapara-Rajbari Silted in many places -do- Shujanagar, Pabna Nasirganj-Belgachi Massively silted -do- Bera, Pabna Notakhola-Nagarbari Silted in many places -do- Dakop, Khulna Bhangon-Nowai route -do- -do- Chalna, Khulna Joarkhali-Purtan Masjid -do- -do- Paikgacha, Khulna Bhangan-Paikgacha -do- -do- Raipura, Narsingdi Chitri-Nabinagar -do- -do- Raipura, Narsingdi Gokon-Nabinagar -do- -do- Raipura, Narsingdi Baluchar-Maijchar -do- -do- Kalapara, Patuakhali Bhangan-Kumirmara -do- -do- Nabinagar, B’baria Bhairab-Belabo Massively silted 30 km closed (Nov.) B’baria, B’baria Chitri-Bhairab -do- 20 km closed (Nov.) B’baria, B’baria Chitri-Narsingdi Silted in many places 30 km closed (Nov.) B’baria, B’baria Chitri Gokan -do- Increased from 17

to 20 km Nabinagar, B’baria Baish Mouza-Bhairab

(Meghna) -do- Increased 8 km

Nabinagar, B’baria Baish Mouza-Nabinagar (Titas)

-do- Increased 10 km

Nabinagar, B’baria Baish Mouza-B’Baria (Pagla)

-do- Increased 20 km

Nasirnagar, B’baria Nasirnagar-Kuliar char -do- Increased from 30 to 55 km

Rajbari, Rajbari Dhawapara-Rajbari -do- Increased 8 km Rajbari, Rajbari Nazirganj-Belgachi -do- Increased 8 km Mitamoin, Kishoreganj

Sadhana-Ajmeer -do- 25 km closed

Mitamoin, Kishoreganj

Banglapara-Adampur -do- 8 km closed

Bajitpur, Kishoreganj Chatalganj-Mitamoin -do- 3 km silted Sarankhola, Bagerhat Sarankhola-Badaghat -do- 1 km silted Tahirpur, Sunamganj Fazilpur-Durlavpur -do- closed in late

October Tahirpur, Sunamganj Fazilpur-Lalpur -do- -do-



(vi) The semi pucca pilothouses were seriously damaged. (vii) The buoys drifted off and the shore beacons became dislocated due to

erosion and strong currents. The navigational markings made of bamboo and bamboo mats were washed off.

(viii) The problem went out of control when the floodwater rose so high as to make the driving of the bamboo poles impossible.

(ix) Low-cost waiting sheds constructed near the launch stations or kheaghats got damaged by floods. In some cases, they were even washed away by flood or bank erosion.

(x) All RCC jetties were inundated by floodwater by a meter or more, but their use for the purpose of handling food grains, relief materials, etc. continued. The bottoms of loaded country boats and engine boats dragging over them damaged the decking. Dashing of the vessels damaged the columns and outer beams.

(xi) The pontoons of the port terminals were floated up beyond their design limits and reached extreme conditions making them dangerous. Any further increase in flood level could have made them drift off and cause the gangways supported on them to slip into the river. This could have resulted in almost irrecoverable damage in terms of time and money.

(xii) The Decca stations were flooded for an inordinately long period and water submerged the floors and foundations of the machines. Lighter equipment could be removed to higher places. However, the operation of the hydrographic survey was kept undisturbed by taking emergency measures but the station at Chandpur could not operate for about 10 days when the power unit had to be disconnected and shifted to safer places.

(xiii) The lifeline between the capital city Dhaka and the country's northern and southern region through the Aricha-Notakhola, Aricha-Daulatdia and the Mawa-Charjanajat ferry routes were maintained in extreme conditions with great labor and costs. They went out of operation only when their connecting highways became unusable.

However, the surveys and assessments carried out by the BIWTA were

limited to their installations. Such installations are limited to the defined Class A, B, C and D waterways as defined and fully or partially maintained by the BIWTA. These waterways do not cover practically the bulk of the waterways of the country. The country boats carry more than 50% of the passengers and cargo throughout the country. These boats ply largely on the arterial canals and khals, which is not maintained by the BIWTA except for a limited number of ghats, installed by the authority. Surveys were carried out to assess the extent of damages to the ghats. The objective was not only to assess the physical damages



to the ghats but also the difficulties faced by the passengers and traders due to the dislocation of the ghats or the relocation of the channels or waterways.

In order to assess the same, information were collected on the ghats, which have either been damaged or dislocated or rendered useless/difficult due to the flood. The results of the investigation of such ghats are presented in Table 6. It is seen from the table that a quite large number of ghats have developed wet access ways up to 50 meters which were not present before the flood. As a result, the passengers faced difficulty in embarkation and disembarkation and the cost of loading and unloading of the cargo increased considerably. Moreover, the dry access way became as much a 1 km, which was previously much less. This caused similar effects as the wet access way, albeit to a lesser intensity. It can also be seen from the table that quite a good number of ghats shifted due to the flood by up to 3 km. These shifting were reportedly made to adjust with the shifted river course or siltation during and immediately after the flood.

Table 6: Damage or shifting of ghats due to erosion or

P. Station District Dry Access way Wet Access way Shifting of ghat Mitamoin Kishoreganj 90 meters 40 meters - Bhairab Kishoreganj 15 meter 10 meter - Chilmari Kurigram 1 Km Nil - Rajbari Rajbari 125 meter Nil - Nabinagar B’Baria 200 meter 20 meter - Nabinagar B’Baria 125 meter 20 meter 200 meter B’Baria B’Baria 250 meters 50 meter 125 meter Nasirnagar B’baria - - 2 km Sujanagar Pabna 125 meter - - Bera Pabna 1.5 Km - - Kotwali Barisal 1 Km 20 meter - Kotwali Barisal - - 3 km Swarankhola Bagerhat 500 meter 20 meter 1 km Tahirpur Sunamganj - - 1 km Hatiya Noakhali 50 meter 50 meter - Hatiya Noakhali 100 meter 20 meter - Chilmari Kurigram 1 Km 20 meter 500 meter

CONCLUSIONS AND RECOMMENDATIONS The colossal flood of 1998 has damaged the inland water transport sector as well as other sectors of the Bangladesh economy. Accidents occurred with mechanized country boats during the flood due to a host of reasons, such as excess pressure on the boats, poor maintenance, untrained crew etc. Such



damages could be quantified in monetary terms. However, there are other damages such as those due to siltation of rivers and canals and damaged/dislocation of the ghats. The financial and economic impact of such damages can not be easily quantified. The extents of such damages have been quite serious and directly affected the lives of the poor boat operators. This also caused an increase in the transportation cost.

In order to mitigate the sufferings of the people a series of measures may be taken which include: i) There is a need of restoring the landing stages such as pontoons, jetties, ghats

pilothouses, channel markings, Decca chain equipment etc. ii) The dredging of the river routes with equal emphasis on the large rivers and

the small arterial canals and khals to ensure easy movements of vessels in all river routes of the country throughout the year.

iii) Regulate the country boat operation to ensure minimum technical standard of the boats and efficient operation by skilled crew

iv) Necessary fund should be allocated to repair and rehabilitate the damages caused by the flood.

v) A database should be prepared and maintained on the water transport sector of the country incorporating the mechanized country boat sector. Such database should contain information on the river transport routes, their economic importance, siltation situation etc.

REFERENCES Chowdhury, J. U. and Islam, A. K. M. S., (1999), “Hydrological Characteristics

of 1998 Flood in Major Rivers”, to be published in “BUET Studies on 1998 Flood in Bangladesh”.

IEB, (1999), “Report of Task Force Committee on Flood Management”, the Institution of Engineers, Bangladesh, Ramna, Dhaka.

Rahim, A., Gama, B. A. and Palmar, C., (1993), “Mechanization of Country Boats: The Challenge of Technical Improvements”, Country Boat Pilot Project, Bangladesh Inland Water transport Authority.


Concept of Flood Shelter to Cope with Flood

M. J. B. Alam and Md. Hossain Ali Department of Civil Engineering

Bangladesh University of Engineering and Technology, Dhaka-1000, Bangladesh Abstract Shelter means protection during period of crisis. The people of Bangladesh, a country prone to natural disasters, are familiar with the ‘Cyclone Shelter’. But the concept of flood shelter is relatively new. Traditionally, embankments are most commonly used as shelters during floods. Considering the geography of the country and the amount of floodwater that passes through it during the monsoon period, it may easily be seen that building embankments are not feasible in many areas of the country. In those areas flood shelter may be a very effective tool for coping with flood. In some of the areas of the country the local people have already established flood shelters. This paper investigates the advantages, disadvantages and problems of the flood shelter and proposes measures to improve their usefulness. INTRODUCTION Human being exists in this world in an adapted ecological relationship with the surrounding environment and has to live with a variety of natural hazards, which threaten life and property. River flood is the most common type of global hazard, encompassing a wide range of events from largely unpredictable and localized flash flood to anticipated widespread floods. Sheehan and Hewitt (1969) reported that floods accounted for about 30 percent of all natural disasters and 40 percent of the fatalities.

A flood may be defined as a discharge, which exceeds the channel capacity of a river and then proceeds to inundate adjacent flood plain. Flood hazard is a function of both geophysical attributes and human activities. Since man is still


M. J. B. Alam and Md. Hossain Ali


unable to control the basic atmospheric processes that produce floods, he has attempted to cope with the hazard by taking alleviation measures. Through the application of high technology and massive investment of capital, the flood threat to human life has decreased appreciably in most of the developed countries.

Flood is a perennial problem for Bangladesh. Almost every year flood causes enormous damage to the people and economy of the country. To reduce the problem, a lot of investment has been made to build embankments for protecting important areas. The current records of damages by flood demonstrate that these measures could not bring about the desired results. Considering the geography of the country and amount floodwater in monsoon period, it may be concluded that only the structural measures to cope with flood will not be very successful. Together with structural measures such as building embankments and structures, other evasive measures may also prove to be highly effective. Evasive measures may include flood shelters, which are similar to cyclone shelters (BUET-BIDS, 1992). In this study, the advantages, disadvantages and scope of a flood shelter have been examined. For this purpose two flood shelters in Kurigram district have been selected. During the devastating flood of 1998 these shelters were used by the people of the locality. On the contrary, many people in the vicinity of the shelters had chosen houses of their relatives, located far away, as shelters. This study also investigates the reasons behind their choice. FLOODS IN BANGLADESH AND THE FLOOD OF 1998 Flood is a recurring problem in this flat and riverine country. Some parts of India and Bangladesh experience flood almost every year with considerable damage. Flood statistics for Bangladesh are available since 1954 which is summarized in Table 1. The floods of 1954, 1955, 1974, 1987, 1988 all caused enormous damage to properties and considerable loss of life. During the middle of 1998, Bangladesh experienced the most devastating and prolonged flood in its history, which caused enormous damage to the economy of the country. The extent of damage caused by the flood is estimated to be around 3.0 billion US dollars (Annual Flood Report, 1998).

The prime reason of flood in Bangladesh is heavy rainfall in the upstream of the rivers flowing through the country. Three major rivers of the world flow through Bangladesh before discharging into the ocean. Heavy rainfall over the catchments of these rivers could produce an average runoff of about 1,009,000 million cubic meters. If the whole water were stored, the country would have been flooded to a depth of 8 to 10 meters.

Concept of Flood Shelter

Table 1: Year-wise Major Flood Incidences in Bangladesh

Flood Affected

Areas

Flood Affected

Areas

Flood Affected Areas

Year

Sq. km %

Year

Sq. km %

Year

Sq. km % 1954 36800 25 1969 41400 28 1987 57300 39 1955 50500 34 1970 42400 29 1988 89970 61 1962 37200 25 1971 36300 25 1995 32000 22 1963 43100 29 1974 52600 36 1996 35800 24 1968 37200 25 1980 3300 22 1998 100250 68

During the monsoon of 1998, due to excessive and intermittent rainfall in the country and in the upper catchment areas from July to September, all the rivers of the country experienced significant increase in flow far above the danger level. The flood situation turned worse from the middle of July and by this time the low-lying areas of the country had already gone under water. At that time, about 45,000 sq. km of 37 districts of the country were affected by flood. Although flood situation started improving in early August, the flow of the two main rivers of the country- Padma and Brahmaputra-Jamuna increased significantly during the middle of August. This was caused by heavy rainfall in the upper catchment areas. By the end of August flood situation became worse and about 60,000 sq.km area of 42 districts were affected. During the early September the flow of the major rivers increased abruptly, worsening the condition. The flood situation became worst during the second week of September and about 75,000 sq.km area of 52 districts were affected during that time. The flooded condition existed for about three months, from early July to the last week of September, in different magnitudes at different places. Thus flood of 1998 became the most prolonged flood in the history of the country. The total flood inundated area was about 1,00,250 sq.km (68 percent of the total area of the country) affecting 53 districts (Annual Flood Report, 1998). CONCEPT OF FLOOD SHELTER The implementation of any flood alleviation scheme has four basic aims – (i) to reduce flooding, (ii) to reduce damage, (iii) to save lives, and (iv) to save property. A particular scheme may cover all four of these e.g., building embankments to protect vulnerable areas. On the other hand, small-scale projects such as flood shelters may help in saving lives and properties. These shelters can be used to manage relief and rehabilitation activities in an organized way. The




shelters can also be used as schools and community centers when there is no flood.

Clearly, floodplain evacuation is neither socially desirable nor economically viable, particularly in densely populated and large areas. But providing shelter in the most vulnerable areas, which cannot be protected by structural measures due to practical reason, seems to be a plausible solution.

The study area selected in this research is an area that is struck by flood every year. The geography of the area is such that any measures, such as building embankments, is not economically or technically feasible. Considering the fact, the local people, administration and NGOs have taken steps to build shelters for the affected people. These shelters are used as school, community center, medical center and offices of charity organizations during periods other than flood. During flood these shelters store emergency medicine and relief materials other than providing shelter. In the present study, the advantages and disadvantages of these flood shelters are investigated so that it can be improved further and implemented in other areas of the country.

THE STUDY AREA The study area isa part of Kurigram district which is one of the worst victims of the flood of 1998. The research is aimed at evaluating the necessity and acceptability of flood shelters in this area.

The prolonged and devastating flood of 1998 caused serious damages in the whole Kurigram district and the area under study is one of the worst victims of this flood. In this area, flood is caused by heavy rainfall and overflow of the excess water of the rivers Dharla and Dudkumar which have their sources in India.

In 1998, floodwater remained in the study area for about 90 days. This flood affected about 40 thousand people of the study area. In this area, floodwater rose up to 10 to 15 feet above the ground level. The floodwater went up to the top of the houses. People took shelter on high roads, embankments, and in the houses of the relatives. Thousands of them rushed to the cities. Ripe crops in the fields were submerged and cattle, poultry and household belongings were washed away. Lack of pure drinking water and lack of hygienic latrine facilities caused various diseases and many lives were lost. An assessment of the damages caused by the flood 1998 in the study area is presented in Table 2.


Table 2: Assessment of the Damages Caused by Flood of 1998 in the Study Area

Area (Upazila) Kurigram Nageswari No. of Union

No. of villages Affected family Affected people

Damaged houses (full) Damaged houses (partial) Damaged crops (hectare)

Affected damaged embankment (km) Damaged bridge (No)

Damaged educational institutes (Nos)

3 10

7366 31283

168 516 546 10

120 2

1 1

1000 5000 100 200 79 1 2 2

(Source : Zibika, A local NGO) FLOOD SHELTERS IN THE STUDY AREA There are two flood shelters in the study area. One is at Zatrapur and the other is at Mogal Basa in Kurigram District. Their present conditions are described in the following sections. Zatrapur Flood Shelter

This flood shelter is located in the village Ghanoshampur under Zatrapur Union of Kurigram Upazilla beside the Kurigram-Zatrapur Road. The district police station is 8 km to the east and Zatrapur market is half a kilometer to the south of it. It is 700 m away from the Brahmaputra/Dharla river. The Zatrapur flood shelter was constructed for the purpose of giving temporary shelter to the people of Panch Ghasi, Zatrapur and Ghogadah unions, who are the victims of the flood caused by the overflow of the Buamaputra/Dharla river. The Zatrapur flood shelter was built by LGRD and Co-operative Ministry during 1975-76 fiscal year. The coverage area of this flood shelter is 2.5 acres. There are four sheds each 220ft by 20ft, two katcha kitchens, ten latrines and two tube-wells in the shelter area. RCC pillar and wooden trusses have been used as framing elements, G.I. sheet has been used in the roof and the houses have been enclosed by bamboo. The construction cost was 6,50,000 Taka and 1500 kg of wheat. Later, in 1978-79, ten inch-brick walls were constructed up to plinth level and five-inch brick walls were constructed as partition walls. During the devastating flood of 1998, more than 2000 people with their domestic animals took refuge in this flood shelter.




According to the people of the area there is no doubt about the necessity of flood shelter of adequate capacity along with sufficient facilities. But the flood shelter of Zatrapur in in a poor state due to lack of maintenance and supervision. Besides it is not clear who is responsible for its maintenance and as a result its properties and materials are being stolen. There are no health and utility facilities in the shelter. It is necessary to take proper measures to maintain and rehabilitate the flood shelter so that it could be used by the people in this area in the event of a future flood. Zibika Flood Shelter

This flood shelter is located in the village Char Shetaizar under Mogol Basa Union of Kurigram Upazilla beside the Dharla river. The district police station is 6 km to the south-east of it. This flood shelter was constructed for the purpose of giving temporary shelter to the people of the village of Char Shetaizar during the flood caused by the overflow of water of the Dharla river. It was built by Zibika, a local NGO in April,1996. The area of this flood shelter is about 1.5 acres and it has two permanent sheds of size 30ft by 15ft. There are provisions to build temporary sheds, if needed. There exist a public toilet, a bathroom and a separate toilet for the women. The elevation of the shelter is about 8ft above the ground level. There is a pond in front of the shelter. The design capacity of the flood shelter is 100 families with a total of 300 people. The construction cost of this shelter was Tk. 540,000/-. The sheds are made of RC pillar and G.I. sheet supported by wooden truss. The walls of the sheds are also made of G.I. sheet. Photographs of the shelter are shown in Fig. 1. As the flood shelter is maintained and supervised by Zibika, its present condition is good. People who took shelter here during the 1998 flood were quite happy with the facilities provided to them. In this shelter, there are arrangements for keeping domestic animals. There is also provision for individual cooking. Special care is taken to provide hygienic condition. People are able to get pure water and the bathroom/toilet facilities are very good. These are supervised on a regular basis. Two sheds are used as a primary school and a clinic during the non-flood seasons. About 212 families took shelter in it during the 1998 flood. A medical team worked in the shelter during that time and provided health treatment to 110 male, 210 female and 105 children. But the supply of food was insufficient. In the future, this shelter should be expanded and its elevation should be increased.


Figure 2: Photograph of Zibika Flood Shelter




Usage of the Flood Shelter and People’s Opinion

This section presents the results of a survey to grasp people’s perception about the flood shelter. For this purpose a survey was conducted after the flood of 1998. From the survey it is evident that the people who live in katcha houses are the main users of the shelters. Among the users, about 75 percent are farmers and laborers (Fig. 2). In the study area, about 74 percent of the houses are katcha and the rest are tin shed (Fig. 3). During the flood of 1998 about 27 percent of the houses were washed away, 18% were completely damaged and 55 percent were partially damaged (Fig. 4).

OCCUPATION OF PEOPLE STUDIED

44.4%

2.7%30.4%

3.5%

12.5%6.6%

Farmer Service Holder Day LabourerHouse Wife Others Shop Keeper

Figure 2: Occupation of People Survey in the Study Area

TYPES OF HOUSES OF THE STUDY AREA

26.2%

73.8%

Kancha Tin house Building Others

Figure 3: Types of Houses in the Study Area


DAMAGES OF HOUSES CAUSED BY THE FLOOD OF 1998

26.5%

18.1%

55.3%

Washed away Fully damaged Partially damaged

Figure 4: Damages Caused by Flood in the Study Area

Figure 5 shows that about 87 percent of the people had to leave their houses during the flood of 1998, among them 76 percent knew about the existence of the flood shelter in the locality beforehand. Of the people who left their houses during the flood, 58 percent went to flood shelter and 39 percent took shelter on high roads and embankments (Fig. 6). Most of the people took their livestock with them, as it was their only asset (Fig. 7).

PEOPLE'S AWARENESS ABOUT FLOOD SHELTERS

76.7%

23.3%

People know about the flood shelter in their locality People do not know about the flood shelter in their locality

Figure 5: People’s Awareness about Flood Shelter in the Study Area




PLACES USED BY THE AFFECTED PEOPLE DURING

THE FLOOD OF 1998

57.9%

36.6%

2.8%

0.9% 1.9%

Relative's house Flood shelter High road Embankment Others

Figure 6: Places Used as Shelter by People during the 1998 Flood

PLACES USED BY THE PEOPLE FOR KEEPINGTHE DOMESTIC ANIMALS

38.7%

4.2% 3.66%1.6%

51.8%

Relative' house Flood shelter High road Embankment Others

Figure 7: Places Used by People to Keep Livestock during the 1998 Flood Most of the users of flood shelters had no other alternatives for taking shelter

as shown in Fig. 8. Although many of the users did not mention any specific problem, only few of them were satisfied with food supply and toilet facilities. A substantial portion of the people demanded separate arrangement for the women (Fig. 9). Although there were some complaints, almost all the people mentioned that the shelters were of great help to them, which assisted them to survive during the flood and to get rehabilitated afterwards.


ADVANTAGES OF FLOOD SHELTER

5.8%

16.5%2.9% 1.9%

72.8%

There was no alternative Minimum food supply maintainedNo special problem Good toilet facilityOthers

Figure 8: Advantages of Flood Shelters as Suggested by the People

DISADVANTAGES OF FLOOD SHELTER

11.8%

14.2%

15.0%1.2%13.0%

15.4%

29.3%

Lack of Privacy of Women Too low capacityInsufficient food supply Difficulty in cookingDirty surrounding OthersThe Shelter is Vulnerable to Flood

Figure 9: Disadvantages of Flood Shelters as Suggested by the People




CONCLUSIONS Flood is a perennial problem for Bangladesh and the people of the country must live with flood. Although by building embankments the menace of flood can be reduced to some extent, it cannot be eliminated for all the areas. For the areas of frequent flooding evasive measures to reduce the sufferings of the people seem to be effective. Flood shelter is a form of evasive action. These shelters may provide the victims of flood with place to survive during flood and get rehabilitated afterwards. The shelters can also be used as the nucleus of relief and rehabilitation activities in a broader perspective. During periods other than flood, the shelters may be used as schools, community centers, health centers etc. To investigate the effectiveness of flood shelters, a survey was conducted in Kurigram district, one of the worst hit areas of the country during the flood of 1998. This area was selected because flood shelters, constructed by the local people and administration, existed here for more than 20 years. Most of the people mentioned that the shelters were of great help to them. The people provided some suggestions to improve the condition and usability of the shelters, which include improvement of toilet facility and separate arrangement for the women. From the results of the study it is evident that the shelters can play a significant role in the survival and rehabilitation of the flood affected people. The study can further be extended to incorporate technical specifications regarding the design and construction guidelines on ground elevation, structure type and layout of the flood shelters. REFERENCES Annual Flood Report (1998), Flood Forecasting & Warning Centre, Processing

and Flood Forecasting Circle, Bangladesh Water Development Board. Sheehan, L. and Hewitt, K. (1969), A Pilot Survey of Global Natural Disaster of

the Past Twenty Years. Natural Hazard Research Working Paper No. 11, Department of Geography, University of Toronto.

Smith, K. and Tobin, G. (1979), Human Adjustment to the Flood Hazard. Longman Group Ltd., UK.

BUET-BIDS (1992), Progress Report-II, Multipurpose Cyclone Shelter Programme, World Bank/UNDP/GOB Project.


A Study on the Disaster Response for Shelters

During the 1998 Flood in Dhaka City

K. M. Maniruzzaman and B. M. Alam Department of Urban and Regional Planning

Bangladesh University of Engineering and Technology, Dhaka-1000, Bangladesh Abstract Provision of emergency shelter for disaster victims is an important aspect of post-disaster response. This happens in Dhaka City through an informal and ad hoc process in the absence of any disaster plans. The present paper reports an evaluation of the process in the wake of the specific instance of the 1998 flood, and some managerial and operational issues associated with it. Problems are identified based on field visits and a small survey, and some suggestions for improvement are given in conclusion. INTRODUCTION Shelter is recognized as a basic human right. While it may not be possible, under the prevailing political system, to ensure shelter for each citizen, it is nonetheless the responsibility of the government to create an environment where most people can fulfill this basic need. The government has a more direct responsibility to provide emergency shelter though, when environmental disasters render people homeless temporarily.

Disasters may physically destroy or damage homes. Or, in some cases, the environment of an area may degrade to an uninhabitable level in the aftermath of a disaster. In yet other cases, the government may force residents to leave their homes to ensure public safety in anticipation of an impending disaster. In any case, the government have to provide alternative arrangements for shelter until the displaced residents can return to their own homes.


K. M. Maniruzzaman and B. M. Alam


When we speak of these alternative arrangements, we mean not only an enclosed space protecting the inmates from the elements of weather, but also the basic services and utilities that people need for a healthy life in a livable environment. The minimum of such provisions should include potable water supply, sanitation and medical facilities. Other supplies may be required depending on the specific situation.

Providing shelter for disaster victims is an important task in disaster management. Disaster management is a complex process that can be conceptualized as consisting of four non-linear, overlapping stages (Maheshwari, 1997): (i) preparedness, (ii) response, (iii) recovery, and (iv) mitigation.

The act of shelter provision takes place in the response phase, but this is ideally done by local/central government agencies according to a disaster plan prepared in advance in the preparedness phase. Disaster plans should spell out, among other things, designated emergency shelters with adequate capacity and within reasonable distance for each residential neighborhood.

The present paper looks into the response from different quarters to the temporary homelessness of victims of the 1998 floods in Dhaka City. Floods periodically affect Dhaka City and the hazard of other forms of disasters also exists. However, there are no disaster plans to mitigate and cope with the effects of disasters. In the absence of a disaster plan in Dhaka, it is not clear who is responsible for the arrangements of temporary shelters and provisions for city dwellers who fall victim to disasters. Yet numerous shelters were opened in the city during the 1998 floods (as well as during previous instances of disasters) providing temporary relief to a large number of victims who had to leave behind their homes. In the present study, the process and different aspects of shelter provision and management have been investigated and measures for overcoming identified drawbacks have been suggested.

The objective of the study was to understand, in the light of the 1998 floods, how temporary shelter is provided to disaster victims in Dhaka City in the absence of any disaster plans. The specific objectives of the study were to: (i) identify who take the initiative to set up emergency shelters, (ii) identify who take the responsibility to operate, manage and provide services to the shelters; and (iii) assess the problems associated with the present informal system of emergency shelter provision from the point of view of the operators, shelter seekers and the authorities of the premises. It is expected that the insights drawn from the study would help formulating the emergency shelter aspect of a much needed disaster plan for the city.

The authors visited a number of shelters while they were in operation. Key figures in the bodies responsible for the premises where the shelters were set up, or owners, in case of private property, were interviewed. Data on the location of shelters and number of shelter seekers were obtained from the Dhaka City

Disaster Response for Shelters


Corporation (DCC). A random sample of 51 shelters out of a total of 301 in the list were selected for a small scale questionnaire survey to obtain information on the management and operational aspects of the shelters. The respondents of the survey were the key figures i.e., owners of private residences, heads or senior teachers of educational institutions, managers of industrial establishments etc. It was recognized that a survey of the shelter seekers could give a better understanding of the issues at hand, but since they had already started to return to their homes by the time our survey could be launched, such a survey was not done. THE SHELTER SEEKERS The flood of 1998 was arguably the worst in Dhaka City in recent times in terms of duration. However, some of the city areas that were inundated in the preceding serious flood of 1988 were spared this time. This was due, in part at least, to the flood protection embankment that has been built since the last flooding. In any case, the people most affected⎯those who fall victims first and are relieved last⎯are mostly from the lowest socio-economic tiers of society. The competition for land forces them to live on the most vulnerable land. People who flock to the shelters come from this level of society. Victims from better socio-economic backgrounds shun the shelters because of psychological barriers and lifestyle differences across class divisions.

Those who took refuge in the shelters included many members of the same family. The number of resident families as well as the total number of people was known for 251 shelters. Analysis of that data revealed a mean family size of 5.61. Male-female ratio for adults in the 297 shelters for which the breakdown by sex of the resident population was known tilted slightly in favor of females by 1:1.05, although for Dhaka as a whole, the ratio is 1.3:1 (BBS, 1997). This may be due to male members of families staying back to guard their homes and belongings, or the relative ease for single males to arrange alternative shelter elsewhere. SPATIAL DISTRIBUTION OF SHELTERS Figure 1 shows the distribution of the 301 shelters in the DCC list among the 90 wards in the city. A total of 2,38,413 persons sought refuge in these shelters, as of 20 September 1998, according to the same source. Census data on ward populations were not available since the ward boundaries in 1998 were different from what they were during the last population census in 1991. There were 90 wards in 1998 instead of the 75 that were in existence during the last census. In a recent study, the population of 75 wards was redistributed into 90 wards through



GIS-based areal manipulation, with manual adjustments for areas with highly heterogeneous population densities (Management Sciences for Health, 1996). The new ward populations were then projected for 1997. We have used these figures to calculate the ratio of shelter seekers in each ward. This may serve as an indicator for comparison of the relative impact of the flood on the residents in different wards, at least in terms of shelter, as shown in Fig. 2. The figure shows that the worst affected wards were in the east and south-east of the city. The refugee-to-population ratio was lower in the central wards, which are generally located on somewhat higher grounds, and in the west thanks to the flood protection embankment. It must be borne in mind, though, that many of the refugees in a shelter in a certain ward may come from outside that ward. During our visits to the ward, we have seen people staying in shelters that were not closest from their homes. In many cases the nearest shelter had already filled up and they had to move further on to find refuge. In some cases the shelter seekers decided to stay in shelters far from their homes because they were nearer to their places of work. While visiting a shelter in the Shukrabad area, we encountered a family that had just arrived from distant Barisal. Also, many ‘floating people’, who were not victims of the 1998 floods in the proper sense, found temporary refuge and other free benefits in the shelters. ORIGINAL USE OF SHELTER BUILDINGS One of the primary matters of interest was where the shelters had been set up. It was found that the overwhelming majority of the shelters (76.7 percent) were located in buildings of educational institutes such as schools, colleges and madrasahs (see Table 1). These are convenient locations for emergency shelters because they can provide large spaces indoors, and usually outdoors as well, under public or communal ownership and/or management. Among the educational institutes, 60.6 percent are non-government. Only one, Jagannath University, is for tertiary education. Twelve community centers, owned by the DCC, provided shelter to the flood victims. The private residences and commercial buildings that acted as flood shelters were mostly under construction and therefore not in use when the flood occurred. THE SAMPLE SURVEYED

As stated earlier, 51 shelters were randomly selected for a small questionnaire survey regarding the operational and management aspects of the shelters. The original use of the shelter buildings are given in Table 2. As expected the majority of the shelters were educational institutes. The two shelters included in the ‘Other’ category were a club and a market. The spatial distribution of the sample is shown in Fig. 1.



Figure 1: Distribution of shelters among the 90 wards



Figure 2: Wards categorized by ratio of refugees to total population during the 1998 flood



Table 1: Original use of building(s) used for flood shelters

Original Use of Shelter Number % School 200 66.4 College 28 9.3 Madrasah 3 1.0 Community Centre 12 4.0 Residence 20 6.6 Office 6 2.0 Factory 5 1.7 Institution/Hospital 3 1.0 Other 24 8.0

Total 24 8.0 Source: Calculated from DCC data. Table 2: Original use of buildings of sample shelters

Original Use of Shelter Number % Educational Institute 40 78.4 Community Centre 2 3.9 Residence 4 7.8 Office 1 2.0 Factory 2 3.9 Other 2 3.9

Total 51 100.0 INITIATIVE TO OPEN SHELTERS

The authorities or owners of the buildings used as shelters, depending on their original use, were asked to indicate who took the initiative to open a flood shelter in their respective buildings. The respondents were presented with a set of given answers as shown in the legend of Fig. 3. 39.22 percent of the respondents informed that the shelter was opened at the request of local political leadership. Although the word ‘request’ was given in the questionnaire, discussions with many respondents suggested that in the given context they were rather obliged to heed to those requests. ‘Request’ may therefore be considered as a euphemism rather than taken in its literal sense. 7.8 percent of the respondents cited directives from higher authorities and 15.7 percent pressure of shelter seekers as the reason for opening the shelter. Almost a quarter (25.5 percent) did so at their own initiative out of a sense duty.



7.84%

7.84%

25.49%

43.14%

15.69%

At the directives of higher ups

At the pressureof shelter

At the request of local political leadership

Self

Other

Figure 3: Initiative to open shelters

OPERATION AND MANAGEMENT The shelters, which were opened at the request of politicians, were mostly operated by them either directly or indirectly. Shelters run by them numbered 22 in the sample surveyed (see Fig. 4). The respondents were involved in overall management of 12 shelters. Some of the shelters were managed almost entirely by NGOs (falling under the ‘Other’ category in the chart).

25.49%

43.14%

5.88%

25.49%Self

Local politicalleadership

Shelter seekersthemselves

Other

Figure 4: Party responsible for operation and management of shelters



Visits to the shelters revealed dismal conditions in most of them. Most of the shelters lacked minimum facilities like adequate water supply and were full of filth and squalor. The toilets were particularly filthy and unhygienic. The shelters did not have the resources to cope with the maintenance of the premises with large numbers of round-the-clock inmates. DCC provided some cleaning services on an irregular basis. The shelters looked after by some NGOs were in relatively better shape. CARE, for example, took responsibility for food, water (Fig. 5) and toilet facilities for 15 shelters around the city that were well maintained. They supplied food and water on a regular basis and installed extra toilets as necessary. Medical teams from DCC, the Army, NGOs and other organizations provided inoculation against diseases and other preventive and curative treatment.

Figure 5: Provision for water arranged by an NGO at a flood shelter In about three-fourths of the surveyed cases (37 out of 51 cases, to be precise) the shelters were not opened out of the free will of the respondents. We were therefore interested to know how those who took the initiative to open the shelters supported them. As far as the provision of potable water was concerned, the responsibility was borne by the party that took the initiative to open the shelter in exactly 14 cases (Fig. 6). However, thanks to WASA and other agencies, the party represented by the respondent had to bear the responsibility in only 9 (17.64 percent) cases.



15.69%31.37%

7.84%

17.65%27.45%

Self

DCC

WASA

Initiator

Other

Figure 6: Provision of water

Most of the shelters had not enough sanitary facilities to cope with a large

resident population, since they were not designed that way. Half the shelters had to manage with their own insufficient facilities (Fig. 7), resulting in overflowing toilets or sewers and unhygienic conditions and filth (Fig. 8). Many residents found it more convenient to relieve themselves elsewhere, further exacerbating the situation. The situation was better in shelters where extra temporary toilets were installed (Fig. 9). When the initiator of the shelter did not take the responsibility for cleaning, it was mostly left to the respondents to manage the job with their own resources as shown in Fig. 10. This involved not only the employment of sweepers and cleaners, but also the procurement of sterilizing agents, bleaching powder etc.

35.29%

50.98%

5.88%

7.84%

Self

Initiator

Other

DCC

Figure 7: Provision of sanitation



Figure 8: Poorly maintained, unhygienic toilets

Figure 9: Temporary toilets installed in one shelter

RELIEF SUPPLIES Thanks to their accessible urban locations, the shelters received abundant supplies of relief for the victims of flood. Relief material came from many different sources: the government, local government (DCC), NGOs and other sundry sources (mainly different social or community-based organizations,



business firms etc.). The most common relief material was food and medical supplies, received by all shelters from one or more of the four categories of sources. Food was the most common material donated by the public (47 cases), followed by medical supplies, presumably with a fair share of ORS (34 cases), clothing (24 cases) and potable water (20 cases). Food and medical supplies also topped the list of relief goods from government and NGO sources. There was virtually no coordination in the distribution of relief from different sources. In some shelters, supplies were not regular according to complaints from the inmates and published reports, while in some other centers (or at certain times) the supplies were rather generous and lavish (Sarker et al., 1998). While food was the most common donated item, many families were found arranging their own meals. They had temporarily lost their homes only, not their livelihood.

49.02%

35.29%

3.92%

11.76%

Self

Initiator

Other

DCC

Figure 10: Provision of cleaning service

CONCLUSIONS Dhaka City has no disaster management plans according to which shelter-related response can be directed and coordinated. Yet numerous emergency shelters were set up based on spontaneous and ad hoc decisions. Although these shelters provided succor to a large number of people whose homes were engulfed by the flood, there were problems that requires attention. As stated earlier, many shelter seekers had to take refuge far off from their homes because adequate shelters were not available nearby. Relief distribution in the shelters was arbitrary, inequitable and not always matching needs. Management and maintenance in many shelters were far from satisfactory, because of lack of skills, resources or the will on the part of the owners/managers of the premises. The premises where the shelters were set up suffered considerable loss in terms of both environmental degradation and damage of building and furniture. Most owners/managers



bitterly complained about this aspect irrespective of whether they opened up their premises willingly or not. Surprisingly though, 70% of all respondents in our survey declared their willingness to offer their premises for emergency shelter if the need arises in future. We can thus be assured that if another disaster strikes this capital city, disaster shelters would spring up again in different parts where the victims would huddle together for sometime until their homes are fit for habitation.

However, in order to instill some discipline, order, coordination and predictability, there must be some prior planning. This planning should be done in the context of an overall disaster plan with a wider range addressing not only shelter provisions, but also other key issues such as lifelines, traffic, law-and-order etc. A cell in the local government (DCC) can be trained and entrusted with the coordination job with representations from concerned government and non-government agencies as required. The shelter-related recommendations are as follows: (i) Possible emergency shelters should be identified and designated, based on

hazard and risk mapping of the city. The mapping can be done with GIS giving the probable number of homeless victims under different scenarios for different neighborhoods. The space available in the shelters must match the needs assessed in the GIS analysis.

(ii) The public must be made aware of the designated shelters as a measure of preparedness, and the opening of shelters must be promptly announced publicly.

(iii) Admission to the shelters should be strictly based on need and location of residence of admission seeker. Voters’ identity cards may be used to determine if a person is from the locality of the shelter and thus eligible to take refuge there.

(iv) The authorities responsible for the premises must have instructions on how to manage the shelters and their inmates. Standardized registration and record keeping of shelter seekers and their losses and needs should be enforced. Some compensation should be forthcoming for both the time and effort, and especially for the damage to the premises.

(v) All relief material and services to the shelters should be channeled through the cell so that they can be dispatched to the shelters in a coordinated manner and in accordance with assessed needs.

(vi) There should be a stock of water tanks, dispensers, etc. that can be easily dispatched and set up at the designated premises when they start functioning as shelters.



ACKNOWLEDGEMENT The authors wish to acknowledge the financial assistance for the study from BUET, and the assistance received from graduate students of the department of URP, BUET, including Monirul Alam and Fazlul Hoque, and research associates Shehzad Zahir and Provash Kundu REFERENCES BBS (Bangladesh Bureau of Statistics) (1997) “Bangladesh Population Census

1991: Urban Area Report”. Maheshwari, S. (1997) “GIS from a Disaster Management Perspective:An

Overview”,http://www.spatial.maine.edu/ucgis/testproc/maheshwari/disaster. html.

Management Sciences for Health (1996) “Bangladesh Urban Primary Health Care Project”, report prepared for GoB and Asian Development Bank.

Sarker, A., Haider, M., Islam, S., Ahmed, L. and Amin, M. (1998), “Flood Relief Shelters”, Star Weekend Magazine, November 11 issue, Dhaka.


Impact of the 1998 Flood on the Morphology

of Rivers around Bridges

M. Mozzammel Hoque, Sujit K. Bala, Syed Mohib Uddin Ahmed, M. Anisul Haque and Saifullah Al Mamun

Institute of Water and Flood Management Bangladesh University of Engineering and Technology, Dhaka-1000, Bangladesh Abstract The unprecedented flood of 1998 had severe impacts on bridge structures along with river channel processes. This paper presents an evaluation of flood impact on the morphological changes of the river Meghna, upstream and downstream of the Meghna Bridge and around Turag-Bhakurta Bridge on the river Turag. The Meghna river reach around Meghna Bridge has undergone significant changes in the river channel process during the 1998 flood. The morphological conditions around the Meghna Bridge were evaluated by comparing the results of detailed survey carried out in May (pre flood) and in October (post flood) in 1998. The bed level has changed at section from L10 to R1O showing about 20m depth near left bank while deposition has taken place at right bank. In the middle of the river near the section from R7 to L7 a sand bar existed before and significant erosion has taken place at the right side of the sand bar with a depth of about 20m. The erosion has extended up to 2 km. At some locations, no bed changes have occurred. Local scouring at right bank from R1O to R8 shows deposition of silt during flood time with scour depth varying from 2 to 22m, while at left bank from L4 to L3 shows a progressing scour depth of more than 20m and mostly 18m along the major part of the reach. Local scouring along left revetment at upstream and downstream of the bridge shows no significant changes and maximum scour at upstream is about 25m while at downstream it is about 22m. Maximum local scour depth around piers 7, 8 and 9 is about 20m. The Turag-Bhakurta Bridge on Turag-Bhakurta road was washed away by the 1998 flood.


M. M. Hqoue, S. K. Bala, S. M. U. Ahmed, M. A. Haque and S. Al Mamun


INTRODUCTION Flood is an annual event in Bangladesh. But unusual devastating floods like 1954, 1987, 1988 and 1998 are considered the worst ones on record, which caused widespread sufferings and loss of lives. The 1998 flood was unprecedented both in terms of magnitude and duration. The flood of 1998 being more serious in nature due to its longer duration than other past floods might have caused enormous impact on the major bridge structures of the country and morphological changes at the vicinity of the bridges. This has posed a threat to the sustainability of the bridges. Therefore, a study was undertaken to evaluate the impact of the 1998 flood on major bridges. A total of eleven bridges have been studied (Hoque et al., 1999), from which two bridges - the Meghna bridge over the river Meghna and the Turag bridge over the river Turag near Amin bazaar, Dhaka have been studied in detail. This paper presents the results of this detailed study. The location map of the bridges covered in this study is shown in Fig. 1. METHODOLOGY A detailed field survey around the Meghna Bridge has been carried out. Detailed cross-sectional survey and bed form measurement at 10km upstream and 2km downstream of the bridge have been made using the electronic distance meter (EDM) and echo sounder. The local scours at the piers, revetment and at another three locations, which are highly prone to erosion, have been evaluated by measurement and comparison with the previous records. The Turag-Bhakurta Bridge constructed by Local Government Engineering Department (LGED) on Turag-Bhakurta road over the Turag river has been studied in detail to compare the hydrological and morphological changes that had occurred during 1995 and 1998 floods. The relevant data were collected from LGED office, Dhaka. All these data have been graphically presented and a comparison has been made with the previous results to evaluate the impact of 1998 floods. THE MEGHNA BRIDGE The Megima Bridge is located on Dhaka-Chittagong Highway over the river Meghna about 25km south east of Dhaka. The location of the bridge is shown in Fig. 1. Dhaka-Chittagong Highway is the nation's 'lifeline' and construction of the bridge has added new dimension to this 'lifeline' eliminating the time killing ferry service at Meghna Ferryghat. The bridge is a pre stressed concrete structure and has 10 piers with eleven spans. The total length of the bridge is 930 m and its

Impact of Flood on Morphology of Rivers around Bridges


construction was completed in 1991. Some parts of Dhaka-Chittagong Highway went under water during the 1998 flood. Discharge at Bhairab Bazaar in the Meghna river for 20 years return period is 19124 m3/s. Water surface slope of the river at bridge site in respect to Bhairab Bazaar is about 0.0000158 (MPO, 1991).

Figure 1: Location map of selected bridges



To evaluate the morphological changes at the vicinity of the bridge, a river reach of about 1km downstream of the bridge and 5km upstream has been considered. This reach has been studied by Japan Bangladesh Joint Study Project on Floods (Hoque et al., 1997). The selected river reach is shown in Fig. 2.

Figure 2: Location map of the study area



The 1998 pre-flood morphological conditions are available for this area from a previous study (Hoque and others, 1997) in terms of (1) bed topography, (2) local scouring along the bank and (3) local scouring around the bridge piers 7, 8 and 9 from Dhaka side. To evaluate the changes in the river channel processes that occurred due to the flood of 1998, a detail survey was conducted during October 1998 on the river cross-sections, bed levels, scour around the piers, scour along the left revetment at upstream and downstream of the bridge and scour along the bank. A comparison between situations in May (pre-flood) and in October (post flood) 1998 has been made to evaluate the impact of 1998 floods. River Bed Changes

With riverbed cross-sectional data surveyed during May 1998 (pre-flood) and October 1998 (post flood) river, bed contour maps have been developed as presented in Fig. 3 and Fig. 4, respectively. From these figures significant differences have been observed in several places showing both erosion and deposition occurring during the 1998 flood.

Figure 3: River bed during May 1998 (before the 1998 flood)



Figure 4: River bed during October 1998 (after the 1998 flood) At section from L10 to R10, significant erosions have taken place on the left

side showing depth around 20m near the bank. A comparison of Fig. 3 and Fig. 4 reveals that at right bank some deposition has taken place during the 1998 floods. Downstream to this section on the right bank, the highest color shows further deposition where the depth is close to only Sm. However, further downstream to this section at the middle of the river where a sand bar existed before, significant erosion has taken place on the right side of the sand bar and the depth has reached to about 20m. This erosion has extended up to a reach of about 2km. On the left side of this section (left bank) further deposition has taken place. Further downstream up to the bridge and to the end of the reach, both sides show no change between the pre-flood and post-flood conditions. In general a significant change was found to have occurred in the bed level of the river Meghna upstream of the bridge during the flood of 1998.



Local Scour at Bank

Three locations in the reach in question have been identified with local scour at the bank and have been surveyed during October 1998. This survey has been made in addition to the general survey of cross-sections conducted for developing bed level contour as discussed above. This localized detail survey has been done between R10 and R8 on the right bank, between L4 and L3 on the left bank and along the revetment, upstream and downstream of the bridge on the left side. The bank in this section is convex in nature and the comparison between the Fig. 3 and Fig. 4 shows that during pre-flood time (May 1998), the depth in this section was deeper than post-flood time (October 1998). River bed condition along the right bank between section R10-R8 (Hoque et al., 1999) shows that the depth at this section along the bank varies from 2 to 22m, which may be regarded as one of the deepest sections in the reach considered for the present study. The scour conditions along the left bank between L4 and L3 (Hoque et al., 1999) shows that the deeper part of this section is along the bank, which confirms the results in Fig. 3 and Fig. 4. So scouring is progressing at this section and the maximum depth of scour is more than 20m but mostly 1 8m along the major part of the section. The bank is also convex at this section, which increases the near-bank flow velocity causing transport of sediments. At the downstream of the Meghna Bridge, the maximum depth of scour is about 22 m, which is almost same as in February 1997 (Hoque et al., 1997). For upstream, this scour depth is about 25 m at several places which is about the same as in February 1997. This result also confirms that there is no significance difference in bed level between pre-flood and post-flood conditions. But in comparison with the results of February 1997, several scour holes have been observed in October 1998. Local Scour Around the Piers

Detailed survey of bed level around the piers 7, 8 and 9 of the bridge was conducted during October 1998 to see the changes that occurred during the 1998 flood around these piers. Similar measurements were made during February 1997 and the results have been reported by Hoque et al. (1997). Conditions of the local scour in 1998 at piers 7, 8 and 9 are presented in Hoque et al. (1999). The maximum depth of scour holes observed is about 20m, which is less than the depth observed in February 1997 (Hoque et al., 1997). This may be due to the impact of the sand bar at upstream of the bridge, which has extended further down close to the bridge piers during the floods. THE TURAG-BHAKURTA BRIDGE

Local Government Engineering Department (LGED) constructed the Turag-Bhakurta Bridge on Turag-Bhakurta road over the Turag river in 1994 at Savar,



Dhaka. The bridge connects a number of villages e.g., Bhakurta, Kamrangirchar, and Keranigonj with Dhaka. The construction of the bridge together with other structures on Turag-Bhakurta road has made communication easy towards capital Dhaka for the people of these areas especially during rainy season. The bridge consists of 5 spans with length 67m each and width 3.70m. After the construction of the bridge, about 1km portion of Turag-Bhakurta road was raised and upgraded. The raised portion of the road used to go under water every year during flood and floodwater flowed freely over it during flood season. Following the rise of road level, the bridge came under heavy water pressure during the 1995 and 1998 floods due to obstruction of free flow of floodwater. During 1995 flood, the first pier of the bridge from Dhaka-Aricha side settled down by about 1.61 m (as seen in Fig. 5). After recession of flood, the bridge was rehabilitated by providing additional piles with length of up to 30m at the affected pier. During the 1998 flood the bridge again came under flood attack and was washed away except for the pier rehabilitated after the 1995 flood (as seen in Fig. 6).

Figure 5: Settlement of Pier-1 of Turag-Bhakurta bridge on Turag- Bhakurta road near Amin Bazar close to Dhaka-Aricha road during the

1995 flood (Date: 25-02-95)



Figure 6: Washed away Turag-Bhakurta bridge on Turag-Bhakurta road near Amin Bazar close to Dhaka-Aricha road at Savar during the 1998 flood

(Date: 27-10-98) Morphological Changes

The cross-section along the central line of the bridge was measured after the 1995 flood by the local office of LGED at Savar. After the 1998 flood, cross-section was again measured along the central line of the bridge. These data were collected from LGED office, Dhaka to evaluate the morphological changes that had occurred due to floods of 1995 and 1998.

Information about water level, velocity etc. were also recorded during the cross section measurement in the year 1995 (September 24). Water level at the damaged bridge site was 6.50m during cross-section measurement and flood level was 7.88m during the 1995 flood. The surface velocity of water through the bridge section was about 2.OOm/s, which was measured by LGED. Measurements of cross sections at inflow and outflow points were carried out during survey work. Cross-section of outflow (damaged bridge) section is shown in Fig. 7. Figure 7 also shows the comparison of cross section before and after 1995 and 1998 floods at damaged bridge section. The water surface slope during the 1995 flood was quite high, as high land of Bhakurta village made spill water of the Dhaleswari river to make a substantial head at inflow section. Moreover, the catchment is a low-lying area bounded by highland and road network.



Discharge at damaged bridge section during the 1995 flood was about 1047 m3/s while during the 1998 flood it was about 1370 m3/s. For the passage of 1995 flood discharge through damaged bridge section, opening length should have been about 154m. But actual length of ridge was 67m. Hence, substantial morphological changes or scour had occurred at affected bridge section during the 1995 flood reaching a scour depth of (-) 6.60m as shown in Figure 7. For the passage of discharge through the bridge section during the 1998 flood, opening length should be about 176m. So significant morphological changes had occurred during the 1998 flood reaching a scour depth of (-) 13.40m as shown in Fig. 7. As a result, the bridge was washed away. Causes of Failures of the Bridge During 1998 Flood

Design elevation of pile end at bed of the bridge was (-) 6.75m. It is evident from Fig. 7 that scour depth at pier-1 during the 1995 flood reached almost the pile length. Maximum scour was about (-) 6.60m below bed elevation. As a result, the pier got settled down from the effect of load of deck slab and girder. But during the 1998 flood, scour depth reached the extent of about two times the pile length of the bridge. Maximum scour is about (-) 13 .40m below bed elevation (Fig. 7). Hence, total bridge had been washed away except for the portion of the bridge rehabilitated after the 1995 flood. The pile length provided for the rehabilitated pier is about 30m. During the 1995 flood the outflow discharge at the damaged bridge section was about 1047 m3/s. During the 1998 flood the outflow discharge at bridge section was about 1370 m3/s. As a result, enormous pressure was exerted on the structure. For the passage of the discharge of the 1995 and the 1998 floods through the bridge section, the adequate opening length should have been 154m and 176m, respectively. But the bridge length is only 67m. Hence, due to enormous thrust of floodwater heavy morphological change occurred during flood period. The failure of the bridge is mainly due to a severe morphological change. Flood level in 1998 was 9.70m and the bridge was partially submerged. So, enormous hydraulic pressure was exerted on the bridge during the flood of 1998. As a result, the bridge was washed away.

CONCLUSIONS

The major floods, like the one of 1998, have significant impact on the river morphology at the vicinity of the bridges. The impact on a smaller bridge is more severe than that on a larger bridge. Significant changes took place in the river channel process due to the 1998 floods at the vicinity of the Meghna Bridge. Heavy scouring occurred under the Turag-Bhakurtia bridge during 1998 flood.



Figure 7: Cross-section at damaged bridge section before and after the 1995

and 1998 floods



REFERENCES MPO (1991), Surface Water Hydrology Report, MPO, Dhaka. Hoque M.M. et al. (1997), Study of Morphological Behaviours of the river

Meghna, Final Report of Japan Bangladesh Joint Study Project, Topic 3, IFCDR, BUET.

Hoque, M. M., Bala, S. K., Ahmed, S. M. U., Haque, A. and Al-Mamun, S. (1999), Impact of 1998 Floods on the Morphology of the Rivers Around Bridges, BUET.


Impact of the 1998 Flood on Groundwater Recharge in Dhaka

M. Mozzammel Hoque, Syed Mohib Uddin Ahmed and Md. S. A. Hossain

Institute of Water and Flood Management Bangladesh University of Engineering and Technology, Dhaka-1000, Bangladesh Abstract Floods have severe adverse impacts on agriculture, communication network, river channel process, and public health. Despite its adverse impact, floods have some positive impact also and one of the major positive impacts of floods is the groundwater recharge. The impact of 1998 flood on the recharge of Dhaka City aquifer has been evaluated. The Old Dhaka city close to the river Buriganga received a reasonable recharge during 1998 floods. The Sutrapur area received a net recharge of about 1.5m in 1998 as observed in the well at Jagannath College compound. The Lalbag area received a net recharge of about 0.9m during the 1998 flood. The central part of the city received insignificant net recharge due to paved area and high withdrawal. This includes Motijheel commercial area, which is fully paved and also distant from the river. The low lying unpaved areas on the periphery of the city where both the vertical and the horizontal components of recharge have played important role, received significant net recharge during the flood period. The maximum net recharge to this area during 1998 is about 6.6m. In summary, it may be concluded that the 1998 flood has made significant contribution to the recharge to the aquifer of Dhaka city except to the central part of the city.

INTRODUCTION


Due to its geographical location, each year about 18% of Bangladesh is flooded. During severe flood the affected area may exceed 36% of the country and almost 68% of the net cultivable area. The catastrophic flood of 1987 and 1988

M. M. Hqoue, S. M. U. Ahmed and Md. S. A. Hossain


submerged more than 50% of the country causing huge economic loss. The flood of 1998 was more severe due to its long duration and this flood caused enormous damages to the infrastructure, industry, agriculture, and human health. In spite of many adverse effects of floods, in some areas, floods bring positive impacts. One of such positive impacts is the groundwater recharge (Hoque, et al., 1999; Hoque and Shahabuddin, 1998; Shahabuddin, 1996). In Bangladesh, recharge occurs primarily through direct infiltration and percolation, mostly from the huge amount of rainfall and floodwater during the period from June to September. During the dry season, groundwater becomes a major source of domestic, industrial, and agricultural water supply. But in many areas of the country the groundwater level is declining gradually posing a threat to the availability of water during the dry season. Therefore, recharge to the groundwater is considered an important phenomenon in water resources system. The city of Dhaka is growing very fast. The present water supply system of Dhaka almost entirely depends on groundwater. As surface water bodies near the City are becoming increasingly polluted and costly to purify, public water utilities and other urban water users are turning to groundwater as potential source of supply. But exploitation of groundwater has its limit and depends on how much water is replenished during the monsoon. The aquifer of Dhaka city is recharged by direct rainfall, river water, and floods (MPO, 1987). The current study evaluates the extent of recharge due to floods, especially the flood of 1998 which had a long duration compared to other past floods.

STUDY AREA The study area, shown in Fig. 1, is bounded by the Buriganga river to the south and west and Lakhya river to the east. The convergence of Turag and Balu rivers limits the western, northern and part of eastern boundaries. The metropolitan area lies approximately between 23o40’ and 23o53’ North Latitude and between 90o20’ and 90o31’ East longitude. The city is situated on flat plain land. The lowest land is': located in the Balu river and the highest land is located in the Mirpur area. The periphery of the city is low-lying in comparison to the central part. The core of the city falls within 6 to 8 m contours with reference to mean sea level. Dhaka city is expanding rapidly through urban and industrial development. The rapidly increasing paved area is affecting the recharge considerably through the change of runoff length, evapotranspiration, etc. In densely built up areas, the natural recharge is significantly reduced. However, urbanization may produce other form of recharge such as leakage from water distribution and sewerage system. The Buriganga, Balu, Turag, Tongi Khal and the Lakhya are the rivers surrounding the greater Dhaka city. The surface water system of Dhaka, comprising several depression storage (roads, lakes and

Impact of Flood on Groundwater Recharge


submerged low-lying lands) and khals (channels), is linked to these surrounding rivers. The city rainfall run off is accumulated in the depression storage and is discharged to the surrounding rivers through the khals.

Figure 1: Location of the study area



METHODOLOGY The study is based on the secondary data. The groundwater level and the river water level data have been collected from the relevant organizations. For groundwater observation data, several observations wells maintained by the Groundwater Circle of Bangladesh Water Development Board (BWDB) have been selected. The river water level data for the same period from 4 stations, one on each river, have also been used for analysis. The groundwater observations wells and the river water level gauges selected for this study are shown in Fig. 2. The groundwater observation wells under the influence of different rivers have been grouped together as listed in Table 1. The analysis has been done mostly graphically. A comparison has been made between the impacts of 1998 floods and that of the other big floods. Table 1: River water level stations and groundwater observation wells with locations

Water level Rivers Wells

Ordinate Name Measuring Station

No. Well No. Place Latitude

(N) Longitude

(E) Tongi Khal

Tongi 299 DA-103 Maniknagar, Cantonment

23°58'05" 90°38'10"

DA-70 Cheragali Market, Tongi

23°42'00" 90°24'45"

Turag Mirpur 302 DA-108 Mohammadpur 23°45'00" 90°22'55" Balu Demra 7.5 DA-A12 Banani \WDB,

Gulshan 23°51'58" 90°19'26"

DA-112 Malibagh, Motijheel

23°43'35" 90°23'35"

DA-123 South Basabo, Motijheel

23°44'05" 90°25'30"

DA-124 South Khilgaon, Motijheel

23°45'00" 90°25'15"

Buriganga Millbarak 42 DA-13 Jagannath College, Sutrapur

23°42'00" 90°24'45"

DA-A13 BUET, Lalbagh 23°45'00" 90°17'04" DA-111 Charakghata,

Mohammadpur 23°45'20" 90°21'25"



Figure 2: Groundwater observation wells and the river water level gauges



RESULTS AND DISCUSSIONS

Fig. 3 shows a variation of groundwater level at DA-13, an observation well located at Jaganath College compound under Sutrapur thana and the river water level in the river Buriganga at Milbarak (Gauge Station 42) for the flood years of 1995 and 1998. In July 1998, before the beginning of the flood, the water level at the well was much lower (about 0.6 m) than that of 1995. For both the years, the groundwater levels continued to rise until late October and reached the same level (about 1.2 m) from the base flow. It is observed that at the beginning of flood seasons, the groundwater level of 1995 was about 0.6 m higher than that of the 1998, but finally they reached the same level in the middle of October. So, during the period from July to late October the groundwater level was replenished by about 0.6m in 1995 and by 1.2m in 1998. The river water levels of 1995 and 1998 as shown in Fig. 3, indicate that the prolonged and higher flood water level of 1998 caused the higher recharge in 1998 than 1995. The lag time between the peak flood and the peak groundwater level of 1995 is about 10 weeks, but the lag time for 1998 is about 6 weeks. The recession part shows that the falling rate of flood level for 1995 is slightly higher than that of 1998. In old Dhaka city, the paved area remained more or less unchanged between the years 1995 and 1998. The other impacts such as vertical recharge due to rainfall may be considered constant. So, the lateral recharge from the river has made a major contribution to 1998 recharge.

Figure 3: Responses of groundwater level at Well DA-13 located at Jagannath College to the water level at Millbarak of Buriganga River



Figure 4 shows the variation in groundwater level at the well DA-A 13, located at BUBT campus under Lalbag thana for the year 1987 and 1998 and the corresponding water level in the river Buriganga at Gauge Station 42, located at Millbarak. The rise in water level started from the middle of July in both the year from the same level and at the beginning the rise in 1987 is faster than that of 1998, although the water table reached the same level during middle of October. The total recharge during the flood seasons of 1987 and 1998 is almost same (about 1 m). But the river water level in 1998 is significantly higher than that of 1987. The recession curve at the beginning is steeper in 1998 than in 1987. The lag time between the highest peak in groundwater level in observation well DA-A13 and the river water level at Millbarak is about 31 days in 1998 and about 45 days in 1987. In 1987 this longer lag time is possibly due to a rise in river water level during middle of September and a very slow recession of river water level in comparison to 1998. The flood of 1998 has contributed to the recharge of groundwater at BUET campus area to some extent but the total replenishment is not enough to raise the water to the level of the previous year. This is mostly a paved area, and there is no surrounding water body. Therefore, whatever replenishment has taken place is possibly due to the lateral recharge from the river.

Figure 4: Responses of groundwater level at Well DA-A13 located at BUET Campus to the water level at Millbarak of Buriganga River



Figure 5 shows the variation in groundwater levels at observation well DA-103, located at Maniknagar, Cantonment for the flood years 1995 and 1998 and the corresponding flood water levels at Gauge Station 302 in the river Turag at Mirpur. For the year 1995 the groundwater level continued to rise until the middle of September but in 1998 the groundwater level continued to rise until the end of September. At the beginning of the flood season, during the third week of July, the groundwater level in 1998 was almost 1 m lower than that of 1995. At the end of replenishment, the total rise in groundwater level is 2.0 m in 1998 and about 0.5 m in 1995. So, the total replenishment in 1998 is 1.5 m (300%) higher than that of 1995. When compared the flood levels in the river Turag at Station 302, the flood level was found to be increasing until the middle of September in 1998, but in 1995 there were three peaks each lower than those during 1998. Therefore, the constant rise (duration) and higher peak have caused the higher recharge in the year 1998. The lag between the peak in river water level and the groundwater level is almost 2 weeks in 1998. This area is characterized by a rapid increase in population as well as paved area due to heavy construction. The declination rate is very high as the vertical recharge area is decreasing and the withdrawal rate is increasing. Therefore, it can be concluded that the 1998 flood has made a significant contribution to the groundwater recharge in this area due to its high magnitude especially due to its long duration.

Figure 5: Responses of groundwater level at Well DA-103 located at

Maniknagar, Cantonment to the water level of Turag River at Mirpur



Figure 6 shows the rise of groundwater table in well DA-70, located at Charagali Market, Tongi, Gazipur District, for the flood years 1988, 1995 and 1998 and the corresponding water levels of Tongi Khal at Tongi (Gauge Station 299). The figure shows that during April - May the water level is almost at same level in three years, which are very similar to the river water level. The river water level has significant influence on the groundwater level. At the beginning of flood season (mid-July) the river water level is at the same level in 1998, 1995 and 1988, but the groundwater level is lower in the year 1998 compared to the other two years. For 1988, the water level continued to rise until September and then receded sharply followed by a rise again until the middle of November. This is possibly due to the second peak in the river level. But in 1995 the groundwater level continued to be at same level from early September until late October. This is possibly due to water level rise and fall, and rise again. However, in case of 1998, the groundwater level continued to rise sharply until the second week of September and then receded continuously unlike other two years. The recharge rate in 1998 is slightly higher than the other two years, which has similarity with the flood level in the river Tongi Khal at Tongi. It is observed that the lag time between peak water level in the river and the peak groundwater level in the well is significantly shorter than the other wells, which are located in the metropolitan areas of the Dhaka city. It is interesting to see that the rate of rise of groundwater level in this area is almost the same as the rise in the river water level. This is possibly due to the fact that this well is located in such a place where the man-made intervention is less than that of the city area and the aquifer is hydraulically well connected with the river. Therefore, the response of the aquifer to the river level is very fast and the peaks at the river level and groundwater level were attained on the same day. The vertical and lateral recharge rates are significantly high due to favorable conditions. Therefore, the flood of 1998 had tremendous impact on the recharge to the aquifer where both the vertical and lateral recharge is possible. Therefore, it can be concluded that the vertical recharge also plays a vital role to recharge the aquifer system and the long duration of flood has significant contribution to the recharge.

Figure 7 shows the groundwater level variations for the flood years 1995 and 1998 at observation well DA-l 11, located at Charakghata, Mohammadpur with corresponding river water level at Millbarak (Gauge Station 42) in the river Buriganga. In 1998 the groundwater level started rising from April and continued the trend of rise until peak was achieved. But in 1995, the groundwater level started rising in the middle of May and had two peaks, one in August and the second one in November. At the beginning of flood period (first week of June) the groundwater level in 1995 was higher than that of 1998 by almost 1.5 m. This indicates that the rate of rise in 1995 is higher than the rate of rise in 1998.



Figure 6: Responses of groundwater level at Well DA-70 located at Cheragali Market to the water level at Tongi Khal

Figure 7: Responses of groundwater level at Well DA-111 located at

Charghata, Mohammadpur to water level at Millbarak of Buriganga River



The river water level in 1995 was also seen slightly higher than the river water level of 1998. But the difference between the water level of 1995 and 1998 is much lower than the difference between groundwater rise between 1995 and 1998. So, there is a significant contribution to the recharge from other sources in the year 1995. But in the year 1998 major contributions came from the river as the figure indicates that there is a continuous rise in the groundwater level until October with similar pattern to the rise of river water level in 1998. The influence of river water level to the recharge in 1998 is less than that of 1995 and this may be due to the presence of embankment during the 1998 flood. The lag time between the peak water level in the river Buriganga at Millbarak and the peak level of groundwater is about two weeks. This area is very close to the river Buriganga and Turag, but despite its location, the withdrawal rate is more than the rate of replenishment by flood. This is due to significant increase in paved area with the construction of buildings and other infrastructures. Thus as a result of the combined effect of a decrease in area of vertical recharge and the increase in withdrawal, the flood of 1998 could not show a higher recharge compared to the previous years. However, the figure shows that the recovery is more during the flood of 1998 in comparison to the recovery of 1995.

Figure 8 shows the groundwater levels recorded at observation well DA-A12, located at Banani under Gulshan thana for the flood years 1987 and 1998 with the corresponding river water level of the river Balu at Demra (Gauge Station 7.5). For the year 1987 the groundwater level continued to rise until the middle of October, but in 1998 the water level continued to rise until the end of October. In 1987 the response of the groundwater rise to the river water level is found to be faster than that of 1998. The lag time in 1987 is almost 5 weeks whereas the lag time in 1998 is more than 5 weeks. Total replenishment in 1987 is about 3.5 m whereas in 1998 total replenishment is about 1 m. So, the recharge conditions in 1987 were much better than those of 1998. This may be due to the increase in the rate of withdrawal and decrease in the recharge area due to the increase in pavement area. The pavement area decreases the rate of vertical recharge and the vertical recharge plays a significant role to the aquifer recharge.

Figure 9 shows groundwater levels variation with time at observation well DA-124, located at south Khilgaon, Motijheel for the flood years 1988, 1995 and 1998 with corresponding river water levels at Demra (Gauge Station 7.5) on the river Balu. It is seen that the river water level at different years is almost the same, but there is a big difference in groundwater level at different years. The groundwater is falling with time at a very high rate. The influence of the river water level on the groundwater level is decreasing gradually and the influence of other factors responsible for replenishment of groundwater level is decreasing as well. In the year 1988, the difference between the highest groundwater level and the lowest groundwater is only about 2 m and in other two years the difference is



even smaller. In this well the groundwater table is constantly decreasing and therefore it is not possible to determine the influence of 'the peak river water on the groundwater table. This indicates that in this area there is a major mining of groundwater level due to excessive withdrawal and that possibly the recharge is very insignificant compared to the withdrawal.

Figure 8: Responses of groundwater level at Well DA-A12 located at Banani to the water level of Balu River at Demra

Figure 10 shows the groundwater variation with time at observation well

DA-l12, located at Malibag, Motijheel for the years 1988, 1995 and 1998 with corresponding river water levels on the river Balu at Demra, Gauge Station 7.5. The figure shows that there is a significant decrease in water level with time and the flood level has insignificant influence on the replenishment of the groundwater table. Possibly during flood season the influence of river water level is compensated by the overdraft and as a result there is no significant resultant increase in the groundwater table. Similar to the well DA-124, the groundwater table in well DA-112 is constantly decreasing and it was not possible to identify the influence of the river water level in this well in terms of depth or lag time. Therefore, it can be concluded that in the Motijheel area, recharge is insignificant even during the very big floods, like the one of 1998.



Figure 9: Responses of groundwater level at Well DA-124 located at South Khilgaon to the water level of Balu River at Demra

Figure 10: Responses of groundwater level at Well DA-112 located at Malibagh to the water level of Balu River at Demra



CONCLUSIONS

In general the flood of 1998 has contributed significantly to the recharge of the Dhaka city aquifer. The impact of 1998 flood to the recharge is observed more at the periphery of the City where the area is relatively low, unpaved and close to the river. In densely populated and fully paved centrally located areas, the recharge from 1998 flood is insignificant. The net recharge is significantly decreasing with increasing urbanization due to an increase in paved area. The impact of flood on recharge is a function of the distance from the rivers, unpaved/paved area and rate of withdrawal. Even the major flood like one of 1998 could not protect the groundwater level from mining or declination with time at the central part of the city. REFERENCES Hoque. M. M. and M. Shahabuddin (1998), An Evaluation of Groundwater

Conditions in Dhaka City, Hydrology in a Changing Environment, Proceedings of the British Hydrological Society International Conference, Exeter, July 1998, Vol.2, pp 175-184.

Hoque, M. M. and Others (1999), Impact of Flood 1998 on Dhaka City Aquifer Recharge, Institute of Flood Control and Drainage Research, Bangladesh University of Engineering & Technology, Dhaka.

MPO (1987), Groundwater Resource Evaluation of Bangladesh, Technical Report No.5, Master Plan Organization, Dhaka.

Shahabuddin, M. (1996), An Evaluation of Dhaka City Groundwater Conditions, Institute of Flo6d Control and Drainage Recharge, Bangladesh University of Engineering and Technology, Dhaka 91 pp.


Hydrological Characteristics of the 1998 Flood in Major Rivers

A. K. M. Saiful Islam and Jahir Uddin Chowdhury

Institute of Water and Flood Management Bangladesh University of Engineering and Technology, Dhaka-1000, Bangladesh Abstract The 1998 flood in the Jamuna, the Padma and the lower Meghna rivers was an extreme event of more than 100-year return period in terms of flow duration. The duration of flood in the Ganges was relatively short, but the return period of peak discharge is greater than 100 years. A numerical hydrodynamic model was used to generate hydrographs of flood discharge in major rivers. The flood flow in 1998 remained above riverbank for nearly 1, 3 and 4 months in the Ganges, the Jamuna and the Padma, respectively. The average flood discharges in these rivers during July to September of 1998 were around 150% of the long-term average discharge for that period. An analysis of inflow and outflow as a function of time indicates that the floodplains stored about 10% of the flow during 1998 flood, and as a result the peak discharge diminished considerably as the flood wave moved downstream. The stage-discharge relationships for flood waves in the Jamuna and the Ganges displayed significant loop characteristics. INTRODUCTION The flood of 1998 is a rare hydrological event in the history of Bangladesh. It caused colossal damage to the socio economy and extreme suffering to the people. The main cause of the flood was spill from the major rivers, particularly the Jamuna and the Ganges, which carry about 85% of the flood flow that enter Bangladesh. This paper investigates some hydrological characteristics of flood flow in the major rivers namely the Jamuna, the Ganges, the Padma, the Meghna




and the Lower Meghna by analysing the data generated by a numerical hydrodynamic model and making comparison with previous large floods. MAJOR RIVERS The flood regimen in Bangladesh is dominated by huge flow carried by three major rivers, namely the Brahmaputra, the Ganges and the Meghna (Fig. 1). The source of floodwater in and around Dhaka city is the Brahmaputra through its distributories. The Brahmaputra has a length of about 2,900 km of which only 270 km lies in Bangladesh, and the reach between offtake of the Old Brahmaputra and confluence with the Ganges is called Jamuna. The catchment of the Brahmaputra at Bahadurabad is approximately 573,500 sq.km. The Ganges has a length of about 2,200 km and the reach in Bangladesh is approximately 230 km long before meeting the Jamuna, and the catchment area at Hardinge Bridge is approximately 1,090,00 sq. km. The travel times of flood waves from the border with India upto the Ganges-Jamuna confluence are approximately 5 and 7 days in the Ganges and the Jamuna, respectively. The reach carrying the combined flow of the Ganges and the Jamuna is called Padma and has a length of about 120 km before meeting the Meghna.

The other major river Meghna, has a length of about 110 km between Bhairabbazar and confluence with Padma. The Meghna receives water flow mainly from 2 big rivers (Fig. 1) and the meeting point of these rivers is approximately 20 km upstream of Bhairabbazar. The catchment area at Bhairabbazar is approximately 77,000 sq.km. The reach carrying the combined flow of Padma and Meghna is called Lower Meghna, which has a length of about 160 km and discharges to the Bay of Bengal. Almost entire flood flow in Bangladesh is drained through the Lower Meghna, which is a tidal river and has a large estuary. ANALYSIS STATIONS Stations where the Bangladesh Water Development Board (BWDB) gauges both water level and discharge in the major rivers are at Bahadurabad in the Jamuna, Hardinge Bridge in the Ganges, Bururia in the Padma and Bhairabbazar in the Meghna (Fig. 1). There is no permanent discharge gauging station in the Lower Meghna. There is a water level gage station at Chandpur, which is approximately 22 km downstream from the confluence of the Padma and the Meghna (Fig. 1). This is the most downstream location in the gage network for flood forecasting. Present study utilizes flood data for these stations.

Hydrological Characteristics of Flood in Major Rivers

Figure 1: River system of Bangladesh

SIMULATION OF FLOOD DISCHARGE Gauged discharge data at Bahadurabad, Hardinge Bridge, Baruria and Bhairab bazar are usually available at about two weeks interval during flood season. For the Lower Meghna, discharge data is not available. A numerical model developed by Chowdhury (1986) was used to generate hydrograph of flood discharge in the major river systems. The model is based on an implicit finite difference solution of the gradually varied unsteady flow equation shown below.

0=−∂∂

+∂∂ q

xQ

thW

0||2

2

=+∂∂

+⎟⎟⎠

⎞⎜⎜⎝

⎛∂∂

+∂∂

kQQgA

xhgA

AQ

xtQ




where, W is the total water surface width (m) , h is the elevation of water surface with respect to a common datum, Q is the flow rate (m3/s), A is the cross-sectional area of flow (conveyance) section (m2), q is the lateral flow rate per unit length of channel, RCAk = , k is the conveyance of flow section in m3/s, R is the hydraulic radius of flow section (m), C is the Chezy resistance coefficient (m1/2/s) and g is the acceleration due to gravity (m2/s). The effective cross-sectional area of flow (conveyance) section is estimated by

rffr YYAAA /+=

where, Ar and Af are water areas for river section and floodplain section respectively ,and Yr and Yf are water depths in river and floodplain respectively.

Hydrodynamic condition during flood flow in the alluvial rivers in Bangladesh is quite different from that during low flow. The cross section changes due to erosion when flow increases from low to high discharge and due to deposition when flood flow recedes. The resistance characteristics also change when low flow changes to high flow in the river –floodplain system. Such changes in the hydraulic condition create difficulty for simulation by numerical model where fixed cross section is assumed. To reduce these difficulties, the simulation was kept confined to the three months period of July to September when about two-thirds of annual discharge occurs.

Schematic representation of the rivers that were included in the model is shown in Fig. 2. It included 1483 km of channels involving 11 rivers, 13 junctions, 4 upstream discharge boundaries and 3 downstream water level boundaries. The rivers were divided into distance steps varying from 1 to 30 km, and the total number of distance steps was 240. A time step of 30 minutes was used in the simulation. The model was calibrated against observed data of previous large flood in 1988 and verified against observed 1998 flood data. Calibrated values of C were in the range 92 to 99 m1/2/s among the rivers. A comparison of predicted hydrographs with observed discharge data during 1998 flood is shown in Fig. 3. FLOOD MAGNITUDE AND DURATION Simulated hydrographs of 1998 flood at Hardinge Bridge, Bahadurabad, Baruria and Chandpur are shown in Fig. 4. The bankfull discharges (after Delft Hydraulics and Others, 1996) are also shown in the figure. Figure 4 indicates that the long duration of 1998 flood in Bangladesh was mainly due to continuous inflow of high discharge for more than two and half months in the Jamuna. There were four major flood waves in the Jamuna while one in the Ganges. The


maximum peak flood discharges at Jamuna and Ganges occurred quite closely at the beginning of September 1998. This feature caused unprecedented high discharge in the Padma and Lower Meghna.

Figure 2: Schematization of river system used in the numerical model

Table 1 shows comparison of three-month (July-September) average

discharge and annual maximum discharge during 1998, 1988 and long-term. It is seen from Table 1 that the average flood discharge at all stations except Bahirabbazar during 1998 was much higher than that during 1988, and around 150% of the long-term average discharge. Data in Table 1 indicates that approximately 55% of total flood discharge during July to September 1998 was carried by the Jamuna while approximately 34% by the Ganges.




Figure 3: Comparison of predicted discharge with observed discharge during the 1998 flood

Figure 4: Predicted hydrographs of flood discharge in major rivers during

the 1998 flood


Table 1: Comparison of three-monthsaverage flood discharge (July to September) at gauge stations of major rivers

Parameter Year Hardinge Bridge

Bahadurabad Baruria Bhairab Bazar

Chandpur

1998 42,808 67,100 1,04,177 12,601 1,06,665 1988 37,844 52,358 82,499 15,513 81,458

Average flood Discharge (m3/s)

28 yrs. Avg.

30,764 43,025 65,900 11,060

1998 80,330 93,658 1,36,481 14,037 1,38,283 1988 71,800 98,300 1,32,000 17,900 1,19,464

Peak flood Discharge (m3/s) 28 yrs.

Avg. 52,358 67,100 89,742 13,946

Duration of 1998 flood is compared with previous large floods as well as

long-term average values. By analyzing data since 1966, the previous floods having large magnitude or long duration have been identified. The duration has been assessed in terms of flow above mean bank level (MBL) and bank-full discharge (BFD). Values of MBL and BFD were determined by Delft Hydraulics (1996). Results for flood level analysis are summarized in Table 2 while discharge analysis is presented in Table 3. It is seen from Tables 2 and 3 that peak flood level and discharge of 1998 flood exceeded previous records for the Ganges and the Lower Meghna but not for the Jamuna and the Meghna. However magnitudes of peak floods in Jamuna and Meghna rivers are quite close to the previous records. Using the results of frequency analysis in IFCDR (1995), the estimated return period of the peak of 1998 flood are 100 years and much greater than 100 years for annual maximum water level at Bahadurabad (Jamuna) and Hardinge Bridge (Ganges), respectively; and 75 years and greater than 100 years for annual maximum discharge. Table 2: Flood level and duration in major rivers

Maximum flood level in specified years (PWD)

(m)

Continuous duration of flood in days above

MBL

Flood of longest

continuous duration

above MBL before 1998

River Gauge Station

Mean Bank Level MBL

(PWD) (m)

Year 1998

Year 1988

Year 1987

28 yrs. Avg

Year 1998

28 yrs. Avg

Year Duration in days

Jamuna Bahdura-bad

18.93 20.41 20.61 19.68 19.74 87 23 1977 59

Ganges Hardinge Bridge

14.01 15.19 14.87 14.79 14.28 28 12 1980

42

Padma Baruria 7.05 10.74 9.8 9.5 8.89 118 84 1990 119




Table 3: Flood discharge and duration in major rivers

Maximum discharge in specified years (103 m3/s)

Continuous duration of

flood in days above

BFD

Flood of longest continuous

duration above BFD before

1998

River Gauge Station

Bank- full dis-charge (BFD) (103

m3/s)

Year 1998

Year 1988

Year 1987

28 yrs.

Avg.

Year 1998

28 yrs.

Avg.

Year Duration in days

Jamuna Bahdura-bad

48.0 93.6 98.3 73.0 67.1 77 21 1966 46

Ganges Hardinge Bridge

43.0 80.3 71.8 75.8 52.3 34 14 1980 40

Padma Baruria 75.0

136.5 132.0 113.0 89.7 115 21 1980 57

The most remarkable feature of 1998 flood is its very long duration as can be

seen from Tables 2 and 3. The long duration of high discharge during 1998 flood in the Jamuna, Padma and Lower Meghna is unprecedented. The flow duration graphs for 1998 flood at Bahadurabad and Chandpur are superimposed on the stage-duration-frequency graphs given in IFCDR (1995) as shown in Fig. 5. The 1998 flood level at Bahadurabad corresponding to duration of 30 days or more has a return period of greater than 100 years. At Chandpur, the return period of both peak magnitude and duration for 1998 flood has exceeded 100 years. STORAGE FUNCTION OF FLOODPLAIN Simulation results show that the flood waves attenuated substantially in a quite short distance as it travelled down the Jamuna river as can be seen from the Fig. 6, where the hydrographs are at two locations of approximately 92 km apart. Large decrease in the peak discharge while increase in the trough discharge indicates that substantial volume of floodwater was retained in the floodplain and river. The floodwater was stored during rising flood level and released during falling flood level. To investigate the retention function of floodplain, an analysis of inflow and outflow using the data generated by the numerical model was made. The analysis was based on integration of the following equation.

dtdSOI =−

where, I is the inflow (m3/s), O is the outflow (m3/s) and S is the storage (m3).


Figure 5: Comparison of daily water level of 1998 flood with stage-duration-frequency curves at Bahdurabad and Chandpur




Figure 6: Simulated flood discharges during 1998 at two locations 92 km

apart along Jamuna River Daily average discharge was used in the integration process and the effect of

travel time was incorporated. The estimated storage, as a percentage of cumulative inflow volume since the first of July, is plotted against time in Fig. 7. Water level at Baruria, which is just at the downstream of Ganges-Jamuna confluence, is also shown. It is seen that around 20% and 10 % of total inflow of floodwater were stored in the river-floodplain system during 1988 and 1998 floods, respectively. Smaller percentage for storage in the case of 1998 flood is because of much larger inflow volume. Larger percentage for storage during early part of 1988 flood season was because of smaller inflow. This analysis indicates that the floodplain plays the role of a huge retention reservoir. Therefore complete prevention of flooding of major floodplains by flood control embankments is likely to cause substantial increase in the flood levels in major rivers. STAGE-DISCHARGE CHARACTERISTICS Using model-generated data for 1998 flood, stage-discharge plots for Jamuna, Ganges and Padma rivers at Bahadurabad, Hardinge Bridge and Baruria,


respectively are shown in Fig. 8. It is seen that there is an anti-clockwise loop in the relationship corresponding to every flood wave. Discharge is larger during rising flood than that during falling flood for a given stage. The loops are distinct and quite wide in the case of Ganges and Jamuna rivers, but not distinct in the case of Padma River. This feature indicates that monotonic stage-discharge relationship (rating curve) may not be appropriate for the Jamuna and the Ganges. Effect of acceleration in flow, resistance from braid-bars and storage function of floodplain are among the main reasons for significant loop characteristic in the stage-discharge relationships.

Figure 7: Estimated storage of floodwater in the river-floodplain system and water level at Baruria

CONCLUSIONS The remarkable feature of 1998 flood was the long duration of very high discharge in the Jamuna, Padma and Lower Meghna. The flood flow remained above the riverbank for nearly 1, 3 and 4 months in the Ganges, the Jamuna and the Padma, respectively. Return periods for the 1998 flood in the Jamuna, Padma and Lower Meghna are greater than 100 years when the flow duration is considered. Close occurrence of peak floods in the Ganges and Jamuna during early September caused very large discharge of around 140,000 m3/s in the Padma and Lower Meghna.




Figure 8: Simulated stage-discharge relationships of flood waves in 1998

for Jamuna, Ganges and Padma rivers


The floodplains played an important role by storing about 10% of the total inflow of flood water through rivers during 1998 flood season, and as a result the peak flood discharge decreased substantially as the flood waves moved downstream. The stage-discharge relationships for flood waves in the Jamuna and the Ganges display significant loop characteristics. REFERENCES Chowdhury, J.U. (1986), An Implicit Numerical Model of Unsteady Flow in

River Network, R 01/86, IFCDR, BUET, Dhaka. Delft Hydraulics, DHI and Others (1996), Floodplain Levels and Bankfull

Discharge, Special Report No.6, FAP24, River Survey Project, WARPO, Dhaka.

IFCDR (1995), Flood Frequency Analysis: Component of the Study on Revision of Flood Danger Levels in Bangladesh, (Prepared by Chowdhury, J.U., Rahman, R. and Salehin, M.), UNDP project of Hydrology Directorate, BWDB, Dhaka.



Effects of Coastal Phenomena on the 1998 Flood

Anisul Haque, Mashfiqus Salehin and Jahir Uddin Chowdhury

Institute of Water and Flood Management Bangladesh University of Engineering and Technology, Dhaka-1000, Bangladesh Abstract The water level during the 1998 flood was unusually high in the lower Meghna, which is the single outlet for the drainage of flood flow in Bangladesh. Using data generated by a finite element model in two-dimensional space as well as observed tide level data, effects of some coastal phenomena on the flood flow have been studied. The unusual high water level in the lower Meghna was mainly due to obstruction to the very large flood discharge by the unexpectedly large spring tide flow from the opposite direction. The monsoon wind can cause a maximum wind set-up of around 25 cm, which may reduce the velocity of flow by about 29%. The study does not indicate presence of earthquake-generated tsunamis in the coastal region during the 1998 flood. INTRODUCTION There was unusually high water level in the Lower Meghna during the 1998 flood. It was remarked in the bulletin of the Flood Forecasting and Warning Center of Bangladesh Water Development Board (BWDB) that there was abnormal tide. A total of 14 tidal gage stations (Fig.1) were considered to investigate the abnormal tidal behavior, if any, which might have been responsible for the unusually high water level at Lower Meghna. Two stations belong to the Mongla Port Authority while the rest to the Bangladesh Inland Water Transport Authority (BIWTA). Besides 1998, data of 1992 and 1988 floods were also used. The flood in 1992 was an average flood while the 1988




flood was second biggest after 1998 flood with respect to magnitude and duration. The expected tidal levels calculated by harmonic analysis were used from the Tide Tables of BIWTA for comparing with observed data. Some results are presented here.

In the months of May, June and August 1998, several earthquakes occurred in the Bay of Bengal close to the Nicobar Island of India (Web Page, USGS), the most severe one occurring near the island (7.329N, 94.277E) on the 10th of August at around 3:50 pm BST with a magnitude of 5.8 in the Richter Scale. The present study investigated if that particular earthquake generated any tsunami and/or had any influence along the coastline of Bangladesh using a two-dimensional hydrodynamic model of water circulation in the Bay of Bengal.

Monsoon wind speed in the Bay of Bengal varies between 5 to 15 m/s. When this monsoon wind blows over the sea surface, apart from the creation of waves (superimposed on the tidal waves), the shear stress at the sea surface causes a rise or surge in water level. Amplitude of surges will only be small in deep water but will become magnified on account of the shoaling effect on entering shallow water. The two-dimensional finite element model has been applied also to study the monsoon wind set-up along the coast of Bangladesh.

Figure 1: Locations of tidal gage stations

Effect of Coastal Phenomenon on Flood


INVESTIGATION OF THE TIDAL BEHAVIOR

Tidal Characteristics

The outlet channels that carry the upland fresh water flow to the Bay of Bengal include Biskhali, Buriswar, Lohalia, Tetulia, Shahbazpur and Hatia channels (Fig.1). The major distributary system includes the Tetulia, the Shahbazpur and the Hatia channels; the Shahbazpur channel being the main flow-carrying river. Tide along the Bangladesh coast originates in the Indian Ocean and approaches the coast of Bangladesh approximately from the south, arriving at the Hiron point and at Cox’s Bazar at about the same time. The tide along the coast is semi-diurnal having an average period of 12 hours and 25 minutes. However, funnel shaped coastal geometry and uneven bottom topography results in distortion of tides at some places. The tide has some daily inequality in high water levels varying between 0.0 m and 0.6 m. Spring-neap tide cycle is about a fortnight and during this cycle the tidal period varies because of the phase inequality. There is considerable variation from neap to spring tides. There is also seasonal variation of the mean sea level that has a direct causal effect on the tidal range. Was There Abnormal Tide?

The water level at Lower Meghna was unusually high (much more than 100-year flood). The question is whether it was due to any abnormal behavior of tides as suggested by the Flood Forecasting and Warning Center of BWDB. A simplified approach was adopted which involved study of hydraulic characteristics along two spatial directions (Fig.2): (i) section 1-3-2 along the coastal belt, and (ii) section 3-4-5 from the river mouth towards upstream in the estuary. Any suspected abnormality in tidal behavior in the estuary must have manifestation of the same along the coastline.

Figure 2: Simplified schematization of tidal network



Figures 3 and 4 show the observed tidal levels along with the expected tidal levels obtained by harmonic analysis along the two selected sections for the month of August of 1998. Evidently, the observed tidal levels matched very well with the expected levels along the coastal belt (Fig.3), whereas the difference between the two got increasingly pronounced upward in the estuary (Fig.4). Similar phenomenon was observed in the months of July and September, and also in the monsoon months of 1992 and 1988 (Haque et al., 1999). The possibility of the presence of any abnormal behavior upstream causing extremely high water level seems to be remote.

The unusually high water level in the Lower Meghna resulted from the interaction between flood waves and spring-neap tidal cycle. The entire flood flow from the upstream drains to the Bay of Bengal through the single outlet of the Lower Meghna. On the other hand, there is an inflow of huge volume of tidal water from opposite direction during rising tide from the Bay of Bengal. The Lower Meghna is the meeting point of upland fresh water flow and tidal flow (Fig.1). During 1998 monsoon, four large flood waves passed Bangladesh in succession, each one coming before the river level could recede. As a result, there was continuous flow of large volume of water for more than two months to the Lower Meghna. This flow was obstructed by the very large spring tides from the opposite direction that resulted in building up of a very large water depth. In other years, the duration of flood wave was short and there was time for recession of flow after a flood or between two flood waves. INVESTIGATION OF TSUNAMI GENERATION

Physics Behind the Formation of Tsunami

'Tsunami' is a series of waves of extremely long wavelength and long period generated in a body of water by a disturbance, primarily the earthquakes in oceanic and coastal regions. As the tsunami crosses the deep ocean, its length from crest to crest may be hundred kilometers or more, and its height from crest to trough only a few meter or less. They cannot be felt aboard ships nor they can be seen from the air in the open ocean. Upon entering the shoaling water of coastline, the wave velocity diminishes and the wave height increases resulting in building up of a large tsunami exceeding 30 m in height strikes with devastating force.

In the deepest oceans, the tsunami behaves as shallow water waves because of long wave length traveling in speed exceeding several hundred kilometers per hour. The water depth to wave length ratio gets very small and waves move at a speed equal to the square root of the product of the acceleration of gravity and water depth. For example, if the typical water depth is 4000 m, a tsunami travels over a speed of 700 km/hr. A tsunami will lose little energy as it propagates, as the rate at which a wave loses its energy is inversely related to its wavelength. Hence in very



deep ocean, a tsunami will travel great distances at high speeds with limited energy loss. As it propagates into the shallower water near the coast, its speed diminishes as the depth of water decreases, but the change of total energy of the tsunami remains constant meaning that the height of the wave grows. Because of this, a tsunami may appear as a series of rapidly rising or falling waves. The terminal wave height will depend on the travel path of the waves, the coastal configuration and the offshore topography. Numerical Simulation

A numerical model has been developed to study whether the earthquake which occurred near the Nicobar Island (7.329N, 94.277E) at around 3:50pm BST on the 10th of August 1998 with a magnitude of 5.8 generated any tsunami along the coastline of Bangladesh. The model developed is a two-dimensional horizontal plane finite element model where the basic equations are the conservation of mass and momentum in two space dimensions. They are:

0 = yV H+

xU H+

t ∂∂

∂∂

∂ζ∂

H - )uv(-

y + )uu(-

x +

xg- =

yUV +

xUU +

tU bx

ρτ

∂∂

∂∂

∂ζ∂

∂∂

∂∂

∂∂

H - )vv(-

y + )uv(-

x +

yg- =

yVV +

xVU +

tV by

ρτ

∂∂

∂∂

∂ζ∂

∂∂

∂∂

∂∂

In the above equations, velocity )V,U( , Reynolds stresses )vv ,uv ,uu( are depth averaged quantities, τbx and τby are the bottom shear stresses, ζ is the surface elevation above the datum, H is the total water depth and ρ is the water density. The pressure gradient term has been evaluated using the hydrostatic pressure assumption. Bottom shear stress in the second and third equations arises due to vertical integration of Reynolds shear stress terms. These stresses have been evaluated by assuming velocity profile to be logarithmic and using a uniform flow approximation. Unknown correlations appearing in the second and third equations are closed using second order closure level of turbulence. Finite elements are used for the spatial discretization and a three level semi-implicit time scheme has been used for the time discretization.

The model has been applied in a region of the Bay of Bengal spanning more than 3000 km in east-western and more than 2000 km in north-southern directions. In the schematized model domain western boundary covers coastlines of Srilanka and India, northern boundary covers coastlines of India and Bangladesh, eastern boundary covers coastlines of Myanmar, Thailand and Malayasia, and the southern boundary covers coastline of Indonesia and open sea.



Figure 3: Observed and expected tidal levels along the coastal belt (section 1-3-2) in August 1998



Figure 4: Observed and expected tidal levels from the coast towards upstream along the esturary (section 3-4-5) in August 1998



Model simulations for the sea surface elevation on August 10, 1998 at 4:00pm, 4:30 pm, 5:00 pm and 9:00 pm (approximately 10 minutes, 40 minutes, one hour and 5 hours after the occurrence of earthquake) were generated (Haque et al., 1999). It was seen that after 10 minutes, the tsunami wave reached the coast of Andaman Island from the north-west direction, the coast of Thailand from the eastern direction and the coast of Indonesia from the southern direction. It did not propagate much in the western direction because of the presence of Andaman Island. After 40 minutes (at 4:30pm), the wave propagation continued, but it did not amplify significantly near the Coast of Nicobar Island, Andaman Island, Thailand and Indonesia. The possibility of the wave reaching near the coast of Bangladesh at this time seems to be a rare possibility, and approximately after 1 hour it could be completely ruled out. Approximately after 5 hours (around 9:00pm), the sea surface came to the original position. Vector plot of tsunami propagation direction at 4:00 pm, 4:30pm, 5:00 pm and 9:00 pm (Haque et al., 1999) indicates that the tsunamis started propagating in almost all directions from its origin 10 minutes after the earthquake. After 40 minutes and approximately 1 hour, the tsunamis were never directed towards the coast of Bangladesh. Approximately after 5 hours, the waves were completely dissipated. Alternative Scenario

Historical evidence showed that earthquake also occurred in the deep sea near the Andaman Island (Webpage, NOAA). Model has been applied to study the probable tsunami generation for this case assuming the epicenter to be near the Andaman Island (10.67N, 93.19E), a location close to which most of the earthquakes occurred in the past (Webpage, USGS). Probable sea surface elevations for a hypothetical large earthquake for this situation were evaluated and the vector plots of tsunami propagation direction were plotted (Haque et al., 1999). It was seen that within 10 minutes of the occurrence of earthquake, waves are initially reflected at the Coast of Andaman Island and propagate towards the south-westerly and north-westerly directions. Unlike the previous case (epicentre close to the Nicober Island), the tsunami did not reach the coast of Indonesia and Thailand. Within 40 minutes of its generation the wave traveled a long distance mainly in the deep ocean, but still remained far away from the coast of Bangladesh. After approximately 1 hour, the tsunami with an insignificant magnitude reached the west coast of Srilanka and India, but the coast of Bangladesh was completely free from any attack. After approximately 5 hours from its generation, waves were dissipated and the sea surface came to its original position. Measured data

Tidal gage station records, if there is any tsunami, should generally indicate an upward swell followed by a rapid drawdown. After that there should be a series of



waves with a period varying from a minute to an hour (Webpage, USGS). A comparison of measured water levels and expected tidal levels (from the Tide Tables of BIWTA) from August 9-12, 1998, at six tidal gage stations along the coast (Hiron Point, Galachipa, Char Changa, Sandwip, Sadar Ghat and Cox's Bazar), situated from west to east, showed that the measured data qualitatively do not show any abnormality compared to the expected behavior (Haque et al., 1999). Quantitatively, the measured values do not show any rapid swell or drawdown or any sign of series of waves indicative of a tsunami. EFFECT OF MONSOON WIND

Model for Wind Set-up

Due to monsoon wind, seawater movement is basically horizontal. In this case, the previously used two-dimensional finite element model in a horizontal plane has been applied. The wind-generated waves were not incorporated in the model. Only the wind induced shear stress was considered. The wind influence was assumed constant over the flow field. A time series of wind velocity components could be incorporated in case of wind varying with time. The transfer of momentum of wind energy to water body is achieved as a shear stress acting on the water surface and is calculated using a square law.

The schematized model domain covers an area consisting of the western coast of India, the coastline of Bangladesh and eastern coast of Myanmar. East-west and north-south boundaries of the model span about 1000 km and 500 km, respectively. The model domain has been discretized into linear elements. Tidal variation of water level has been specified at the southern ocean boundary. These tidal elevations are calculated from the tidal constants for Baruva (India) and Searl Point (Myanmar). The constants were taken from the Admiralty Tide Tables. In intermediate locations, a linear interpolation for the tidal constituents has been made. Simulated Sea Surface Elevation

Model simulations of sea surface elevations for different wind speeds when southern sea boundary is at LW and HW have been developed (Haque et al., 1999). As expected the LW and HW at the sea boundary are not propagated instantly towards the coast. When wind blows over the sea surface, sea level starts rising, particularly near the coast. Although the effective wind stress is the same on deep and shallow water, the funneling shape and shallowness of the Bay near the coast causes sea level to rise particularly near the coast than the deep ocean. The effect is more pronounced with the increase of wind speed. Rise of sea level due to monsoon



wind set-up is particularly significant near the coast of Bangladesh and partly near the Indian coast. But this effect is nearly absent along the coast of Myanmar.

Longitudinal profiles of water level along the Bangladesh coast for different wind speed when sea boundary is at HW and LW show that as the stronger winds blow over the sea, water level rises along the coast. Table 1 shows numerical values of this water level rise. On average, a monsoon wind speed of 10 m/s over the sea raises the water level along the coast by 11cm. The average sea level may rise upto 25 cm, 46 cm and 74 cm if the monsoon winds blow with a speed of 15 m/s, 20 m/s and 25 m/s, respectively. It may be recalled here that monsoon wind speed in the Bay of Bengal varies between 5 to 15 m/s. So this monsoon set-up may result a rise of sea level along the Bangladesh coast upto a maximum 25 cm above the normal astronomical tide that may reduce the velocity of flow by about 29%. Table 1: Water level Rise along the coast of Bangladesh when sea boundary is at HW and LW

Distance from Water Level Rise (m) due to wind speed of Hiron Point 10 m/s 15 m/s 20 m/s 25 m/s

(km) HW LW HW LW HW LW HW LW 0 0.08 0.07 0.19 0.16 0.35 0.31 0.59 0.50 57 0.09 0.08 0.21 0.20 0.40 0.37 0.65 0.59 99 0.10 0.09 0.24 0.22 0.44 0.41 0.71 0.65 148 0.11 0.11 0.26 0.26 0.47 0.48 0.76 0.76 180 0.11 0.12 0.25 0.28 0.46 0.51 0.74 0.81 211 0.11 0.13 0.25 0.30 0.46 0.54 0.73 0.85 245 0.12 0.14 0.27 0.33 0.49 0.60 0.77 0.94 285 0.12 0.16 0.28 0.36 0.50 0.66 0.79 1.03

Average 0.11 0.25 0.46 0.74 CONCLUSIONS The unusually high water level in the Lower Meghna during the 1998 flood resulted from the continuous flow of large volume of water for more than two months coupled with the obstruction by spring tides from the Bay of Bengal. The earthquakes that occurred near the Nicobar Island in the Indian Ocean in 1998 did not cause any tsunami along the Bangladesh coast. Even had there been any earthquake near Andaman Island, the tsunami would not have reached the Bangladesh coast. The monsoon wind set-up may result a rise of sea level along the Bangladesh coast upto a maximum 25 cm above the normal astronomical tide that may reduce the velocity of flow by about 29%.



REFERENCES Haque, A., Salehin, M. and Chowdhury, J. U. (1999), Effect of Coastal

Phenomenon on 1998 Flood, BUET


Performance Evaluation of FCD/FCDI Projects During the 1998 Flood

A.F.M. Saleh, S.M.U. Ahmed, M.R. Rahman, M. Salehin, and M.S. Mondal

Institute of Water and Flood Management Bangladesh University of Engineering and Technology, Dhaka-1000, Bangladesh

and M. Mirjahan

Department of Water Resources Engineering Bangladesh University of Engineering and Technology, Dhaka-1000, Bangladesh

Abstract The performances of six flood control and drainage projects in mitigating the damage from the flood of 1998 have been evaluated. The magnitude of the 1998 flood exceeded the respective design flood levels of Dhaka-Narayanganj-Demra (D-N-D) project, Meghna Dhonagoda Irrigation Project (MDIP) and Narayanganj-Narsingdi (N-N) Project. But the projects were able to withstand the flood because of the freeboard and preventive measures taken by BWDB. In Compartmentalization Pilot Project (CPP), the 1998 flood was also higher than the design flood but the embankment was breached at three locations and cut at one location by outsiders. In Nagor River and Sakunia Beel projects the flood levels were below the design flood levels. But, because of cuts by outsiders, most of the Nagor River project was inundated. Except MDIP and Sakunia Beel projects, all the other projects could not achieve their desired targets of protecting the Aman crop. In D-N-D and N-N projects, the Aman crop was partially damaged due to drainage congestion. Because of flooding in Nagor River project, the Aman crop was totally damaged. In CPP, the compartmentalization concept has not worked and inadequate drainage together with failure by the sub-compartments to store floodwater have resulted in reduction in Aman area by about 50%. But, in spite of the damages inflicted, preliminary analyses on costs of flood fighting and returns from the Aman harvest in MDIP, CPP, D-N-D and N-N projects show that the benefits overwhelmingly outweigh the costs incurred, even without considering the intangible benefits.


A.F.M. Saleh, S.M.U. Ahmed, M. Mirjahan, M.R. Rahman, M. Salehin and M.S. Mondal


INTRODUCTION The flood of 1998 has been a catastrophic flood in terms of recurrence interval, flooded area and duration, especially in the Ganges and lower Brahmaputra basins. Because of the resulting colossal damage and destruction, it has raised a number of issues regarding the planning, design and performance of infrastructure, especially the Flood Control and Drainage (FCD)/Flood Control, Drainage and Irrigation (FCDI) projects. Bangladesh Water Development Board (BWDB), to protect the Aman crop from flooding and to enhance the agricultural productivity, implemented these projects. Over the years, BWDB has constructed more than 400 FCD/FCDI projects covering about 3.7 million hectare, which is about 60% of the total flood vulnerable area and about 40% of the net cultivable area of the country (Chowdhury et al., 1996). It has been estimated that because of the damages to Aman crop due to 1998 flood, the harvest has been reduced by about 2 million tons.

The FCD/FCDI projects are normally designed to withstand the 20-year flood with a freeboard of 0.9 m (BWDB, 1996). For most of the major and medium rivers of Bangladesh, the differences in peak water level between the average flood and the 100-year flood are in the order of 1-1.5 m. The difference between 20-year and 100-year flood is less than 1 m (Kruger and BCEOM, 1992; Chowdhury et al., 1996). Thus, even though the FCD/FCDI projects are designed for 20-year flood, with the additional freeboard of about 1 m, these projects should theoretically be able to withstand the 100-year flood without being overtopped. Whether the embankments of these projects are actually capable of withstanding the 100-year flood depends a lot upon how well have they been maintained. The specific objective of this study was to evaluate the magnitude of the 1998 flood in the context of the FCD/FCDI projects and to assess the performances of these projects in mitigating the flood damage, especially to Aman crop. The details of the findings of the study are given in Saleh et al., (1998). METHODOLOGY There are lots of variations in the more than 400 FCD/FCDI projects implemented by BWDB, in terms of the hydrologic setting, type, planning and design criteria, size and age of the projects. It is therefore difficult, if not impossible, to select a few projects that would represent the diversity. But, because of limited time and financial resources to complete the study, only six projects were selected for performance evaluation. The two criteria that dominated the selection process were that the projects must be located in the

Performance Evaluation of FCD/FCDI Projects

severely flooded zone in 1998 and should have available past reports. Representation of all the major hydrologic zones, planning and design criteria, size, age and funding agency were also considered. The six selected projects were: (1) Compartmentalization Pilot Project (CPP), Tangail; (2) Dhaka-Narayanganj-Demra (D-N-D) Project, Dhaka; (3) Meghna-Dhonagoda Irrigation Project (MDIP), Chandpur; (4) Nagor River Project, Bogra; (5) Narayanganj-Narsingdi (N-N) Project, Narayanganj; (6) Sakunia Beel Project, Faridpur. The salient features are given in Table 1.

The methodology followed for the performance evaluation comprised of two phases. In the first phase, all existing secondary information (reports and data) of the relevant projects were collected. The information were then analysed and interpreted for developing a framework of parameters (both qualitative and quantitative) to be used in the evaluation. The developed framework was used during the field visits for assessing the physical condition, design appropriateness and effectiveness of the infrastructure. In addition, the design criteria of the infrastructure were also reviewed to check their adequacy in mitigating losses from a flood of the magnitude of 1998 flood. For analysing the magnitude of 1998 flood, all relevant hydrologic data for the selected projects were also collected from BWDB.

Table 1: Salient features of the selected projects

Name of Project Hydrologic Zone

Type Area (ha) Funding Agency

Year of Completion

CPP Central FCD 13,200 FAP-20 1995

D-N-D Central FCDI 8,340 World Bank 1968 Meghna-Dhonagoda SE FCDI 17,580 ADB 1987 Nagor River NW FCD 15,400 EIP 1986 N-N Central FCDI 3,000 JICA 1993 Sakunia Beel SW FCD 5,700 GoB 1985

The field visits to each of the selected projects were undertaken in the second phase and the existing condition of the infrastructure (flood control and drainage) was assessed using the developed framework of parameters. Qualitative assessment of the performance of the project and flood damage (if any), was made by interviewing a wide range of informants, visiting the affected/damaged areas and reaching informed judgements in the field. The information collected from the field were cross checked with the data available with the BWDB project officials and in case of any contradiction the information collected by the team prevailed.




ANALYSIS OF THE 1998 FLOOD From the analyses of flood magnitude it was evident that the 1998 flood was of moderate magnitude in the middle Brahmaputra-Jamuna basin. For Nagor River project and CPP, the 1998 flood magnitude was lower than the previous highest flood (1988 flood) with flood return periods of 10-year and 27-year, respectively. In the lower Brahmaputra and lower Meghna basins, the 1998 flood was catastrophic and exceeded all previous records, both in terms of water level height and in duration. For N-N and D-N-D projects of lower Brahmaputra basin, and MDIP of lower Meghna basin, the 1998 flood could well be classified as the 100-year flood. Even though the 1998 flood was also catastrophic in the Ganges basin, the Sakunia Beel project was not seriously affected as this has now become a compartment of the greater Faridpur-Barisal project and is protected by the Ganges Right Embankment. Because of the regulated flow of the Kumar River, the 1998 flood was like a normal flood for the Sakunia Beel project. The 1998 flood return period and the durations above the danger/design levels for the studied projects are given in Table 2. Table 2: Return period and duration of 1998 flood

Name of Project

Design Return Period (Yrs)

1998 Flood Return

Period (Yrs)

Duration Above Danger/Design

Level (days) CPP Not Available 100 44 D-N-D 20 27 62 Meghna-Dhonagoda 100 > 100 64 Nagor River 20 10 N/A N-N 25 100 33 Sakunia Beel 20 Not Applicable Not Applicable

PROJECT PERFORMANCE Compartmentalization Pilot Project

Flood Control

Contrary to the concept of the traditional FCD projects, where river flooding is both unacceptable and undesirable, the Compartmentalization Pilot Project (CPP), Tangail, was designed to serve the purposes of flood control, controlled or beneficial flooding and flood storage in pre-defined sub-compartments to reduce flood damages elsewhere. Controlled flooding within the compartment would be achieved mainly through the Main Regulator on Lohajang River, and that within


the sub-compartments by the sub-compartment water management committees through other peripheral inlets and water control structures. By lowering the water level of the Lohajang river using the Main Regulator, the river would be allowed to act as an outlet in which the sub-compartments would drain through their numerous internal regulators.

Unlike normal flood years when the Main Regulator along with other inlets are kept open in the pre-monsoon season, the unusual rainfall in May 1998 (103% higher than average) at Tangail forced the Main Regulator to be closed with the outlets open to facilitate drainage. As the water level started rising, the Main Regulator along with other inlets were opened on 1 June to ‘control flooding’ so as to allow the entry of fingerlings into the compartment and siltation on the field. The Main Regulator was partially closed on 9 July in 1998 to effect ‘flood control’ by restraining the inside water level from surpassing the allowable level of 11 m PWD. The water level upstream of the inlet exceeded the danger level of 12.04 m on 15 July and remained over the danger level for 62 days. The level was as high as 13.39 m PWD on 8 September that corresponded to a 27-year flood, only 4 cm lower than the previous highest flood level of 1988. The main regulator was totally closed on 26 August following a breach in the peripheral embankment at Rasulpur on 25 August, in order to lower the water level of Lohajang river and to facilitate ‘drainage’ of flood water.

The water level difference inside and outside CPP during the flood was more than 2 m threatening a major failure of CPP embankment. Although as per the original concept, the embankment was to be overtopped in case of a more than 20-year flood with all the inlet structures open, this was not allowed to happen by the project people. The freeboard of 0.3 m beyond the 1988 flood level did not allow overtopping of the embankment, but at some locations (approximately a total of length of 9 km) overtopping was prevented by raising the crest level with sand bags.

Much effort was given to prevent the CPP area from flooding by taking protective measures along its peripheral embankment, and by operating its inlet structures carefully. But, despite all efforts made by the sub-compartment water management committees, individuals and project officials, some of the areas of the CPP suffered from flooding due to breaching at three locations (Rasulpur, Passbetur and Indrabelta) and public cut at one location (Pauli) along the peripheral embankment. The public cut and the breaches except the one at Rasulpur were closed in two days, while the breach at Rasulpur that occurred on 25 August could not be closed due to high head difference. About 200 m of the embankment was badly damaged by erosion of Pungli River at Birnahali during the flood. Thus, although designed against a 20-year flood, the project was partially successful in mitigating damages from 27-year flood of 1998.




Drainage

The unusual rain in May and July (103% and 68% above average, respectively) caused severe inundation and restrained seedbed preparation, destroyed seedlings and delayed transplantation of T. Aman. Because of the breaches and the public cuts, there was flooding and crop damage in sub-compartments 6, 7 and 8. There was backwater effect starting from 9 September at the southern part of the project that could not be controlled because of the absence of any such provisions, and hence it caused considerable flooding and crop damage in sub-compartments 13, 14 and 15. Thus, ‘compartmentalization concept’ did not work during the 1998 flood.

The sub-compartmental embankments being porous also resulted in drainage congestion during the heavy rainfall of May and July that caused crop damage. With adequate control on water in watertight sub-compartments, damages due to rainfall, breaching and backwater flooding could have been averted. Dhaka-Narayanganj-Demra (D-N-D) Project

Flood Control

The elevation of the top of the flood wall (where it exists) of the D-N-D project varied from 7.47 m to 7.62 m PWD and the highest water level during the 1998 flood at the project’s pumping station was 6.49 m. So the flood wall was not overtopped at any place of the project and was always at least about one meter above the highest flood level at all points. The railway embankment top between I.T. School and Chashara Railway Station where there is no flood wall was set at about 6.8 m to 7.4 m PWD after the 1988 flood. The highest water level recorded at Narayanganj was 6.92 m PWD. Though the maximum difference between the highest water level at Narayanganj and the embankment top was 12 cm, the embankment was overtopped at a number of points by water depths of up to 65 cm. This might have happened due to inadequate maintenance or settlement during the last 10 years. Several slope failures due to seepage, piping, leakage, sliding and overtopping were observed during the field visit.

It can therefore be concluded that the embankment of the project having floodwall has been successful and effective in protecting the area from the 1998 flood. The problems encountered in this part of the embankment were mainly due to 18 drainage pipes of WASA and one old culvert at Ranimahal. These buried pipes and old culvert should immediately be removed or be adequately sealed up for the safety of the embankment. The embankment between I.T. School and Chashara is a very weak railway-cum-flood protection embankment and its performance during the flood was very poor. The crest level, crest width and side slopes should be corrected and revised for the safety of the project.


Drainage

The D-N-D project has provisions for pump drainage but the four pumps of the project were out of operation from 4 September to 14 September, 1998, for a total of 11 days because of rise in river side water level beyond the pump operation limit (6.2 m). The drainage situation was further deteriorated due to increase in drainage volume through rainfall, leakage, seepage, piping, overtopping, industrial and domestic wastes etc.. It was assessed that in about half of the gross project area the crop loss due to drainage congestion was at least 40%. In fact, in about 1915 ha (40% of the gross land), which are located below the elevation 2.64 m PWD, and where the submergence was more than 15 days, the damage to Aman crop was total. Farmers also corroborated these findings during the field visit in October 1998. Meghna-Dhonagoda-Irrigation Project

Flood Control

The MDIP was successful in withstanding the 1998 flood, which was unprecedented in the history of the project, both in terms of magnitude and duration. The highest1998 flood water level of 5.64 m at Chandpur on Meghna corresponded to a more than 100-year flood. The previous highest level of 5.08 m in 1988 corresponded to around 30-year flood. The water level was above the danger level for 88 days and above the 100-year flood level for 6 days. So it was the freeboard that prevented the embankment from being overtopped.

The project started experiencing multifarious problems since the second week of July when flood water first exceeded the danger level. There were four flood peaks, on 29 July, 10 August, 25 August and 9 September. Each one came before the river level could recede, and so there was continuous flow of large volume of water for over two months, causing huge damages to the embankment. About 250 seepage holes, 320 boiling points, 83 piping points, 3610 m of sliding of country side (c/s) and 2525 m sliding of river side (r/s) slopes of the embankment were reported by BWDB.

The main reason for such damage may well be attributed to persistently long high water level beyond the design level, head difference between the inside and outside water levels, weak sub-soil properties of the embankment, and inadequate embankment design (less than design crest width). Field visit revealed that most parts of the embankment were constructed with sandy soils that were vulnerable to piping and seepage problems. The portion resectioned (12 km) with proper width and r/s slope suffered little, while the old portion suffered most, especially in areas adjacent to launch terminal and where there were borrow pits on the c/s. Besides, the embankment was endangered by thousands of trees along the embankment. Wave action and wind caused the trees to get dislodged leading to piping and seepage at many places.




The protective measures undertaken by BWDB during the flood were: (i) piling and filling of slopes by sand bags to prevent sliding (3380 m); (ii) share key to prevent sliding (at 151 locations); (iii) ring filter for protection against boiling (at 99 locations); (iv) control of erosion due to wave action by packing water hyacinth (13 km); and (v) feeding of water to the irrigation canal to minimize the head difference.

As the Meghna near MDIP carries the combined flow of both the Meghna and the Padma (the Ganges and the Brahmaputra), riverbank erosion is a major problem. The setback distance designed considering 100 years of river erosion has proved to be inadequate as the present erosion rate is much higher than the rate estimated at the feasibility stage. In the western part of the Project (Mohanpur and Dashani), river bank erosion assumed a serious turn because of 1998 flood. Around 3 km of such embankment along Meghna was under threat, so was around 1.5-2 km of the Dhonagoda side in the south-eastern part (Gazipur, Shibpur, Amirabad and Torki). Drainage

The MDIP has provisions for pump drainage and there was very little drainage congestion problem inside the project area during the flood, as reported by BWDB. The water level inside the project was kept slightly higher than the design level by pumping water from the rivers to the irrigation canals with the purpose of minimizing the head difference between the inside and outside of the project. Nagor River Project

The return period of 1998 flood in Nagor River was about 10 years and it was not an exceptionally high flood both in terms of magnitude and duration. The average flood water level remained about 1 meter below the embankment top and embankment was not overtopped at any place during the 1998 flood. Overtopping had rarely been observed since the implementation of the project in 1986. However, since its completion, Nagor River Project has regularly experienced flooding due to public cuts on its embankment made by the inhabitants of adjacent Nagor Valley Project. During the flood, the neighbouring Nagor Valley Project (on the right bank of Nagor river and just opposite to the Nagor River Project) was submerged due to heavy rainfall and overflow of floodwater from the Raktadoha-Lohachura Project. Consequently, the Nagor Valley Project inhabitants cut the embankments on both sides of the Nagor river so that water finds a direct route to south-eastern lower floodplain through the Nagor River Project. This phenomenon of public cut has become a routine event. Public cuts were made during 1987 and 1988 floods and after reviewing the


prevalent situation, FPCO (1991) considered the project as a total failure. The project has been rehabilitated in 1997/98 as per “Redesign Project” undertaken in 1993/94. Nagor Valley project has also been rehabilitated in the mean time. But the situation remains unchanged. Public cuts were made in 1996 and 1997. The 1998 flood year was no exception. Public cuts were made very early in the flood season inundating the entire project area. During the field visit in November 1998, six public cuts were seen.

Field level discussion with Nagor Valley Project inhabitants revealed that they made the public cuts to save their homesteads from flood damage. They said that given the hydrological situations of the area they have no other alternatives and they will continue to do so in future under similar circumstances. Inhabitants of Nagor River Project patrolled their embankment regularly to deter the public cuts but of no avail.

According to the Nagor River Project inhabitants, the damage to their Aman crop would have been much lower in absence of any embankment at all. Before the project, water level used to rise slowly and B.Aman could grow gradually with the floodwater. Because of the public cuts, the damage to the Aman crops is total as the water level increases suddenly in the project area and the water velocity remains very high.

In Atrai basin, series of projects have been implemented since late seventies without basin-wide planning. As a result, problems like waterlogging, confinement, back flow and backwater effects, siltation, public cutting etc. have become commonplace. Among the projects, Nagor River Project being situated in the downstream end is the worst affected. Given the hydrological situation in the region, modification of the entire project into a submersible one with adequate post-monsoon drainage facilities so that the area remains under free flow condition during monsoon, appears to be the only viable option. BWDB project officials and local people also supported this option during the field visit. Under this option, damage to Aman crop will reduce and the gain in Boro production will be retained. Narayanganj-Narsingdi Irrigation Project

Flood Control

The crest level of the embankment was set at 7.5 m PWD and the highest water level during the 1998 flood was at 7.12 m. Thus, the embankment was not overtopped at any place of the project and was always at least 30 cm below the crest level at all points. But, the persistently long high flood level beyond the designed level of 6.6 m for more than a month had its toll on the stability of the embankment. BWDB officials reported more than 300 major and minor leakage points along the embankment. But during field visit on 11-12 October 1998,




seven severe slope failures due to seepage, piping and leakage problems were observed.

Random sampling of cross section measurements at a number of points of the embankment showed that the side slopes of both the r/s and c/s have been adequate and maintained as per design. The crest width was never found to be 6 m (the designed value) and varied from 4.5 to 5.2 m. The embankment was heavily encroached, especially on the c/s slope by shops, homestead, trees etc.. The turf on the r/s has been damaged by flood but where they existed in the c/s, they were in good condition.

It can therefore be concluded that the embankment of the project has been successful and effective in protecting the area from 1998 flood even though the flood level was 0.5 m higher than the designed flood level. The problems encountered during the 1998 flood, which threatened the embankment and the project, were not because of inadequate design but because of improper construction. The borrow pits/ponds/depressions on the c/s adjacent to the toe of the embankment should be immediately filled up and the buried pipes/culverts (not belonging to the project) should be immediately removed or adequately sealed up for the safety of the embankment. Drainage

The designed pumping capacity of the project was adequate to meet the internal drainage demand of the 1998 flood. But, even then, there were significant damages to Aman crop during the 1998 flood due to severe drainage congestion. Because of interruption in power supply the pumps could not operate at the desired time and rate. The average daily power interruption during the flood season (from 1 July, 1998 to end of September) was 2.17 hours with a maximum of 8.9 hours. Even when there was power, the pumps could not be operated because of low voltage (the pumps require 400V for their operation). Moreover, higher than designed flood level at the r/s resulted in a total shut down of the pumping station for 24 days (from 25 August to 17 September) with disastrous effect on the standing Aman crop. As per the operation manual, the pumps are not to be operated if the water level in the r/s exceeds 6.8 m PWD.

From the analyses of water levels in the c/s and r/s at the pumping station during the 1998 flood and area-elevation curve of the project, it was evident that in more than one third of the total arable land of the project area the crop loss due to drainage congestion was at least 40% or more. In fact, 25% of the agricultural land, where the submergence was for more than 15 days, the damage to Aman crop was total. The farmers also corroborated these findings during the field visits.


Sakunia Beel Project

Flood Control

The Sakunia Beel Project, consisting of a number of small beels including Sakunia as the main one, is now considered as one of the several sub-projects of the large scale Faridpur-Barisal Project. The Ganges right embankment of the project now protects the Sakunia Beel Project along with other sub-projects; and the Madankhali regulator at the mouth of Kumar at Faridpur regulates the flow of water in the Kumar inside the project area.

Contrary to the unprecedented magnitude and duration of the 1998 flood elsewhere in the country, the project was almost completely free from flooding of Kumar River. The highest water level of the Ganges at Goalundo upstream of the Madankhali regulator was 10.19 m corresponding to a flood with well over 100-year return period and the water level remained above the danger level for 68 days. However, efficient operation of the Madankhali regulator ensured that the water level of Kumar at Faridpur was always maintained far below the danger level of 7.5 m. So there was no question of the Kumar river embankment being overtopped.

The flood control embankment along the Kumar performed well against the regulated flooding in the Kumar. However, the embankment suffered some damages to the r/s slopes at several locations due to the non-existence of proper setback distance. The FAP 12 study (FPCO, 1991) as well as field inspection revealed that almost no setback distance was provided for about 85% of embankment reach.

The BWDB did not take any emergency protective measures in the project as the extent of damage was not as threatening as that of other severely affected flood control projects. Their flood fighting activities were mainly concentrated along the large scale Ganges embankment as prevention of any breaches would automatically reduce the risk of any significant damage to the sub-projects of the Faridpur-Barisal Project. Drainage

The project experienced severe backwater flow in the Gabra khal on the southern part of the project, which overtopped the adjacent dwarf embankment (from Joyjhap to Gotti bridge, poorly constructed with earth spoil with no defined crest level) over most of its parts causing breaches at two locations. Consequently, about 20% area in this extreme southern part of the project was inundated, and drainage congestion prevailed over one month as reported by the local people. The backwater flow seems to have resulted from the unusual high water in the lower reaches of Kumar river due to serious bank erosion by flood water that washed away a 5 km reach of the Ganges embankment near Char Harirampur and Char Bhadrashan Unions.




Drainage in the greater upper part of the area was reasonably effective, if not very efficient. Defective lifting device of one gate of the Mridhadangi regulator, siltation of the drainage canal and cracked wing wall were observed during the field visit. Although the operation was not affected much, the damages need to be repaired soon as it may pose the risk of getting worse with time. COSTS AND BENEFITS OF FLOOD FIGHTING Except MDIP and Sakunia Beel project, all the other projects could not achieve their desired targets of protecting the Aman crop. In D-N-D and N-N projects, the Aman crop was partially damaged due to drainage congestion resulting from persistently high river water level above the design level. Because of flooding in Nagor River project, the damage to Aman crop was total. In CPP, the compartmentalization concept has not worked and inadequate drainage together with failure by the sub-compartments to store floodwater has resulted in reduction in Aman area by about 50%.

But, in spite of the damages inflicted on Aman crop, preliminary analyses on costs of flood fighting and returns from the Aman harvest in MDIP, CPP, D-N-D and N-N projects show that the benefits overwhelmingly outweigh the costs incurred, even without considering the intangible benefits. The costs of flood fighting and the gross return from Aman harvests for the studied projects are given in Table 3. Table 3: Costs of flood fighting and returns from Aman harvests

Name of Project

Cost of Flood Fighting*

(Million Taka) Return from Aman Harvest

(Million Taka)

CPP 4.78 5.64 D-N-D 7.00 38.20 Meghna-Dhonagoda 7.40 234.40 Nagor River 1.07 0 N-N 4.00 19.20 Sakunia Beel N/A N/A

*Source: BWDB (1998) CONCLUSIONS For N-N and D-N-D projects of lower Brahmaputra basin and MDIP of lower Meghna basin, the 1998 flood was catastrophic with 100-year return period,


which exceeded all previous records, both in terms of water level height and duration. Even though the 1998 flood was catastrophic in the Ganges basin, Sakunia Beel project was not seriously affected, as this has now become a compartment of the Greater Faridpur-Barisal project. For both Nagor River project and CPP, the 1998 flood magnitude was lower than the previous highest flood (1988), with return periods of 10-year and 27-year, respectively.

Even though the 1998 flood magnitude exceeded their respective design flood levels, D-N-D, MDIP and N-N projects were able to withstand the flood because of the freeboard and preventive measures taken by BWDB. In CPP also, the 1998 flood was higher than the design flood but the embankment was breached at three locations and cut at one location by outsiders, thus, partly depriving the people of the benefits of the embankment. In Nagor River and Sakunia Beel projects the flood levels were below the design flood levels. But, because of public cuts by outsiders, most of the project was inundated.

Except MDIP and Sakunia Beel project, all the other projects could not achieve their desired targets of protecting the Aman crop. In D-N-D and N-N projects, the Aman crop was partially damaged due to drainage congestion resulting from persistently high water level above the design level. Because of flooding in Nagor River project, the damage to Aman crop was total. In CPP, the compartmentalization concept has not worked and inadequate drainage together with failure by the sub-compartments to store flood water has resulted in reduction in Aman area by about 50%.

In spite of the damages inflicted on Aman crop, preliminary analyses on costs of flood fighting and returns from the Aman harvest in MDIP, CPP, D-N-D and N-N projects show that the benefits overwhelmingly outweigh the costs incurred, even without considering the intangible benefits.

From the review of the design criteria of the studied projects, analysis of the 1998 flood data and performance evaluation of the projects during the 1998 flood, it has become evident that the present embankment design criteria is adequate to withstand even the 100-year flood and there is no need to revise or upgrade the design flood level. In none of the studied projects the embankment was overtopped and breaches that occurred were not because of faulty design but because of public cuts, poor construction and/or inadequate maintenance.

The repercussions of FCD/FCDI projects on the flood level in the unprotected areas deserve a holistic analysis. Unless this is done the disgruntled outsiders will continue to cut the embankment and curtail the benefits of the FCD/FCDI projects. Piecemeal implementation of FCD/FCDI projects should be discontinued and projects should be implemented only after rigorous regional analysis.




REFERENCES BWDB (1996).Standard Design Manual, Vol. 1, Design Section, Bangladesh

Water Development Board, Dhaka. BWDB (1998). Report on damage to infrastructure during 1998 flood and the

probable cost rehabilitation. Monitoring and Evaluation Office, Bangladesh Water Development Board, Dhaka.

Chowdhury, J.U., M.R. Rahman and M. Salehin (1996). Flood control in a flood plain country: Experiences of Bangladesh. Institute of Flood Control and Drainage Research, BUET and Islamic Educational Scientific and Cultural Organization, Rabat, Morocco, p.135.

FPCO (1991). FAP 12 FCD/I Agricultural Study: Rapid Rural Appraisal of Nagor River Project, Flood Plan Coordination Organization, Dhaka.

Kruger Consult and BCEOM (1992). Flood Modeling and Management, Flood Hydrology Study, Main Report, FAP 25, Flood Plan Coordination Organization, Dhaka.

Saleh, A.F.M., S.M.U. Ahmed, M.M. Miah, M.R. Rahman, M. Salehin and M.S. Mondal (1998). Performance Evaluation of FCD/FCDI Projects During 1998 Flood. Institute of Flood Control and Drainage Research, Bangladesh University of Engineering and Technology, Dhaka.


Experiences with Flood Management Practices During the 1998 Flood

Mohammad Rezaur Rahman and Jahir Uddin Chowdhury

Institute of Water and Flood Management Bangladesh University of Engineering and Technology, Dhaka-1000, Bangladesh Abstract Bangladesh currently employs a variety of flood management options such as flood control, controlled flood, flood proofing and flood forecasting. This paper discusses the experiences with these flood management options during the unprecedented flood of 1998. In the light of these experiences, this paper suggests a few courses of action, which are expected to improve the performance of these practices during such large floods in future. THE FLOOD OF 1998 Bangladesh is a floodplain country located on the delta of three mighty rivers, the Ganges, the Brahmaputra and the Meghna. Most of the floods in Bangladesh are caused by the bank overflow from these major rivers and their tributaries and distributaries. These rivers carry enormous volume of flood water generated by heavy rainfall in upper catchments outside the boundaries of Bangladesh, which occupies only 7% of 1.72 million sq.km of total catchment area of Ganges-Brahmaputra-Meghna river system. The cause of the flood of 1998 is also the heavy rainfall in upper catchments of the Ganges and the Brahmaputra. During the 1998 flood season, trans-boundary flow contributed to 5.95 meters of water and internal rainfall contributed to 1.66 meters of water (see Table 1). Such amount of water caused flooding in more than two-third area of the country.



Table 1: Comparison of Trans-boundary Flow and Internal Rainfall During 1998 Flood

Month Trans-boundary run-off (meters)

Internal rainfall (meters)

July 1.82 0.77 August 2.25 0.59 September 1.88 0.30 Total 5.95 1.66

Table 2 shows the water levels and duration of flood at three major rivers. As can be seen from Table 2, the 1998 flood level crossed recorded highest flood levels in the Ganges and the Lower Meghna rivers. The most damaging feature of 1998 flood was its long duration. The flood remained above mean river bank level continuously for record duration at Bahadurabad on the Brahmaputra and Hardinge Bridge on the Ganges. Table 2: Comparison of 1998 Flood Level and Duration with Previous Record

Highest water level (meters PWD)

Duration (days above bank

level)

Station and river

1998 flood

Previous record and year

1998 flood

1988 flood

Hardinge Bridge on Ganges 15.19 14.87 (1988) 29 27 Bahdurabad on Brahmaputra 20.37 20.61 (1988) 76 30 Chandpur on Lower Meghna 5.62 5.16 (1988) 75 87

The 1998 flood hydrograph of Lower Meghna river at Chandpur is shown in Fig. 1 and is compared with 1988 flood hydrograph. Stage-duration-frequency analysis carried out by Islam and Chowdhury (1999) indicates that at Chandpur both the highest level and the duration were unprecedented and were higher than that of 100 year return period. Salehin et al. (1999) stated that continuous inflow of very large amount of water caused abnormal tide behaviour at Chandpur. The study found no effect of spring tide, monsoon wind set up or tsunami on the 1998 flood. 268 Engineering Concerns of Flood

Experiences with Flood Management Practices


The 1998 flood caused enormous damage in the country, which can be summarized after IEB (1998) as follows:

Total area affected : 100,000 sq.km Districts affected : 52 Thanas affected : 366 Number of affected people : 30,916,351 Crop loss : 2.2 million tons Highways and roads damaged : 15,927 km. Bridges and culverts damaged : 6,890 Embankments damaged : 4,528 km

Date

Figure 1: Water Level Hydrographs of Meghna at Chandpur

FLOODPLAIN PROCESSES An important physiographic feature of Bangladesh is that except Chittagong region, rest of the area mainly consists of floodplain. Fig. 2 shows the areal extent of inundation each year since 1954. The areal extent of the floods in 1988 and 1998 is evidence that almost the entire country is basically a floodplain. A declining trend in the flooded area is observed in Fig. 2. This is due to construction of around 400 flood control projects over the years. Chowdhury et. al (1996) estimated that since 1964,


there has been a growth of flood control coverage of 120,000 ha/year resulting in flood free area of 80,000 ha/year. Almost 60% of the potential area for flood protection is now under flood protection. Such large scale intervention in a floodplain environment is likely to produce repercussions. It is seen from Fig. 2, that despite the declining trend in flooded area, the recent floods in 1998, 1988 and 1987 are the largest recorded floods. It is also seen that year-to-year variability of flooded area has increased in recent years indicating an unstable system.

Figure 2: Extent of Inundation During Different Years

One of the reasons for poor performance of many structural interventions within the floodplain including roads, flood control projects etc. is the lack of recognition of floodplain processes and functions and their relationship with the rivers. It has been seen that population do desire some form of protection especially from large floods. Consequent land and water use interventions creates opportunities for social upliftment. However, while flood control interventions bring economic benefit to one section of the society, cause economic hardships to another section especially to those poorer sections who are dependent on many free resources of floodplain. The resultant conflict of interest in a densely populated country often compromises project performances. It has been observed that while projects affect the floodplain 270 Engineering Concerns of Flood



functions, the natural processes also affect the intended functioning of the projects. Such socio-economic and environmental interactions in the floodplain are illustrated in Fig. 3.

Figure 3: Schematic Diagram of Interactions in a Floodplain Environment An important beneficial hydraulic function of floodplain is that it moderates the

flood flow by acting as storage. Storage function of floodplain is also very helpful in maintaining the channels in the coastal region of Bangladesh where tide occurs twice a day. Another beneficial function of floodplain is that it augments the post-monsoon river flow by gradually releasing water from its flood storage. Rainfall and flood water over the floodplain infiltrate and recharge the unconfined aquifer. Lateral recharge also occurs from rivers at high water level. During dry season when the


rivers are at low water levels, major portion of their flow comes from groundwater discharge from the upper aquifer.

While large floods such as those of 1998 and 1988 cause enormous damage to economy, normal floods have been proven to be beneficial in many ways. The river-borne sediments, which are dispersed over the floodplain, are valuable sources of soil nutrients. There is empirical evidence of bumper harvest of Boro rice after every major flood. In Bangladesh, fish is second only to rice as a source of food. Breeding, multiplication and sustenance of the inland water fish and prawn populations are intimately bound to the sequence of annual flooding. The great majority of our wildlife species are also directly or indirectly more intimately associated with the aquatic habitat. A number of globally endangered species depend upon the floodplain wetlands of Bangladesh. FLOOD MANAGEMENT PRACTICES OF BANGLADESH Structural Measures

Many conventional flood mitigation measures like flood control reservoirs, flood diversions or flood bypasses are not feasible inside Bangladesh because of its extreme flat topography and high population density in the floodplain. The principal structural flood management measure that has been adopted in Bangladesh is construction of embankments parallel to the riverbanks. Drainage of floodwater is facilitated by re-excavating drainage channels and constructing drainage regulators and sluices. Most of the flood control projects in our country are intended for protecting agricultural land against river flood. Such projects are designed for protection against a 20-year flood. As of 1993, there were 372 flood control and drainage projects giving protection to a total of 3.72 Mha.

In recent years, due to many adverse impacts of flood control projects on environment, most notably on fisheries, controlled flooding concept has been advocated and administered in a pilot scale in Tangail. According to this concept, flood water is not completely eliminated, but rather it is allowed to enter the project area in a controlled fashion in order to retain many benefits of normal flood while protecting the project area from abnormal floods. Non-structural Measures

Importance of non-structural measures of flood management such as flood forecasting and flood proofing is becoming evident especially during large flood




events. The Flood Forecasting and Warning Center (FFWC) of BWDB, established in 1972, is responsible for making flood forecasts and flood warning during the flood season. At present, the FFWC issues forecast of river stages for 46 stations on major and medium rivers, formulated for 24, 48 and 72 hours.

Flood proofing of homestead is a tradition of the rural settlements in Bangladesh. Rural homesteads are generally raised above usual flood level. As per Bangladesh National Building Code, any area having a potential for being flooded under at least 1m deep water due to flooding should be delineated as flood prone area (FPA). The code specifies that the lowest floor, including the basement, of any building within the FPA shall not be located below the design flood level, and the roof of one or two story buildings and the floor immediately above the design flood level for three or more story buildings shall be accessible with an exterior stair. EVALUATION OF FLOOD CONTROL MEASURES Agricultural Land Protection

High water level and long duration of the flood threatened many of the flood control projects. About 5,000 km of embankment was damaged. As a typical example, damages suffered by Meghna-Dhonagoda Irrigation Project (MDIP) are shown in Table 3. Damages were extensive due to long duration (more than two months) of the flood. The damages reported in Table 3 are due to typical floodplain soil characteristics and earthen nature of embankment. Risk of failure of embankment is omnipresent in the complex geo-morphological setting of the floodplain. Table 3: Damage Assessment of MDIP by BWDB (Source: Saleh et al., 1998)

Sliding Types of damages

Seepage Boiling Piping due to rat hole c/s slope r/s slope

Extent of damage

250 nos 322 nos 83 nos 3611 m 2525 m

Fortunately, large flood control projects, such as the Brahmaputra Right

Embankment, the Meghna-Dhonagoda Irrigation Project and the Chandpur Irrigation Projects were not breached due to constant monitoring and timely intervention by BWDB engineers. The projects were not overtopped either, despite the fact that flood level often exceeded design flood level. Saleh et al. (1998) surmised that the additional freeboard of 3 feet (0.9 meter) saved many


projects from overtopping. Indeed, Chowdhury et. al (1996) showed that due to extreme flatness of the country, the difference between 20-year flood level and 100-year flood level is less than 1 meter. Therefore, it can be stated that the existing design crest level of the embankments is able to protect the projects from 100-year floods.

However, duration of flood, which produces continuous seepage pressure on the earthen embankment, need to be made an important design consideration so that structural integrity of the earthen embankment can be maintained during long duration floods. Long duration flood also causes morphological changes in the river, which may compromise the safety of structures. MDIP was seriously threatened by river erosion during the 1998 flood.

Typical floodplain hydrology played a role in failure of some flood control projects. For example, Nagor river project failed due to public cuts by outsiders. This occurred as outsiders perceived that water level outside the project was higher due to this project. Their apprehension is not without ground. HTSL (1992) claimed that 13 projects out of 17 projects evaluated by it have given rise to higher water levels outside the projects. Harza (1991) asserted that all the eight projects that it evaluated have produced higher water levels outside. External impact of flood control projects can not be avoided especially in floodplain setting. But moderating such impacts through appropriate design modification is an important pre-condition for sustainability of projects.

The flood control projects saved considerable amount of foodgrain (MDIP alone saving more than 40,000 tons of Aman rice) which otherwise would have to be imported at the expense of scarce foreign reserve. Saving crops from such large floods can form a sounder justification for flood control projects. Urban Protection

The importance and need of structural measures for flood control in urban areas have been underlined during 1998 flood. Flood protection for major urban areas are provided against 100-year flood while secondary towns are provided with protection from 50-year floods. During the 1998 flood, most of the western part of Dhaka City was saved from inundation by the embankment, which was built after the 1988 flood. Some part of the western Dhaka specially the diplomatic enclave and eastern Dhaka was flooded. While the agricultural flood control projects are solely managed by one agency namely BWDB, Dhaka flood protection works are managed and operated by three agencies namely, BWDB, WASA and RAJUK. Lack of coordination between these three agencies has caused flooding in the western part of Dhaka city (Chowdhury et al., 1998). 274 Engineering Concerns of Flood



The government has decided to protect the eastern part of Dhaka City by constructing a 60-km embankment. While constructing this flood protection work, it is advisable that the planners look at the deficiencies of flood protection works in Dhaka west and take appropriate measures so that the same mistakes are not repeated. Among these deficiencies, Chowdhury et al. (1998) reported drainage congestion, lack of integrated development of urban infrastructures and formulation of appropriate land use regulation. EVALUATION OF CONTROLLED FLOODING MEASURES The Compartmentalization Pilot Project (CPP), Tangail was developed under Flood Action Plan (FAP-20) to promote the concept of ‘controlled flooding’ from both outside and inside the embankment or compartment. A compartment has been defined as an area in which effective water management, particularly through semi-controlled flooding and controlled drainage, is made possible through structural and institutional arrangements. It was envisaged in the FAP-20 that the CPP would serve three purposes: (i) flood control in order to reduce flood damages to human lives, economic assets and agricultural production; (ii) controlled flooding to capture the benefits of flooding for agriculture, fisheries and navigational purposes; and (iii) flood storage to store flood water and/or rainwater in pre-defined sub compartments to reduce flood damages elsewhere in the floodplain. Contrary to the concept of the traditional FCD/FCDI projects, where river water flooding is both unacceptable and undesirable, the compartmentalization concept recognizes the beneficial effects of river flooding on agriculture (silt deposits and nitrogen fixation) and on fishery. The CPP was commissioned in 1995 and the 1998 flood was the first major test for this pilot project.

The 1998 flood water level corresponded to a level with 27-year return period while the design level is 20-year flood level. The water level difference between inside and outside of the CPP during 1998 flood was more than 2m. As per the original concept, the embankment was to be overtopped by floodwater with all the inlet structures open, if the flood exceeded the 20-year flood. However, this could not happen, as the project people did not allow it. During the flood, the embankment was not overtopped, but there were three breaches and one public cut by outsiders. Saleh et al. (1998) expressed that ‘compartmentalization concept’ has not worked during the 1998 flood, but the project is deriving benefits from controlled flooding.


EVALUATION OF FLOOD PROOFING MEASURES Properties and infrastructures have suffered substantial damage during 1998 flood, which would in turn affect the national economy profoundly. A program of flood proofing and flood preparedness can reduce vulnerability of the society to floods. Highway links of Dhaka with Chittagong, Aricha, Sylhet and Tangail were cutoff for quite a considerable time due to submergence of the highways at several places. Disruption of communication lines and essential utility services result in considerable suffering of the population. The national highways should be of adequate height so that transport is not disrupted during large floods. After the 1998 flood, it has been estimated that the Dhaka-Chittagong highway need to be raised by at least 1 meter at places to keep the highway free from submergence during such large floods. However, such costly decisions need to be made on the basis of appropriate risk based analysis.

Vital installations and other infrastructures should also be flood friendly. Food godown, domestic water supply sources and capital assets should be secured by making them flood resistant. During 1998 flood, many industries especially in Narsingdi-Narayanganj-Munshiganj belt were submerged hampering the industrial production. Daily economic and employment activities can be kept functioning by making industrial installations and business centers flood proofed.

Although Bangladesh has a building code, it is yet to be institutionalized. Due to this lacking, many houses in the floodplains have been built without due consideration to the flood risk. First floor of almost all the houses in the Balu floodplain of Dhaka city were inundated and remained submerged for almost two months during the 1998 flood causing immense suffering to the inhabitants. If the building code was institutionalized and strictly adhered to, then sufferings of the floodplain inhabitants could be considerably mitigated.

Flood control embankments, roads and other infrastructures affect the water regime in the floodplain. Construction of rural roads and highways is growing every year. Urban areas, industrial areas and rural growth centres are expanding rapidly. There should be co-ordinated planning and construction of flood control projects, roads and other water regime affecting infrastructures. A floodplain land use regulation can be formulated so that planning, design and construction of infrastructures take into consideration the flood risk, keep adequate provision for unimpeded drainage and also account for the preservation of environment.




FLOOD FORECASTING AND WARNING Presently flood levels are forecasted at river gauging stations. During the 1998 flood, water levels at different stations were adequately forecasted. However, flood forecasting and warning process need to be made more useful and meaningful to the people. Currently no information is given about areal extent of flooding. The forecasting method should be such that it can give sufficiently advance warning for Thanas those are at the risk of flooding. It must be able to forecast the position of flood level in next two or three days at critical locations of vital infrastructures; for example, probable position of flood level with respect to the top of roads, highways and flood protection embankments. The current lead-time of 72 hours can also be increased. Chowdhury et. al (1998) have shown that flood level at Dhaka city can be predicted 4-6 days ahead. CONCLUSIONS The experiences during the 1998 flood in relation to various flood management practices that are currently in use in the country have been discussed in this paper. The experiences show that each of the options has scope of improvement either in planning, design or execution, which are summarized below. Flood Control

During the 1998 flood, it has been observed that flood control measures have been able to protect large areas from inundation. The untold suffering of the people in the unprotected area will raise the call for further protection as was evident in case of eastern part of Dhaka City. Even before the 1998 flood, it was a well-known fact that people in general demand for protection against flood. The case will only be stronger after the 1998 deluge and probably there is no alternative to structural measures in ensuring protection to urban and industrial areas.

Regional hydrology play an important role in sustainability of projects. It has been seen that physical interventions by interfering with the floodplain processes have created social tension among different sections of the floodplain dwellers hampering the project performances. These experiences clearly show the need for maintaining harmony with the floodplain processes to the extent possible.

It was the unprecedented duration, not the high level of flood, that caused much more damage to the embankments. Thus duration of flood should be an important design parameter in future flood control projects. Since flood control projects saved large crop areas, saving of crops during large floods should get appropriate


consideration in benefit-cost analysis at the planning stage of future projects.

Compartmentalization Pilot Project (CPP)

The CPP was severely tested during 1998 flood. It was apparent that compartmentalization concept did not work. However, controlled flooding concept has gained popularity in CPP and may be repeated in other existing and future projects to derive benefits of normal flood. Flood Proofing

An approach needs to be adopted whereby large floods can be managed while enjoying benefits of normal flood. In managing large floods, structural approach will of-course play its due role. But non-structural measures should get serious consideration considering their cost-effectiveness and environment friendliness. Especially during large floods, non-structural measures are the best option in containing damages.

Various flood-proofing measures have to be institutionalized. Lack of institutionalization caused untold sufferings in the Dhaka East. Flood-proof measures of infrastructures should be based on risk based analysis. Flood Forecasting

Flood forecasting need to be more user-friendly. In addition to forecasting of river water levels, flood level with respect to vulnerable areas and key linkages need to forecasted so that people can plan their activities ahead of the time. REFERENCES Chowdhury, J. U., Rahman, M.R., Bala, S. K. and Islam, S. (1998), Impact of 1998 Flood on Dhaka City and Performance of Flood Control Works, IFCDR, BUET. Chowdhury, J. U., Rahman, M. R. and Salehin, M. (1996), Flood Control in a Floodplain Country: Experiences of Bangladesh, IFCDR, BUET. Harza (1991), Evaluation of Historical Water resource Development and Implications for the National Water Plan, National Water Plan Project - Phase II, Water Resources Planning Organization, Dhaka. 278 Engineering Concerns of Flood



HTSL (1991), FAP 12, FCD/I Agricultural Study, Flood Plan Coordination Organization, Dhaka. The Institution of Engineers, Bangladesh (1998), Report of Task Force Committee on Flood Management, Dhaka. Islam S. and Chowdhury, J. U. (1999): Hydrological Aspects of 1998 Flood, Paper presented at the 43rd Convention of Institution of Engineers, Bangladesh, Dhaka, March. Saleh, A. F. M., Ahmed, S. M., Mirjahan, Md., Rahman, M. R., Salehin, M. and Mondal, M. S. (1998), Performance Evaluation of FCD/FCDI Projects during 1998 Flood, BUET. Salehin, M., Hoque, A. and Chowdhury J. U. (1999): Investigation of Issues related to Tides and Tsunamis during 1998 Flood, Paper presented at the 43rd Convention of Institution of Engineers, Bangladesh, Dhaka, March.


Remote Sensing Imagery to Assess the Environmental Impacts of Flood

M. J. B. Alam, M. H. Rahman and Md. Mujibur Rahman

Department of Civil Engineering Bangladesh University of Engineering and Technology, Dhaka-1000, Bangladesh Abstract Developments in remote sensing technology have opened a new horizon in disaster management, particularly in the field of assessment of flood damages. Now a days, various kinds of satellite images have become readily available. Using these images, ground conditions can be assessed on a pixel by pixel basis enabling identification and analysis of land surface. The abundance of information available from these images makes it possible to address quarries in numerous environmental issues. This study focuses on the applicability of satellite images in the monitoring and analysis of environmental impacts including changes in land coverage and extent of flooding. A software has been developed for the analysis of satellite images. The software has been applied to analyze the images for predicting the changes in land coverage. The accuracy of the analysis depends on the resolution of the images. Although the images sometimes provide noisy spatial patterns, it has been observed that even the low-cost, low-resolution images can successfully be used for satisfactory analysis. For this purpose, appropriate filtering technique must be utilized to preprocess the image.

INTRODUCTION During the last couple of years a lot of new remote sensing techniques have been developed and numerous satellites have been set up to monitor the changes on the earth’s surface round the clock. Due to this evolution, satellite images have




become readily available and widely used for different purposes including investment decisions, disaster management and even war monitoring and guidance (Eric and Laura, 1997 and Ramsey and Chappel, 1997). In the framework of integrated remote sensing and GIS, satellite images have become an integral tool for the planners and decision-makers. The cost of images vary widely depending on the resolution, accuracy and source. All the images are not suitable for all the purposes. For example, conventional broad band sensors such as Spot-XS, Landsat MSS and Landsat TM are not suitable for mapping minerals or soil properties (Jong, 1998). On the other hand, special purpose high-resolution images are extremely costly. Sometimes appropriately selected low-cost images can provide information, which are almost as good as the high-cost ones. For this purpose the images should be selected and pre-processed carefully. Appropriate merging and filtering techniques must be used for pre-processing the images. In Bangladesh, satellite data is primarily used for meteorological purpose by the Bangladesh Meteorological Department. In the country, SPARSO is the principal organization involved in the collection and study of satellite images. For the analysis of satellite images, costly softwares are imported from other countries. These softwares are usually tailored for specific kind of jobs, which limits the extent of analysis. This study explores the potential of satellite imagery for monitoring the extent of flood and its environmental impacts. The country suffers from flood every year, which cause significant damages to the economy and sufferings of the people. Appropriate monitoring systems will facilitate faster and more accurate monitoring of the extent of flooding then the conventional system of collecting information from the field. For the purpose of analyzing satellite images a software had been developed using Visual Basic Programming Language. Images acquired from GMS and NOAA satellites have been used for the analysis. FLOOD OF 1998 Due to intermittent heavy rainfall in the country and in the upper catchment areas from July to early September of 1998, all the rivers of the country over-flowed and caused severe flood. The flood affected about 68 percent areas of the country. The rivers of the country started experiencing on-rush of flow from the middle of July and by that time the low-lying areas of the country had already gone under water. At that time, about 45,000 sq. km. of 37 districts of the country were affected by flood. Although flood situation started improving in early August, the flow of the two main rivers of the country- Padma and Brahmaputra-Jamuna increased significantly in the middle of August. This was caused by heavy rainfall in the upper catchment areas. By the end of August flood situation

Remote Sensing to Assess Environmental Impacts of Flood


became worse and about 60,000 sq.km area of 42 districts were affected. During the early September the flow of the major rivers increased abruptly worsening the condition. The flood situation became worst in the second week of September and about 75,000 sq.km area of 52 districts were affected during that time. The flooded condition existed from early July to the last week of September, for nearly about three months at different places in different magnitudes. Thus flood of 1998 became the most prolonged flood in the history of the country. The total flood inundated area was about 1,00,250 sq.km (68 percent of the total area of the country) affecting 53 districts (Annual Flood Report, 1998). METHODOLOGY The main objective of the study is to develop a methodology to analyze satellite image for monitoring extent of flood and its environmental impacts. For this purpose the satellite images should be processed initially and then the information from those images is to be extracted.

Automated color based interpretation technique is applied to analyze the images. NOAA and GMS images are used for the purpose of analysis. Images from June to September of 1998 are used in the analysis, which are collected from SPARSO. A typical satellite image is shown in Figure 1. Color differences are considered in order to identify and classify flood-damaged areas. For the purpose of calibration of the software, ground information with respect to specific images of particular dates are utilized. Bangladesh Water Development Board (BWDB) provides the ground information. For a particular application, the land coverage is classified into four classes namely – (I) Severely Flood Affected Area, (II) Moderately Flood Affected Area, (III) Lightly Flood Affected Area, and (IV) Not Flood Affected Area.

In the calibration phase of the analysis, the color value in the form of Red, Green and Blue component is extracted for each of the types mentioned above by using calibration images and ground information provided by the relevant organizations. By analyzing the calibration images, the range of the values of each component for each of the classes is obtained which formulates the classification framework, which is shown in Table 1. This classification scheme is then applied on the satellite images to obtain the amount of areas in different classes and monitor their changes.



Table 1: Classification Scheme to Monitor Extent of Flood and Its Changes

Classification of Land Surface Red Green Blue Severely Flood Affected Area 75-125 35-65 55-140 Moderately Flood Affected Area

40-60 50-80 100-160

Lightly Flood Affected Area 85-190 75-225 85-180 Not Flood Affected Area 30-115 85-170 170-225

DESCRIPTION OF THE SOFTWARE TO ANALYZE THE IMAGES A user-friendly software has been developed to analyze the images using Visual Basic programming language. Different interfaces of the software are shown in Figures 1 and 2. The reasons for the selection of designing the platform for the software is described below: ♦ Visual Basic 5.0 supports various types of common image files, e.g., BMP,

JPG, WMF etc. So, it will be convenient to use the developed software for several types of images.

♦ Color value of any point in images can be read very easily. ♦ Color value of any point can be obtained in two ways – (1) read the value as

a long color value, (2) read the value in RGB form (split the value in Red, Green and Blue color components).

♦ The developed software can be made efficient in browsing picture files and in loading them from anywhere of the hard drives of the computer.

♦ The output can be made in a presentable form and can be linked with any database file.

♦ More flexibility can be associated with the software so that it becomes more user-friendly.

Description of the Software

As stated earlier, the software has been designed to analyze images to extract data. The satellite images are not like normal photographs. The color of any object in a satellite image is not the same as in a normal photograph. For example, in a satellite image forest will not be as green as it is in a normal photograph. This variation is caused by noises due to air, temperature differential, suspended particles and sensor capabilities. To make the analysis practically meaningful, color property of various objects in this special image type needs to be identified and a range of color is to be set for each object. This



phase of the analysis is called calibration phase. Afterwards the calibrated classification schemes can be utilized to analyze the images. Calibration Phase In this phase, color property is calibrated to identify the objects during the analysis. The calibration phase consists of a simple process to set a color range for selected objects. In this phase the objects are identified and the range of color value is established for each object. Color values or properties are read for a certain type of object from a satellite image for which the ground conditions are known. Three components in the color value i.e., Red, Green, Blue (RGB), are used to identify the color. After taking a number of color values (RGB) for a certain object they were scrutinized and compared with each other and a color value range is established which represents the color of that particular object. In this way, the color value ranges are established for all the selected objects and a “Decision Table” is prepared. This “Decision Table” is then used in developing the software for the analysis. Analysis Phase This is the main component of the software. After establishing the color value component (Red, Green, Blue) range properly for the selected objects, the analysis phase is implemented. In the analysis phase the red, green and blue color values of each pixel is extracted and the object in the pixel is identified using the decision table prepared in the calibration stage.

The components of the software for calibration and analysis inter-phases are shown in Figure 1 and 2. RESULTS Using the software described above satellite images taken in the months of June, July, August and September of 1998 have been analyzed in the study. Results obtained from the analysis are presented in Table 2.

The results obtained in the study are validated by using data from secondary sources such as Bangladesh Water Development Board (BWDB), Local Government Engineering Department (LGED) and Bangladesh Meteorological Department (BMD). The analysis by the computer program developed in the study is in close agreement with the results predicted by another program written in C++ programming language, which has been developed independently.



Figure 1: Software Inter-Phase for Calibration Stage

Figure 2. Software Inter-Phase for Analysis Stage



Table 2: Extent of Flood and Its Changes As Obtained from the Software

Date of Imaging Classification of Land Surface End of

June, 1998 Middle of July, 1998

End of August,

1998

Beginning of September,

1999 Severely Flood Affected Area

17631 58670 49199 76565

Moderately Flood Affected Area

6292 6569 3534 324

Lightly Flood Affected Area

56364 20595 35917 35874

Not Flood Affected Area

63713 58165 55349 31237

CONCLUSIONS The satellite based remote sensing techniques proved to be appropriate methods for documenting and analyzing damages caused by flood. Remote sensing provides an overview of a regional problem, particularly in extensive areas with scarce information. These techniques also provide a basis for monitoring the natural environment because the analysis of remote sensing data at regular intervals reveals the trends and changes. Due to low resolution and noise the extracted information may not be very accurate. But it acts as an instantaneous tool for decision-making. Also by using appropriate techniques and filtering the low-cost satellite images, the accuracy of analysis can be greatly improved. Compared with field survey, the remote sensing satellite images provide a lot of more information at much cheaper cost.

In this study, low cost satellite images have been successfully applied to monitor the impacts of the devastating flood of 1998. For this purpose, an integrated software has been developed to analyze the images. From the analysis it has been found that during the middle of July, end of August and Early September of 1998 about 65000, 53000 and 77000 sq. km. of the country was under flood. The accuracy of the analysis is checked by collecting information from organizations responsible for monitoring flood situation. It has been observed that the differences between the results of this study and relevant information from the sources mentioned above are less than 10 percent. The software developed in this study can be used to analyze any image for the purpose of studying land use, vegetation cover, soil properties, etc.



The study can further be extended incorporating GIS framework. GIS proved to be an excellent tool for studying the environment, combining different maps, layers, variables and data from diverse sources. Prediction analysis and trend assessment could be realized using a GIS framework.

The accuracy of the analysis can be improved by using images of higher resolution. For this purpose Airborne Visible/InfraRed Imaging Spectrometer (AVIRIS ) or Landsat TM images can be utilized. Landsat TM records data in seven bands with a spatial resolution of 30mX30m. AVIRIS acquires images at a spatial resolution of 20m x 20m having 224 bands with a nominal band interval of 10 nm. ACKNOWLEDGEMENT The authors acknowledge the contributions of Mr. J.R. Khan and Ms. N. Ferdous as research assistants in the project. REFERENCES Eric, S. K. and Laura, L. B. (1997), Monitoring South Florida Wetlands Using

ERS-1 SAR Imagery. Photogrametric Engineering and Remote Sensing, Vol. 63 No. 3.

Jong, S. M. De (1998), Imaging Spectrometry for Monitoring tree Damage Caused by Volcanic activity in the Long Valley Caldera, California. ITC Journal 1998(1).

Ramsey, E. W. and Chappel, D. K. (1997) AVHRR Imagery Used to Identify Hurricane Damage in a Forested Wetland of Lousiana. Photogrametric Engineering and Remote Sensing, Vol. 63 No. 3.

Jurio, E. M. and Juidan, R. A. (1998), Remote Sensing, Synergism and Geographic Information System for Desertification Analysis: An Example from Northwest Patagonia, Argentina. ITC Journal 1998-3/4.

Annual Flood Report (1998), Flood Forecasting & Warning Centre, Processing and Flood Forecasting Circle, Bangladesh Water Development Board.


The Socio-Economic Impacts of the 1998 Flood in Dhaka City

Sarwar Jahan

Department of Urban and Regional Planning Bangladesh University of Engineering and Technology, Dhaka-1000, Bangladesh Abstract This paper presents the results of a study that was undertaken to analyze the socio-economic impacts of the 1998 flood in Dhaka city. The study indicates that the flood affected the people of different income groups in different parts of the city. The low-income people living mostly in low-lying areas, however, suffered more heavily than the middle or upper-income people. The flood caused heavy damage to housing, health, job and business income. Overall damage to households has been found to be dependent on income. Statistical analysis has shown that there is a positive correlation between the level of income and the extent of damage, but a negative correlation between the level of income and the burden of such damage. The study also indicates that majority of the people tried to make up the losses or repair the damages with their own savings while the poorer sections of the people had to depend on others to cope with the flood damage. In their efforts to cope with the disaster, the low-income people, however, received more help from friends, relatives and voluntary organizations than the governmental or non-governmental organizations. INTRODUCTION As a natural hazard, floods are common phenomena in Bangladesh. About 18 percent of the land area is flooded during the monsoon season every year. The problems of flood in Bangladesh came to the forefront after the two consecutive floods of 1954 and 1955. Since then a large number of flood studies were completed and quite a few flood control measures implemented. But the


Sarwar Jahan


devastating floods of 1987, 1988 and 1998 gave rise to a feeling that much more needed to be done.

An analysis of flood damage statistics indicates that the extent of damage has increased along with the increase in the intensity of flooding since 1954. Table 1 presents the estimated damages due to severe floods during 1970’s, 1980’s and 1990’s. In 1988 about two-thirds of the country were inundated, affecting 50 million people and killing 1600 of them. This catastrophic flood hit the greater Dhaka area during the months of August and September. About 56 percent of the greater Dhaka area was submerged affecting about 1.9 million people (JICA, 1990). While no official figures of flood damages in Dhaka are available, the Dhaka city corporation estimated that some 400 km. of roads were damaged. From the estimates of JICA for an area of 137 kms2 which includes the major built-up part of greater Dhaka, flood damage was estimated in the order of Tk. 500 millions to Tk. 1000 millions.

The 1998 flood was an unprecedented event of its kind in terms of duration, inundation of areas and damages (DMB, 1998). The overall duration of the flood throughout the country was 65 days while the longest duration was 73 days at a single point. The flood inundated nearly 100,000 sq. km. of 52 districts affecting more than 30 million people. Total economic damage amounted to nearly 3 billion dollars (see Table 1).

Table 1: People affected and overall damage by the severe floods in Bangladesh since1970

Year

People affected (Millions)

Overall damage (Millions Tk.)

1974 30 600 1980 20 120 1984 20 130 1987 41 1000 1988 50 1200 1998 30 2900

Source: Elahi(1988), DMB(1998)

Dhaka city was also severely affected by the 1998 flood. Seventy out of ninety wards of Dhaka City Corporation went under water of various depths, which lasted for more than eight weeks. The flood affected almost all aspects of human life. It affected not only the physical assets of the people, but also their

Socio-Economic Impacts of Flood

income, health and occupation. People of various income and occupation groups suffered in varying degrees due to the flood. There were also significant spatial variations in the impacts of the flood.

This paper presents the results of a study that was undertaken to determine the nature and degree of impact of the flood on various socio-economic groups in Dhaka city and the mechanisms through which people coped with the flood. More specifically, the aims of the study were: (i) to investigate the extent of damage to lives and properties, income, job, health etc. across income groups; (ii) to study the coping mechanism of the people during the disaster; and (iii) to determine the help and assistance received by the affected people from governmental and non-governmental organizations and private individuals.

The study was carried out in four areas in the eastern part and one area in the southern part of the city. The areas in the eastern part were Meradia, Basabo, Anandanagar and Gulshan while in the southern part the area was Kamrangir Char. Data were collected from a total of 294 households out of which 66 were in Kamrangir char, 73 in Meradia, 32 in Basabo, 96 in Anandanagar and 27 in Gulshan. For the purpose of questionnaire survey, each area was divided among 8 groups of investigators. Each group was then assigned with a small cluster within each sub-area. Households were then selected from each cluster following a systematic sampling procedure.

The questionnaire was designed to collect information so as to fulfil the objectives as mentioned above. The questionnaire, therefore, included such aspects as the socio-economic and demographic characteristics of the people affected by the flood, level and duration of the flood, coping mechanism of the people during the disaster phase, sufferings of the people, extent of damage in terms of housing, health, income, occupation, clothing, furniture etc, the extent of recovery after the flood, and help and assistance received from different sources. SOCIO-ECONOMIC ATTRIBUTES A total of 294 household heads were interviewed from five different areas. About 95% of them were males. Majority of the respondents (30.61%) belonged to the age group 30 to 40 years; while 22.55% were in the age group 41 to 50 years or older. 17.68% were in the age bracket 51 to 60 years while the rest belonged to the age group 21 to 30 years.

Nearly 31 percent of the respondents were illiterate, 9.2 percent passed SSC or HSC examinations while 15.6 percent earned Bachelor’s or Master’s degrees. The rest attended schools at primary or secondary levels. Business was the occupation of about 18 percent of the respondents followed by service (15%), petty business (13.7%), rickshaw pulling (13.3%), and daily labor (10%). About 23 percent of the respondents were engaged in various other types of jobs


Sarwar Jahan


including professional activities, and factory work. Nearly 7 percent were unemployed.

The flood affected people of different income groups. Figure 1 presents the distribution of people by income groups. It shows that 19.4 percent of the respondents belonged to the lowest income group having income upto Tk. 2999. Largest number of respondents (40.8%) belonged to the group having income between Tk. 3000 and Tk. 5999 while the respondents earning Tk. 12000 or more constituted 17.6% of the total number of respondents.

There were also spatial variations in income as is evident from Figure 2. The highest average income was recorded in Gulshan, while the lowest was in Meradia. Average income in Kamrangir Char and Ananda Nagar areas were found to be almost the same while the average income in Basabo was a little bit higher than these areas.

Distribution of Respondents by Income

19%

41%22%

18%Upto Tk. 2999

Tk. 3000 to 5999

Tk. 6000 to 11999

Above Tk. 12000

Figure 1: Percentage distribution of respondents by income LEVEL AND DURATION OF FLOOD People in the flood–hit areas were affected in varying degrees depending on the level of flooding. More than 80 percent of the houses went under 3 ft or more water. The water level reached the roof in about 19 percent of the houses and up to half the dwelling height in about 31 percent of the houses.

There were, however, variations among the areas in terms of water level. In Gulshan area the maximum height of water was 3 ft above the plinth level of dwelling units, while in Kamrangir char, water level reached the roof of about 45 percent of the houses. In other areas proportions of houses submerged up to the roof varied between 10 to 17 percent.


The duration of the 1998 flood was one of the highest in recent history. Majority (60%) of the respondents mentioned that their houses remained submerged for more than 60 days while the houses of about 25 percent of the respondents were under water for about 51 to 60 days. Only 10 percent of the respondents mentioned that the duration of the flood was 30 days or less.

0

10000

20000

30000

40000

50000

60000

Income (Taka)

KamrangirChar

Basabo Gulshan

Areas

Average Monthly Hosehold Income

Figure 2: Average monthly household income in different areas

EXTENT OF EVACUATION Although more than 80 percent of the houses went under 3 ft or more water, majority of the people did not leave their houses. The present survey indicates that nearly 32 percent of the people left their homes along with other family members and took shelter in relative’s houses, nearby high–rise buildings or schools or madrashas. Majority (50%) of those who left their homes took shelter in relative’s houses in and outside the area they live in. Most of them used boats for the purpose of evacuation.

About 68 percent of the people did not leave their houses. Nearly 34 percent of those who did not evacuate stayed on the roof of their houses while about 60 percent stayed on an elevated platform inside the house. For the rest (about 6%), the floodwater did not pose serious problems to make such arrangements. People gave different reasons for not evacuating. Majority of them (51%) stayed back home to guard their properties. About 18 percent of the people mentioned that there was no shelter nearby or the available shelter was not suitable for staying. The remaining 31 percent gave various other reasons such as that the house was flood-resistant, problems were not too serious, the family members were ill, etc.


Sarwar Jahan


LIVING CONDITION DURING THE FLOOD People suffered heavily due to the flood. They faced many problems while struggling for survival. The problems such as shortage of drinking water, getting wet by rainwater, shortage of food, possibility of snakebite etc. hit the people quite hard. Most of the people (86.8%) mentioned the shortage of drinking water as the main problem followed by shortage of food (62.5%) and getting wet by rainwater (56.6%). Rain posed serious problems for those who stayed on the roof of their house.

There were, however, spatial variations in the problems faced by the people who did not evacuate. In Kamrangir Char area rainwater and shortage of food were considered as serious problems by 80% and 82% of the people, respectively. This, however, was not unexpected given the fact that Kamrangir Char is a low–income and flood–prone area where most of the houses were submerged. Most of the people here are daily labourers, rickshaw–pullers or low–paid factory workers. Consequently, their jobs and income were badly affected.

Various types of diseases also broke out during the flood. About 76 percent of the respondents mentioned that one or more of the family members suffered from diseases like diarrhea, dysentery, virus fever, jaundice etc. Diarrhea was widespread and nearly 24 percent of the respondents mentioned that one of their family members suffered from this disease. Families of about 17 percent of the respondents had 2 or more members suffering from this disease. There were also an epidemic of virus fever and at least one member of nearly 51 percent of the families suffered from this disease. Dysentery or jaundice also affected about 31 percent of the families.

People also suffered heavily due to increases in household expenditures during the flood. Expenditure on flood, medicine and transportation was considerably higher during the flood than before. Average household expenditure on housing, food, medicine and transportation together was Tk. 7568 during the flood compared to Tk. 6367 before the flood (Table 2) indicating that there was nearly a 19% increase in household expenditure during the flood. The increase in expenditure, however, was not uniform across different items. Transportation expenditure registered the highest increase. The reason for this increase was that people who moved on foot before the flood could not do so during the flood. They had to take rickshaws or boats for moving. Rickshaw fair also increased during the flood as the rickshaw-pullers in most cases could not ply through floodwaters and had to pull the rickshaws by hand. The increase in medical expenditure was about 62% mainly because of the various diseases, which broke out in the flood-affected areas. The increase in food expenditure was, however, modest (15%) compared to transportation and medical expenditures.


Table 2: Average monthly expenditure before and during flood

Average Monthly Expenditure

Items

Before Flood (Tk.) During Flood (Tk.) Difference (Tk.) Housing 1586 1526 -60* Food 3811 4368 557**

Medicine 278 451 173**

Transport 202 1223 1021**

Total 6367 7568 1201**

* Not significant; ** Significant at .01 level COPING WITH THE DISASTER The increase in household expenditures and reduction or loss of income during the flood put many people in a precarious situation. They were compelled to borrow for survival. Nearly 36 percent of the respondents had to take loan for various purposes. Buying food was the main reason for nearly 86 percent of those who took loans. About 28 percent of the people borrowed money for the purpose of medical treatment. People also borrowed money for instant repair of the house (12.1%) during the flood or for renting a house when they had to move to a flood-free area.

The main sources of borrowing were relatives (38.7%), neighbors (22.6%) and friends (11.3%). People (23.6%) also bought food and daily necessities on credit from shop–keepers. It is interesting to note that the proportion of people taking loan from NGOs or Mahajans was very insignificant. Only 6.6 percent of the borrowers got money from the NGOs while another 6.6 percent went to the Mahajans. The results of the study confirm previous findings that largest proportion of the affected people received financial help and credit during and after flood from non-institutional sources such as friends, relatives and neighbors (Elahi, 1988; Hossain, 1990). Some people also sold or mortgaged jewelry mainly to buy food. 8.16 % of the people surveyed sold jewelry while 3.4% mortgaged the same. Almost all of those who sold jewelry mentioned that they did not receive fair price.

Many organizations, however, came forward to provide the flood-affected people with material help. About 34 percent of the people received relief goods from Government (GOs), Non-Government (NGOs) and Voluntary organizations (VOs) as well as various other sources. Figure 3 presents the percentage distribution of people by relief materials received and by sources of such


Sarwar Jahan


materials. The materials they received included food, clothing, medicine and water purification tablets. Some people also received money. It is interesting to note that largest number of people received food, medicine and water purification tablets from voluntary organizations. More people received medicine and water purification tablets from NGOs than from GOs, but the number of people receiving food from GOs was higher than the NGOs. The data once again indicate that flood-affected people received more help from sources other than Governmental and Non-Governmental Organizations.

0

20

40

60

80

100

120

Percent of Households

Mon

ey

Food

Wat

erP

urifi

catio

nTa

blet

s

Med

icin

e

Clo

thin

g

Relief Materials

Others

VOs

NGOs

GOs

Figure 3: Percentage distribution of respondents by relief materials received

and sources of relief materials EXTENT OF FLOOD DAMAGE The 1998 flood caused extensive damage to lives and properties throughout the country. Almost all sectors of the economy were affected. The people in the flood–affected study areas suffered heavily due to the damage to housing, clothing, furniture, job, business, health etc. Table 3 presents the average household damage due to the flood.


Table 3: Average food damage in different study areas

Item

Average damage

(Tk.)

Kamrangir Char (Tk.)

Meradia

(Tk.)

Basabo

(Tk.)

Ananda Nagar (Tk.)

Gulshan

(Tk.) Food 285 480 243 275 191 38 Clothing 455 383 428 226 610 1 Health 694 1238 419 436 626 538 Housing 6277 14514 4143 3616 3277 7186 Furniture 2126 2737 1385 1734 2483 3223 Job 1396 1382 1140 2022 1419 75 Business Income

4253 4930 3946 3688 4313 4040

Overall 15486 25664 11704 11997 12919 15101

The flood caused extensive damage to housing averaging about Tk. 6277 for a household. Loss of business income, damage to furniture and loss of jobs were also quite significant and amounted to Tk. 4253, Tk. 2126 and Tk. 1396 per household, respectively. There were also spatial variations in damage. Households in Kamrangir Char suffered the heaviest damage where average damage was nearly twice that of Meradia, Basabo or Ananda Nagar.

Table 4 presents the distribution of the people by the extent of house-damage and the level of income while Table 5 presents the distribution of people by the extent of income-damage and the levels of their income. About 25 percent of the lowest income people had their houses fully damaged by the flood compared to only 5 percent of the highest income group whose houses were fully damaged. Similar is the picture in case of income-damage. About 43 percent of the lowest–income people suffered total loss of income during the flood compared to about 4 percent of the highest-income group.

Table 4: Distribution of respondents by income-group and extent of housing-damage

Extent of Damage* Level of Income Fully Partly No Damage Total

Up to Tk 2999 14 (24.60) 28 (49.10) 15 (26.30) 57 (19.60)

Tk 3000 to Tk 5999 24 (20.30) 72 (61.0) 22 (18.60) 118 (40.50)

Tk 6000 to Tk 8999 4 (8.20) 28 (57.10) 17 (34.70) 49 (16.80)

Tk 9000 to Tk 11999 4 (26.70) 11 (73.30) 0 (0.0) 15 (5.20)

Tk 12000 and Above 3 (5.80) 38 (73.10) 11 (21.20) 52 (17.90)

Total 49 (16.80) 177(60.80) 65 (22.30) 291 (100.0)

* Number within the bracket represents the percentage.


Sarwar Jahan


Table 5: Distribution of respondents by income-group and extent of income-damage

Extent of Damage* Level of Income Fully Partly No Damage Total

Up to Tk 2999 24 (42.90) 18 (32.10) 14 (25.0) 56 (19.20)

Tk 3000 to Tk 5999 45 (37.80) 51 (42.90) 23 (19.30) 119 (40.90)

Tk 6000 to Tk 8999 8 (16.0) 25 (50.0) 17 (34.0) 50 (17.20)

Tk 9000 to Tk 11999 3 (20.0) 6 (40.0) 6 (40.0) 15 (5.20)

Tk 12000 and Above 2 (3.90) 20 (39.20) 29 (56.90) 51 (17.50)

Total 82 (28.20) 120 (41.20) 89 (30.60) 291 (100.0)

* Number within the bracket represents the percentage.

Chi-square tests were performed to determine whether flood damages were independent of income. The results of the tests are presented in Table 6. The results of the chi–squre tests are highly significant for damages to housing, furniture, business income and job but insignificant for food, clothing and health damages indicating that housing, furniture and income damages could be estimated in terms of income. Table 6: Results of Chi–square tests of independence between income and damages.

Types of Damage Chi-sqare Value Degrees of Freedom

Significance Level

Business Income* 59.0372 24 .00009 Food 26.5032 20 .14983 Furniture* 43.2893 24 .00922 Clothing 25.9242 24 .35702 Housing* 48.58131 24 .01110 Job* 43.0825 24 .00973 Health 25.9076 20 .16888

* Significant ESTIMATING HOUSEHOLD DAMAGE For estimating the curve relating household damage to income, total household damage was obtained combining all types of household damages. It was hypothesized that there existed a positive correlation between total household


damage and household income but a negative correlation between the burden of such damage and household income. The burden of damage was obtained by calculating total household damage per Taka of household income. Computation of Pearson’s correlation coefficients confirmed the hypothesis. The Pearson’s correlation coefficient between total household damage and household income was found to be 0.24 and significant at 99.99% confidence level while the correlation coefficient between the burden of household damage and household income was found to be .20 and significant at 99.7 % confidence level.

For estimating the equation three functional forms were considered: linear, logarithmic and inverse. On the basis of ‘R-square’ and confidence level, logarithmic form was found to be most appropriate for estimating total household damage while inverse functional form was found to be most appropriate for estimating burden of damage. The estimated equations are as follows:

THDAM = -47390+7676 ln (THINCOME)

where, THDAM = Total household damage in Taka THINCOME = Total household income in Taka ln = Natural Logarithm

BHDAM = 1.5057 +8466/ THINCOME

where, BHDAM = Burden of household damage in term of Taka per Taka of Income

THINCOME = Total household income in Taka.

The first equation indicates that total household damage increases with increase in income but at a decreasing rate while burden of damage decreases with increase in income. Estimated curves relating flood damage and burden of flood damage to income are presented in Figs. 4 and 5.

EXTENT OF RECOVERY The present survey was carried out about 3 months after the flood. The respondents were asked if they recovered from various types of damages inflicted by the flood. Table 7 presents the distributions of respondents by the extent of recovery from various types of damages. Majority of the households did not recover even after three month of the flood. In case of housing, only about 25 percent of the respondents recovered completely while nearly 26 percent indicated that they could not do anything about their damaged houses. Percentage of respondents recovering completely from job loss and loss of business income was higher but still less than 50 percent. The situation with respect to health was also not much better, indicating that people needed more help.


Sarwar Jahan


Damage-Income Relationship

05000

1000015000200002500030000

0 5000 10000 15000 20000Income (Taka)

Dam

age

(Tak

a)

Figure 4: Estimated curve relating flood damage and income

Burden-Income Relationship

0

2

4

6

8

10

12

0 5000 10000 15000 20000Income (Taka)

Bur

den

(Dam

age

in T

aka

per

Taka

of I

ncom

e)

Figure 5: Estimated curve relating burden of flood damage and income


Table 7: Percentage distribution of respondents by the extent of recovery from flood damages

Extent of Recovery

Housing Business Income Job Health

Completely 26.85 38.09 48.53 51.06 About 75% 10.36 17.76 12.10 24.73 About 50% 18.39 22.28 15.16 15.50 About 25% 17.02 9.89 10.07 3.61 Not Recovered 27.38 11.98 14.14 5.10

An attempt was made to assess how people tried to cope with the flood

damage. Majority of the respondents mentioned that they tried to face the damage without taking any help from others (Table 8). About 9 percent of the respondents took help of relatives to recover from damage to their houses while about 12 percent of the respondents took help of relatives to make up the loss of business income. People received very little help from the government or the NGOs in their efforts to recover from flood damage. Table 8: Percentage distribution of respondents by sources of help for recovery

Sources of Help Housing Business Income Job Health Own Resource 61.04 63.00 73.68 78.00 NGO 1.96 2.84 0.0 0.0 Neighbours 1.96 2.27 2.63 3.63 Relatives 9.13 12.37 3.95 6.69 Government 0.70 0.0 1.31 3.06 Others 25.21 19.52 18.43 8.62

Nearly 20 percent of the respondents borrowed money from various sources

after the flood. Majority of them borrowed from relative (38.1%). Friends (19%) and neighbors (14.3%) were other major sources of credit. Repairing the damaged house was the main reason for borrowing for nearly 43 percent of the respondents who borrowed money. Nearly 34 percent of the borrowers needed money for buying food reflecting the fact that many people could not recover from the loss of jobs or income. Other reasons for borrowing were treatment of patients, repairing of house etc.

CONCLUSIONS The 1998 flood has left considerable socio-economic impacts in Dhaka city. It has not only damaged houses and infrastructure but also caused considerable


Sarwar Jahan


damage to business, job and health. The findings of the study indicate that the burden of flood damage was borne more by the poor than by the non-poor. The poor suffered heavily due to the loss of employment, housing and property. In many cases, they sold their assets or borrowed heavily for survival.

The study has also shown that the people tried to cope with the damage on their own but the weaker sections of the community had to take help from others as they lost their jobs and income due to the flood. The study reveals that the largest proportion of flood-affected people received material help and credit from non-institutional sources such as friends and relatives, neighbors and voluntary organizations. The results of previous studies carried out in rural areas also corroborate the findings of the present study.

From this study it appears that poverty or low income is a major determinant of flood damage at the household level. Improvement in income and living condition of the people, therefore, would greatly reduce the vulnerability of the population to natural disasters like flood. The government should also play a more active role at different stages of the flood. Post-flood rehabilitation measures assume special importance in view of the fact that poor people need help and assistance to recover from the flood damage. In the absence of any financial assistance or credit facilities from institutional sources, the poor become compelled to depend on friends and relatives. From this study it is evident that financial help or credit from such sources is not sufficient; majority of the affected people could not recover from the flood damage despite getting help from friends and relatives. The vast majority of the flood-affected poor people would suffer more unless the government and non-government organizations come forward to their assistance. REFERENCES DMB (1998) “Report on Bangladesh flood, 1998: chronology, damages and

response”, Disaster Management Bureau, Dhaka. Elahi, K. M. (1988) “ The strategy for living with flood and flood rehabilitation”,

paper presented at the seminar on Floods in Bangladesh: Bangladeshi Views, held on January 24, 25 and 27, 1990 in Dhaka.

Hossain, M. (1990) “Impact of the 1988 flood on the rural economy of Bangladesh”, paper presented at the seminar on Floods in Bangladesh: Bangladeshi Views, held on January 24, 25 and 27, 1990 in Dhaka.

JICA (1990) “Updating study on storm water drainage system improvement project in Dhaka City”, Japan International Cooperation Agency.


Delineation of Flood Damaged Zones of Dhaka

City Based on the 1998 Flood by Using GIS

Mohammad A. Mohit and Shakil Akther Department of Urban and Regional Planning

Bangladesh University of Engineering and Technology, Dhaka-1000, Bangladesh Abstract The 1998 flood in Bangladesh was an unprecedented event of its kind in terms of duration, inundation of areas and damages. In Dhaka City alone, more than 70% of the city area was inundated and about 60% city population was under inundation for about 10 weeks – the longest time in memory. The city experienced colossal loses in housing, infrastructure, industry, commerce and education sectors. The total damage was estimated at Taka 2.0 billion or US$ 41.0 million. The experience of the 1998 flood suggests that the city has to be saved from recurrent floods by adopting both structural and non-structural measures for flood mitigation. Since structural measures are very expensive and time-consuming, non-structural measures such as flood damage zoning may provide a basis for planning disaster mitigation in the city. Considering this, the present study has attempted to delineate the 1998 flood affected city wards into three flood damage zones based on composite damage value derived from five sectors of the city. These flood damage zones are: low, moderate and high. The planning implications of this zoning exercise are – direct development at safer places of the city, and formulation of land use policies and planning standards to guide development in low and moderate flood damage zones so that the city suffers minimum damages from future floods.

INTRODUCTION Although Bangladesh is predominantly a rural country with 75 per cent people living in rural area compared to 25 per cent of urban population, flooding


Mohammad A. Mohit and Shakil Akther


problems are serious in city areas because of high population densities and inadequate drainage facilities. Along with concentration of population, cities of Bangladesh have concentration of large-scale investments relating to housing, infrastructure, industry and commerce. Dhaka, the capital city, as well as the largest administrative, commercial and industrial centre, alone has a population of over 8.0 million which accounts for about 33.0 per cent of the national urban population (World Bank, 1999) Of the total manufacturing establishments, 44.6 per cent are concentrated in Dhaka region which accounts for 50 per cent of the manufacturing sector employment (BBS, 1998). The metropolitan district of Dhaka accounted for 32.1 per cent of GDP in the manufacturing sector including 36.8 per cent of GDP generated by the large-scale industry sector in 1996-97. In total, 16.7 per cent of GDP of the country are generated by the Dhaka metropolitan city (BBS, 1998). Therefore, the potential of loss of GDP due to flood is very high in Dhaka City compared to other areas of Bangladesh. FLOOD IN DHAKA CITY The occurrence of floods in and around Dhaka City can be traced back to as early in 1787-88 when terrible inundation occurred and the streets of Dhaka were submerged to a depth sufficient to admit boats sailing along them (Hunter, 1877). Again in 1833-34, 1870 devastation due to floods were reported (Hunter, 1877). Major floods also occurred in 1954, 1955, 1962 and 1966, which severely affected the city of Dhaka (Rizvi, 1969). Floods that occurred during 1970, 1974, 1987, 1988 and 1998 also affected the city. Among these the floods of 1988 and 1998 were catastrophic. It was estimated that about 77 per cent of city area were submerged to depths ranging between 0.3 to over 4.5 metres and that about 60 per cent of city population were directly affected in the 1988 flood (FAP, 1991). The return period for a 1988 flood was estimated at 70 years but in just 10 years another flood occurred in 1998.

The 1998 flood of Bangladesh was an unprecedented event of its kind in terms of duration, inundation of areas and damages (DMB, 1998). It was estimated that 79 per cent of Dhaka City area were inundated ranging between 0.3 to over 3.0 metres and that about 60 per cent city population were under inundation for about 10 weeks – the longest time in memory. The city experienced colossal loses in housing, infrastructure, industry and commerce sectors. According to DCC estimates, two-thirds of the city roads and 75% of kutcha and semi-pucca houses were affected in the flood. It was reported that 1000 km of city streets, 400 km of drains, 40 km of foot paths, 400 switch points and 1 lock gate were affected and these were estimated at Taka 4.0 billion or US$

Delineation of Flood-Damaged Zones Using GIS


89 million (DCC, 1998). The total damage in housing sector was estimated at Taka 2310.9 million or US$ 48.2 million (Islam, 1998).

Poor drainage has been identified as the principal cause of flooding in the metropolitan areas of Bangladesh. Flooding in Dhaka City is mainly caused by heavy rainfall, drainage congestion, high surrounding water and overflow of rivers. In whatever way flood occurs, it disrupts city life and inflicts major damages. Local flooding due to poor drainage affects 65 per cent of slums and squatter dwellers and 22 per cent of city dwellers are regularly flooded during minor rainfall (FAP, 1991). FLOOD DISASTER MANAGEMENT IN DHAKA CITY Flood disaster management (FDM) in Dhaka City has been attempted through the construction of embankments along the rivers. The Buckland Bund was the earliest attempt to protect the city from the overflows of river Buriganga. The unprecedented flood of 1988 in the country in general and in Dhaka City in particular, led to the adoption of several structural measures to mitigate future flood disaster in the city. Immediately after the flood, the then Government undertook a plan to protect Dhaka from the intrusion of flood water from surrounding areas and drain out internal storm water run off due to local rainfall. Thirteen projects constituting eight types of flood control facilities including embankment (34 km), flood-wall (37 km), sluice (10 nos), pump house (2 nos), canal cleaning (13 nos), road construction (2.2 km), road raising (8.5 km) and restoration of sewerage, at a cost of US$ 142.6 million.

The effectiveness of these measures became evident during the flood of 1998. While the western part of the city remained flood free, the eastern part of Dhaka suffered the severe devastation due to lack of an embankment. Several studies have been done to examine the prospect of constructing an embankment-cum road by-pass at the eastern side of Dhaka. It will be a huge project in terms of monetary involvement and technical feasibility needs to be reassessed.

Flood plain zoning is basically a non-structural approach used to mitigate flood damages as a precautionary measure. Its necessity as a flood mitigation measure was assessed in the French Consortium Study undertaken after the 1988 flood. Nevertheless, none of the FAP study seriously considered it. Flood zoning is an established disaster mitigation measure widely used in countries such as USA and Japan. In Bangladesh, very few studies have been conducted to apply flood zoning as a policy making tool (BUET/JICA, 1987). These studies are mostly confined to national level analysis. In a few cyclone studies, zoning concept has been used to indicate risk and policy formulations (UNDP/WB/GoB, 1993). Urban flood zoning has not been attempted in Bangladesh. However, the



potential of urban flood zoning remains as a non-structural measure for the mitigation of flood disaster in urban/city areas of Bangladesh. OBJECTIVES The main purpose of the study was to delineate flood damage zones, which will provide a basis for flood disaster management of Dhaka City, and hence the following objectives were set for the study:

a. To collect data on the flood damages that occurred in different sectors of the city;

b. To identify the city areas where these flood damages have occurred; and c. To delineate city areas vulnerable to flood disaster and damages.

METHODOLOGY Methodology of the study was developed with due consideration to achieve the objectives of the research. As such the following methodological procedures were adopted. Identification of Flood Affected Areas/Wards

The Dhaka City Corporation office was visited just after the flood to collect information about areas inundated during 1998 flood. A reconnaissance survey was made to gather preliminary information about the flood-affected areas of the city. Ward commissioners were contacted and details of flooding and extent of damages were recorded. These communications enabled identification of 64 wards (out of 90) which were variously affected during the flood of 1998. Assessment of Sectoral Damages

For assessing sectoral damages of the city, following approaches were adopted:

(i) Collection of Official Records: From the zonal offices of DCC, ward-specific damage data were collected on roads and other infrastructures, persons affected, evacuated and sheltered. Despite this, it became necessary to get flood-specific damage data through field survey.

(ii) Questionnaire Survey: A questionnaire was prepared for collecting ward-wise damage data covering five sectors – housing, educational institutions, commerce, industry and roads. In the housing sector, the sub-sectors were pucca, semi-pucca and kutcha houses. In the education



sector, the sub-sectors were primary and secondary schools and colleges. The sub-sector covered in the commerce sector included shops only. Industry sector included both large and small and the roads sector included all local, arterial and collector roads. Ward commissioners were the respondents for the questionnaire survey and so they were interviewed very thoroughly. In addition to questionnaire survey, discussions were held with local leaders in order to crosscheck information and also to record their suggestions about flood mitigation measures.

(iii) Imputation of Damages: Both ratio method i.e., percentage of units damaged and its monetary value in local currency were used to measure extent of damage. Thus, imputation rather than physical unit was used to record sectoral damages for the study.

Data Analysis

Simple statistical tools such as mean, median, mode, min/max values, were utilised to make the data meaningful. Spreadsheet method such as Excel was used for this purpose. PC version SPSS has been used to calculate correlation matrix among the variables. Digitizing of City Map

A copy of the Dhaka City Corporation map showing 90 wards (Fig. 1a) was collected from DCC. This map was digitised at three covers (layers) – (a) a cover (layer) with ward boundaries, (b) a cover (layer) with water bodies, and (c) a cover (layer) with embankment. Preparation of Sectoral Damage Maps

Based on the percentage of units, sectoral damage maps were prepared. These maps show the extent of damages in different sectors and their spatial distribution by wards of DCC. Such an exercise provided the inputs for the zoning exercise. Preparation of Composite Damage Map and Zoning Exercise

A composite damage map was prepared based on the total value of damages of all the sectors combined. This map shows the extent of total damages and their spatial distribution by wards of DCC. Such an exercise has profound planning and policy-making implications. Nevertheless, the present flood damage zoning exercise has been based on the 1998 damage data only; longitudinal data of damages resulting from floods which occurred at different time periods could provide a better data base for flood zoning exercise.



ANALYSIS OF SECTORAL DAMAGES The water of 1998 flood entered Dhaka City area on 15 July 1998 and continued to rise and the total duration of the flood was 3 months or 90 days. Sixty-four Wards out of 90 were inundated to different extent during the flood. Whereas 24 Wards experienced normal flooding with upto 20 per cent area inundated, 9 Wards experienced high-to-severe flooding and 31 Wards suffered catastrophic flooding. In 11 Wards, 100 per cent area was inundated. The extent of flood inundation presented in Fig. 1(b) shows that the city wards lying in the eastern periphery and those in the western periphery but outside the embankment faced severe inundation from the 1998 flood. The minimum depth of flood was recorded at 13cm while average depth was 1 metre and the highest depth was 3.66 metres. Apart from depth, the duration of 1998 flood was exceptional. The minimum duration of 10 days was recorded in only one Ward, but maximum duration was 90 days and average duration was 56 days. Thirty-seven Wards experienced flood duration above the average. The correlation coefficient of depth-duration relationship has been estimated at 0.56. Both depth and duration of flooding inflicted sectoral damage, which has been the concern of subsequent discussion of the study. Housing Sector Damage And Its Spatial Distribution

The housing sector of Dhaka City covers three types of houses – pucca, semi-pucca, and kutcha. 43 per cent of pucca, 53 per cent of semi-pucca and 71 per cent of kutcha houses of the city suffered damages during the 1998 flood. Average damage value of a pucca house was Tk. 8,001, of a semi-pucca house was Tk. 4,838 and of a kutcha house was Tk. 3,156. The city experienced housing damages in different degrees in its 64 Wards. Sixty-three Wards out of 64 experienced damages in pucca, semi-pucca and kutcha houses. While 1 per cent damage of pucca house occurred in Ward-5, 10 Wards experienced 100 per cent damages of pucca houses and 11 Wards experienced damages of pucca houses above the average of 43%. Similarly, there was a 1 per cent damage of semi-pucca house in Ward-68, but 11 Wards had 100 per cent damages of semi-pucca houses and 15 Wards experienced damages in semi-pucca houses above the average of 53%. Again, there was a minimum of 6 per cent damage of kutcha house in Ward-5, but 28 Wards had 100 per cent damage of their kutcha houses and 7 Wards experienced damages in kutcha houses above the average of 72%. It thus appears that kutcha houses and semi-pucca houses were vulnerable during the 1998 flood.

Damage value data presented in Table 1 shows that while 50% of flood affected wards suffered low housing damages and 19% faced moderate damages, 30% of wards faced severe damages of more than Taka 10 million each. The



spatial distribution of affected wards that experienced housing damages [Fig.1(c)] shows that the wards located at the eastern periphery and those outside the embankment at the western and southern periphery faced severe housing damages compared to the wards located at the centre of the city.

Table 1: Distribution of city wards by housing, education, road, industry and shopping sector damages during 1998 flood.

Sector

Damage (Million Taka)

Housing No. of wards

Education No. of wards

Road No. of wards

Industry No. of wards

Shopping No. of wards

No damage 1(1.6) 2(3.1) 5(7.8) 12(18.7) 12(18.7) Upto 2.5ml 17(26.6) 34(53.0) 25(39.0) 34(53.0) 20(31.2) 2.6-5.0 15(23.4) 10(15.7) 7(11.0) 10(15.7) 10(15.7) 5.1-7.5 8(12.5) 10(15.7) 5(7.8) 3(4.7) 12(18.7) 7.6-10.0 4(6.2) 2(3.1) 11(17.2) 1(1.6) 2(3.1) 10.1-above 19(29.7) 6(9.4) 11(17.2) 4(6.2) 8(12.5) Total: 64(100.0) 64(100.0) 64(100.0) 64(100.0) 64(100.0)

Note: Figures within bracket indicate percentage. Damages in Education Sector And Its Spatial Distribution

The education sector of Dhaka City consists of primary schools, secondary schools, and colleges. These sub-sectors suffered damages both due to floodwater and for being used as flood shelters during flood time. Forty-six percent of primary schools, 62 percent of secondary schools and 36 percent of colleges of the affected Wards of the city suffered damages worth Taka 19.0, 8.0 and 3.5 millions, respectively. Average damage values estimated for a primary school, a secondary school and a college were Taka 70,500/-, 82,500/- and 111,300/-, respectively. Ward-wise damages suffered by different sub-sectors were varied. Whereas four wards did not experience damage in primary schools, Ward-6 suffered a minimum of 4% damage, 24 Wards had 100% of their primary schools damaged and an additional 15 Wards had primary schools damaged above the average. In the secondary school sub-sector, 15 Wards did not experience any damage and Ward-6 experienced the minimum damage of 17%, but 29 Wards experienced 100 percent damages each and an additional 6 Wards experienced damages above the average of 62%. In the college sub-sector, 22 Wards experienced damages. Ward-48 suffered a minimum of 13% damage and 16 Wards faced 100% damage; 4 wards suffered damages above average. It thus appears that damages in the college sub-sector were concentrated to 22 (34.4%) wards.



The loses which occurred in the education sector due to 1998 flood (Table 1) shows that whereas 69% wards experienced lower level of damages and 19% wards suffered medium level of damages, 9% of wards suffered high level of damages whose value exceeded Tk 10.0 million each. The spatial distribution of damages in the education sector [Fig. 1(d)] shows that the wards at the eastern periphery and those lying outside the embankment at the southern periphery including some centrally located wards suffered severe flood damages in their education sectors. One reason for centrally located wards being suffered was their involvement in the flood shelter programme.

Infrastructure Damages And Its Spatial Distribution Infrastructure under the control of Dhaka City Corporation consists of roads, drains and footpaths. The 64 Wards, which were inundated during 1998 flood, had 932 km of roads, of which 303 km or (33%) suffered damages (DCC, 1998). The total value of damaged roads has been estimated at Taka 410 million. However, the extent of road damage by wards was quite varied. The minimum road damage of 2% occurred in Ward-78 and the maximum of 100% damages were experienced in 10 wards. 18 wards experienced road damages above the average of 33%. Average damage value of road has been estimated at Taka 1.62 per kilometer. Minimum road damage of Tk. 37,500/km occurred in Ward-37, but highest per kilometer damage of Tk. 8.0 million was reported in Ward-34. 18 Wards experienced road damages above the average of Tk. 6.96 million. 51 kilometers of drains were damaged during the flood and 27 Wards suffered the damage. The highest damage of 12 km was experienced by Ward-19 and the minimum of 0.21 occurred in Ward-20. Average drain damage was estimated at 1.87 km and 8 Wards suffered drain damages above the average. 11 Wards of Dhaka City Corporation suffered damages of their footpaths. The minimum of 0.26 km damage of footpath occurred in Ward-30 and maximum footpath damage of 5 km occurred in Ward-73. Average footpath damage was estimated at 1.8 km and 5 Wards experienced damages in footpaths above the average.

The distribution of total loss resulting from road damages presented in Table 1 shows that a significant number of wards did not experience any damage in this sector. Whereas 32 (50%) wards suffered low damages and 16 (25%) wards suffered medium level of damages, 17% (11) Wards suffered severe damages in this sector. The spatial damage distribution of roads presented in Fig. 2(a) shows that the wards at the eastern periphery and those lying outside the embankment at the southern periphery including a few centrally located wards suffered severe flood damages in the road sector



Figure 1: (a) 1998 Ward map of DCC, (b) Inundation of Dhaka City during

1998 Flood, (c) Housing damages in Dhaka City, (d) Education sector damages in Dhaka City



Figure 2: (a) Road sector damage in Dhaka City during 1998 flood, (b) Industrial sector damage in Dhaka City, (c) Shopping sector damage in

Dhaka City, (d) Flood 1998 damage zones of Dhaka city



Industry Sector Damages And Its Spatial Distribution

Different types and sizes of industries within the Dhaka City Corporation area suffered damages during the 1998 flood. Industrial damages occurred from the loss of capital value – depreciation that has to be compensated through repair and from loss of production due to closure or production loss. Although the former type of loses can be calculated, it is difficult to impute loses arising due to closure of an industrial unit. Thus, an indicative measure has been adopted to reveal the second type of damages. Of the 64 wards that faced flood inundation, 51 wards were reported to have damages in their industrial sector. A total of 11,718 industrial units were reported to exist in the flood-damaged 51 wards, of which 1954 units or 17% suffered damages from flood. Whereas Ward-34, faced the minimum of 1% of its industries suffered, 7 wards experienced 100% of their industries damaged by the flood. On an average, 43% industries were damaged and 21 wards faced damages above the average. The total loss of the industrial sector was estimated at Taka 18.56 million. While Ward-23 suffered only Tk. 5,000/- damage to its industries, Ward-82 faced the highest damage of Tk. 8.0 million in the industrial sector. Average damage value in the industrial sector was Tk. 0.14 million and 7 wards suffered loses in the sector above the average. In addition to capital loses, industries also suffered production loses due to closure. Due to flood, a minimum of 10 days of closure was faced by the industries of Wards 31, 65 and 78, but the industries of Wards 27 and 30 faced the highest 90 days of closure. Average day of closure faced by the industries was 49 days and 24 wards suffered loss above the average due to closure.

The distribution of total loss from industrial damage presented in Table 1 shows that a significant number of wards did not suffer any damage. While 44 (69%) wards suffered loss of upto Tk. 5.0 million and 4 (6)% wards faced moderate loss, 4 wards (6%) suffered large-scale damages of more than Tk.10.0 million each in their industrial sector. The spatial distribution of damages in the industry sector [Fig. 2(b)] shows that the wards lying at the eastern and southwestern periphery including 4 from the central area of the city suffered severe flood damages in their industrial sector. Damages in Commerce Sector and Its Spatial Distribution

The commercial sector of the city experienced flood damages from both loss of capital assets which have to be repaired and also of profit due to closure, i.e., remaining out of business. 51 wards out of 64 suffered lose in the commerce sector consisting of local/ neighbourhood and corner shops. A total of 1,52,550 shops were reported to be in operation in the affected wards, of which 39,760 or 26% suffered loss due to flooding. Total loses suffered were Taka 391 million. Shops affected by flood were a minimum of 1% in Ward-3 but a maximum of



100% damages were experienced in 8 Wards. A total of 21 wards suffered loses in the shopping sector which was above the average of 26%. The monetary loses suffered by the shopping sector varied significantly by wards. Whereas average loss was as low as Tk. 200 in Ward-30, the maximum loss of Taka 40,000 was suffered by Ward-21. A total of 19 wards suffered shopping loses whose amount exceeded the average of Tk.8,811.

The distribution of total loses arising from damages of the shopping sector (Table 1) shows that a significant number of wards did not suffer damages. Whereas 30 (47%) wards suffered damages of upto Tk. 5.0 million, 14 (24%) wards faced moderate damages, 10 (13%) wards suffered large loses of more than Tk.10.0 million each. The spatial distribution of wards which suffered damages in the shopping sector [Fig.2(c)] shows that the wards located at the eastern periphery and those outside the embankment at the western periphery including 1 in the central area suffered severe damages in this sector. Composite Flood Damage And Its Spatial Distribution

The total damages of Dhaka City within the city corporation area during the 1998 flood have been estimated at slightly more than Taka 2.0 billion. Sector-wise distribution of damages presented in Fig. 3 shows that while the housing sector suffered bulk of the damage of 51% followed by roads (20%) and shops (19%), industry accounted for 9% and education only 1%. Ward-wise distribution of composite damage value (Table 2) shows that whereas 45 wards (50%) suffered low damage of upto Taka 25.0 million, 9 wards or 14% suffered moderate level of damage ranging between Taka 25.1 to 50.0 million, 10 wards or 16% suffered severe damage exceeding Taka 50.0 million each.

Pie-ChartFlood Damages by Sectors

51%

1%20%

9%

19%

Housing

EducationRoad

Industry

Shop

Figure 3: Flood damages by sector



Table 2: Distribution of total damages suffered by the Wards of Dhaka City during 1998 flood

Damages (Taka mil.) Number of Wards

Percentage

Upto 10ml 25 39.1 10.1-25.0 20 31.2 25.1-50.0 9 14.1 50.1-100.0 3 4.7 100.1-above 7 10.9 TOTAL: 64 100.0

In order to provide explanations to the damage factors, correlation

coefficients have been calculated with inundation, depth, duration and the sectoral variables. The results presented in Table 3 shows that although flood damages to different sectors of the city are the outcome of inundation, depth and duration of flooding, their contributions to total loss are different. Among the three factors, the contribution of inundation to flood damage is wide and profound compared to depth and duration. While area of inundation has a significantly high positive correlation with road, housing, shopping, industry and education sector damages, flood duration has a significantly positive correlation with housing, road, shopping and industry sector damages, and depth of flooding has significant correlation with housing, road and shopping sector damages. It is evident from Table 3 that whereas housing sector damages were mostly contributed by inundation, duration and depth, education sector damages were contributed by inundation alone. Again, while road and shopping sector damages were contributed by inundation, duration and depth, industrial damages were caused by inundation and duration. Therefore, it appears that area inundation is a prime factor that, if controlled, can significantly contribute towards reduction of flood damages of Dhaka City.

FLOOD DAMAGE ZONING OF DHAKA CITY AND ITS PLANNING IMPLICATIONS Zoning is primarily a grouping exercise adopted to develop typology, which provide a scientific basis for planning or policy decisions. The zoning exercise undertaken to group city wards which suffered devastation during 1998 flood is based on the aggregate or composite value of damages that occurred in the five sectors of the city. Three damage zones – low, moderate and high, could be identified and their distribution is presented in Table 4.



Table 3: Correlation matrix of sectoral damages with area inundation, depth and duration.

Sectoral Damage Inundated area (%)

Depth (Av)

Duration (Days)

Inundation(%) 1.000 .449** .561** Depth (Av) 0.449** 1.000 .290* Duration (Av) 0.561** .290* 1.000 Percent of houses damaged 0.870** .422** .539** Percent of Pucca houses Damaged 0.902** .304* .496** Percent of Semi-pucca H.D. 0.899** .381* .515** Percent of Kutcha H.D. 0.550** .260* .379** Percent of Education Sector Damaged 0.508** .127 .120 Percent of Primary School Damaged 0.441** .134 .086 Percent of Secondary School Damaged 0.370** .076 .111 Percent of College Damaged 0.117 -0.034 .166 Percent of Industry Damaged 0.703** 0.286 .409** Percent of Shops Damaged 0.797** 0.330** .440** Percent of Roads Damaged 0.895** 0.356** .505** Total loss(Taka) 0.514** 0.248* 0.382**

Note: **Correlation is significant at the 0.01 level (2-tailed). * Correlation is significant at the 0.05 level (2-tailed). Table 4: Distribution of Dhaka City Wards by flood damage zones

Flood Zones Ward Numbers Total of Wards

Percentage

No-Damage Zone 1,7,10,12,13,14,16,33,36,43,44,45,47,49,52,56,57,62,69,79,71,72,74,77,81,88

26 28.9

Low Damage Zone 2,4,6,8,9,11,15,17,19,24,29,30,31,32,35,38,39,40,41,42,50,51,53,54,55,58,59,60,61,63,64,65,66,67,68,73,78,79,80,82,83,84,86,87,90

45 50.0

Moderate Damage Zone

5,85,46,76,20,37,25,34,23 9 10.0

High Damage Zone

3,18,21,22,26,27,28,48,75,89 10 11.1

TOTAL: 90 90 100.0

Table 4 shows that 26 (29%) wards of the city are flood-free and hence constitute the no-damage zone in the city. The table also shows that while 45 (50%) wards have low flood damage risk and 9 (10%) wards have moderate sectoral damage risks from flood, 10 (11%) wards are high-risk areas liable to



severe sectoral damages occurring from flood in the city. The spatial distribution of flood damage zones presented in Fig. 2(d) shows that the wards located at the eastern periphery including one in the north and those outside the embankment at the western periphery of the city suffered severe damages. Wards located at the central part of the city along north-south strip are safe from flood damages. It is thus evident that peripheral wards, if not protected by embankment, are more vulnerable to flood damages than inner city wards.

Damage zoning of Dhaka City based on 1998 flood has the following planning implications: (a) Encourage development in the wards belonging to flood free zone; (b) Improve drainage of centrally located wards of the city; (c) Restrict development in the wards belonging to severe flood damage zone of the city; (d) Formulate policies and standards in order to protect development in low flood damage zone; (e) Adopt land use planning and development control measures in moderate flood damage zone; (f) Protection of the eastern part of the city area is essential in order to protect development from future flooding. CONCLUSIONS This research has endeavoured to delineate flood damage zones of Dhaka City based on damages which occurred in five sectors – housing, education, industry, road and shopping, during the flood of 1998. The composite damage analysis led to the identification of three damage zones – low, moderate and high and a no-damage zone of 26 wards of the city. The identification of damage zones has several planning implications, the most important of which is to protect through embankment the wards lying at the eastern periphery of the city from flooding. In the absence of such an embankment, flooding will remain a regular phenomenon in the city. A no-embankment situation will require considerations for non-structural measures to mitigate flood damages. In such a situation, this zoning exercise will provide a basis to direct development to safer wards of the city and formulate land use policies and planning standards in order to guide development in low and moderate flood zones so that the city suffers minimum damages from future floods. REFERENCES BBS (1998), Statistical Year Book of Bangladesh 1997, Dhaka: Bangladesh

Secretariat, p.224.



BUET/JICA (1997) Japan Bangladesh Joint Study, 1997, Flood Plain Zoning Based on Analysis of Flood Damage to Agriculture, Dhaka, BUET-JICA joint study.

DCC (1998) Dhaka City Corporation (DCC), 1998, Dhaka City Corporation’s Post-Flood Recovery/Reconstruction and Development Programme, Dhaka: DCC, p.1-3.

DMB (1998) Disaster Management Bureau (DMB), 1998, Report on Bangladesh Flood 1998, Dhaka

FAP-8B (1991), Dhaka Integrated Flood Protection Project, Final Report, p.1. Hunter, W.W. (1877), A Statistical Account of Bengal, Vol.V, (Districts of

Dacca, Bakerganj, Faridpur and Maimansingh, London: Trubner & Co. (Reprinted in India by D.K. Publishing House, Delhi), p.103.

Islam, N. (1998), “Flood ’98 and the future of urban settlements in Bangladesh”, CUS BULLETIN 35, July-December, 1998, p.3.

Rizvi, S. N. H. (1969), East Pakistan District Gazetteers – Dacca, Dacca: East Pakistan Govt. Press, pp.33-34.

UNDP/WB/GOB (1993) Planning Commission, 1993, Multipurpose Cyclone Shelter Programme, UNDP/WORLD BANK/GOB PROJECT/BGD/91/025, Dhaka, BUET-BIDS.

World Bank (1999), Towards an Urban Strategy for Bangladesh, Dhaka: World Bank, p.3.


The Role of Small Diesel Engines in Rural Bangladesh During the 1998 Flood

Md. Ehsan, Md. Imtiaz Hossain, Md. Nasir Uddin Miah and Md. Abu Sayed

Department of Mechanical Engineering Bangladesh University of Engineering and Technology, Dhaka-1000, Bangladesh Abstract Bangladesh faced its worst flood in the recent history in 1998. During this flood was that the water level got almost stagnant in large areas of the country for a long period of time. This study is a part of the post-flood research works carried out to assess the effect of flood, focusing on small diesel engines used specially in rural Bangladesh. Most of the small diesel engines are used for irrigation purpose in non-electrified areas of Bangladesh. The damages were found to be far less compared to their electrical counterparts in such applications. The study results showed that small diesel engines, although being easily transportable, were still vulnerable to sudden rise of water levels during flood. This was most apparent in areas which were completely inundated, leaving no safe ground. In most cases such engines were damaged only partially, as the users applied various indigenous techniques to protect them. Lack of funds, rather than spares or expertise, was found to be the major cause hindering the repair of partially damaged engines. Although some of the users made alternative uses of the engines, mostly engines were unutilized resulting in loss of productivity. The situation was, however, different in areas moderately affected by flood. In these areas most of the engines were moved to safety and many of them were used for alternative purposes e.g., in mechanized boats. INTRODUCTION Bangladesh faced its worst flood in the recent history in 1998. This devastating flood affected 49 out of the 64 districts of the country. The worst aspect of this




flood was that the water level got almost stagnant in large areas of the country for a long period of time. In some worst effected areas the floodwater remained for a period of up to 74 days, starting from end of June to the end of September, brining catastrophic consequence to life and property. About a million households were damaged or destroyed and crops of about 1.4 million acres of land was ruined. Even conservative estimates showed that the flood affected about 2.5 million people. About 16,000 km of road networks, 4,500 km of river embankment and thousands of culverts and bridges were damaged. Although the water level did not cross the 1988 mark (the other devastating flood in recent history of Bangladesh) in many cases, the long duration made the 1998 flood one of the worst natural calamities the country has ever faced. Preliminary observations have suggested that unusual rainfall both in amount and distribution pattern inside the country, as well as in the upstream areas of the major rivers were primarily responsible for the flood in 1998.

After the flood, various government and non-government organizations carried out research works to assess the losses suffered by the various sectors during the flood. These research works were aimed at getting a clear picture of both short- and long-term losses caused by the flood, as well as suggesting policies/options for avoiding such losses in the event of a future flood. The present research work is such an effort, carried out by Bangladesh University of Engineering and Technology (BUET), where the role of small diesel engines in rural Bangladesh during the 1998 flood was investigated. OBJECTIVES Small diesel engines are used for multiple purposes such as irrigation, rice-mills, power generation and transportation in rural areas of Bangladesh. In regions experiencing quick rise of floodwater many such units have been damaged. On the other hand, many engines have been used for alternative purposes during the flood. The two major objectives of this study were: (i) to assess the damage of diesel engines and the resulting productivity loss in some flood-affected areas, and (ii) to assess the alternative use of diesel engines in the flood-affected areas. METHODOLOGY The project team consisted of one co-coordinator, an advisor and two field researchers. The team carried out the survey from the end of October until mid December 1998 in various parts of rural Bangladesh. Taking into consideration the limitations of time, fund and logistics, especially during the post-flood period,

Role of Small Diesel Engines

the survey work was carried out only in the worst affected regions where floodwater rose significantly above danger level (FFWC/BWDB, 1998). Nearly one hundred and fifty engines were inspected in twelve districts. Figure 1 shows the worst affected regions where engines were inspected. Instead of relying on information supplied from the government offices or company dealers, first hand information were gathered from the users to ensure reliability of data. This also helped in getting a better understanding of the role of these small diesel engines in the rural life of Bangladesh. Preparation of the Questioner

First a model questioner was developed for gathering necessary information and the field research workers visited few engine sites with it. As most of the end users of the small diesel engines are illiterate people, the inspections needed to be carried out in most informal manner and often the field researcher had to fill up the questioner on behalf of the users. Some modifications were made in the questioner, taking into consideration the feedback from the preliminary field visits. The questioner included the following: Location Information

These included: (i) Name of village/town, post office, upazilla, district; (ii) Details of how the engine was mounted, portable/stationary fixtures, etc.; (iii) Brief history of flood in the location, when the water level rose and when receded, rate of rise/fall of floodwater, duration and approximate height of water level, comparison with 1988 flood level, etc.; (iv) Whether the area has electricity; (v) General geography of the location. User Information

These included: (i) Name and address of the user/owner; (ii) When was it bought and in which condition - new/reconditioned/used; (iii) how the maintenance of the engine is done; (iv) The intended purpose of the engine; (v) Alternative use (if any) of the engine during the flood. Engine Information

These included: (i) Rated engine power and speed; (ii) Brand name, model number, country of manufacture; (ii) Number of cylinders and engine dimensions; (iii) Availability of spare parts and maintenance schedule; (iv) Fuel consumption and fuel cost; (v) Repair costs and difficulties (if any).




Figure 1: Location of the districts covered in the study


Data Collection

The field researchers collected data from different worst affected areas of the country. The small diesel engines are mainly used in rural areas where there is no electricity. Due to time and resource constraints, only a limited number of spots could be covered. In spite of these limitations, efforts were made to cover areas throughout the flood-affected region. Data on a total of 147 engines were collected from twelve flood-effected districts.

In the data collection procedure, the first step in each trip was to identify areas where the diesel engines were affected by the flood. Local engine repairing workshops, district BADC offices and public representatives like union chairman/members were found to be good sources of such information. Once the prospective areas were identified, the research engineers visited the engine sites (often in very remote areas which they had to reach on foot or by using country boats) and gathered first hand information form the users, who were very cooperative. The technical data of the engine was taken from its nameplate in most cases. The users also gave information regarding the productivity loss, repair cost and alternative use of the engines during the flood. Figures 2 and 3 show damaged engines at Chandpur and Manikgonj, respectively. Figures 4 and 5 show alternative uses of engines during the flood. FINDINGS OF THE STUDY The results of the survey are presented in Figs. 6 through 12. Figures 6 through 8 show different aspects of the 1998 flood in the worst affected areas. Nearly half of the engine sites were flooded for 8 to 12 weeks. The peak water level in the flood affected areas reached as high as 20 feet from the ground. Although the water level reached unusually high levels, in many places these were similar to the levels reached during the 1988 flood. Although the people of Bangladesh are used to limited flood during the rainy season, it was the unusually long duration of the flood that caused the devastation in 1998.

Figures 9 through 12 show the effect of flood on the small diesel engines throughout rural Bangladesh. Figure 9 shows the district wise distribution of the inspected engines. The number of engines affected by the flood increased as more and more area became flooded, starting from the beginning of August and continuing to the end of September. During this period almost all the sites were affected.

Most of the engines used had rated power of 5-10 hp with rated speed near 2000 rpm. Although having good transportability, more than 80% of the engines suffered at least partial damage during the flood period. This was partly due to unexpected quick rise of water level, which the owners were not prepared for, and partly due to the unavailability of any dry ground at all in many sites.




Figure 2: Inspection of a damaged engine that was used for irrigation in

Chandpur

Figure 3: A partially damaged engine in Manikgonj


Figure 4: Alternative use of an engine in a rice mill in Narshindgi

Figure 5: Alternative use of an engine in a rice mill in Patuakhali




0

10

20

30

40

50

60

70

80

< 6weeks

6-8weeks

8-10weeks

10-12weeks

Flood Duration

No.

of E

ngin

es /

Perc

enta

ge S

ites No. of Engines

Percentage ofSites

Figure 6: Duration of the 1998 flood at the inspected sites

0

10

20

30

40

50

60

70

upto 8' 8' - 12' 12' - 16' 16' - 20'Highest water Level (ft)

No.

of S

ites

Insp

ecte

d

Figure 7: Peak water levels recorded at the inspected sites


0

20

40

60

80

Upto 3 fthigher

Upto 1 fthigher

Same as1988

Below1988 level

No.

of S

ites

Insp

ecte

d

Figure 8: Comparison of the flood situation in 1998 with that in 1988

0

5

10

15

20

25

Bram

mon

baria

Cha

ndpu

r

Dha

ka

Man

ikgo

nj

Nar

ayan

gonj

Nat

ore

Nor

shin

gdi

Now

abgo

nj

Pabn

a

Patu

akha

li

Raj

shah

i

Sira

jgon

j

No.

of E

ngin

es In

spec

ted

Figure 9: District-wise distribution of inspected engines




0

10

20

30

40

50

60

upto 5hp

5 - 10hp

10 - 15hp

15 - 25hp

above25 hp

Rated engine power (hp)

No

of e

ngin

es in

spec

ted

Figure 10: Distribution of rated engine power

Not damagd

8%

Fully damagd

10%

Partly damagd

82%

Figure 11: Damage analysis of inspected engines

Most of the small diesel engines are used for irrigation purpose in rural areas

without electricity. These are seldom used in areas with electricity, where electric motors running deep tube-well pumps are common. Such electric motors suffered even more damage during the flood. Most of the diesel engines in worst affected areas were not used during the flood (see Fig. 12), although at some places these engines were used for different purposes (see Figs. 4, 5). The main alternative


use has been in small rice mills. But most of the partly effected engines could not be repaired during the flood and many were still out of order due to lack of funds. In worst affected areas, cases were recorded where engines were saved by placing them on top of boats, but could not be used as the owner with his family was also stranded on it. This situation was however different in moderately affected regions were such engines were effectively used in mechanized boats and for other purposes. In this study, the estimates of loss of productivity have been calculated from the owners/users information.

Ultimate use of inspected engines during flood

Not used93%

Used7%

Figure 12: Ultimate use of inspected engines

The important findings from the survey carried out can be summarized as the

follows: (i) The study was limited to the worst affected areas where large rise of water level was experienced; (ii) A total of 147 engines located in 12 districts were inspected in this study throughout the worst flood-affected areas; (iii) The duration of the flood was different at different places; (iv) Water levels as high as 20 ft above ground level was recorded; (v) The 1998 flood exceeded the 1988 flood mainly in respect of duration; (vi) Diesel engines are mostly used in rural having no electricity; (vii) Most of the engines are used for irrigation purpose, although there are some other uses; (viii) The small diesel engines mostly have rated power of less than 20 horse power; (ix) The price of fuel varies from Tk. 13.7 to 15 per liter, depending on location; (x) These engines are in most cases maintained by mechanics from nearby town and spare parts are generally available; (xi) Most of the engines are imported and Chinese brands constitute the bulk of these engines; (xii) Although these engines generally have easy transportability, most (above 80%) suffered partial damage during the rapid rise of floodwater in the worst affected areas; (xiii) Although for most engines, the




degree of damage was limited and some were effectively protected from flood, in worst affected areas only about 7% of the engines could be used for productive purposes during the flood. The situation was different for moderately affected regions where such engines were effectively used in mechanized boats and for other purposes; (xiv) Repair costs for most of the partly damaged engines have been estimated at about Tk. 2000-5000/-. In many cases, lack of availability of funds for repairing kept the engines out of order even after the flood; (xv) Approximate daily productivity of most of such engines ranged between Tk. 100-300/- per day, which was lost in most cases during the 1998 flood. CONCLUSIONS The small diesel engines although being easily transportable, are still vulnerable to sudden rise of water levels during flood. This is most apparent in areas completely inundated, leaving no safe ground. In most cases such engines were damaged only partially, as the users used various techniques to protect them. Many of the damaged engines could not be repaired, even after the flood, mainly due to lack of funds. Most of the engines in the worst effected areas were out of use during the flood resulting in loss of productivity. Some of the users however made alternative use of the engines, mostly in small rice mills. The situation was different in areas moderately affected by the flood. In these areas most of the engines were moved to safety and many of them had alternative uses especially in mechanized boats. Most of the small diesel engines are used for irrigation in rural areas without electricity. The damages to the diesel engines were found to be far less compared to those suffered by electric motors that were in use in the flood-affected areas with electrical connection. REFERENCES FFWC/BWDB (1998) Total Flood Situation 1998, Bangladesh Water

Development, Dhaka.


Damage and Productivity Loss in Industries During the 1998 Flood

Abu Md. Azizul Huq, Md. Imtiaz Hossain, A.K.M. Sadrul Islam,

Md. Arif Hasan Mamun, Mosfequr Rahman and Md. Obaidul Gani Department of Mechanical Engineering

Bangladesh University of Engineering and Technology, Dhaka-1000, Bangladesh Abstract This paper presents estimates of damage and loss in productivity in engineering and manufacturing industries during the 1998 flood. Data on loss and damage were collected for public and private sector industries. Some data were collected directly from the industries and some from Ministry of Industries, FBCCI, DCCI and different corporations. The study shows that industries located in and around Dhaka division suffered most during the flood. In the private sector, maximum loss occurred in leather industry, while among the public sector industries, jute sector suffered the maximum loss. INTRODUCTION


The 1998 flood continued from July to September. The flood affected large areas of 13angladesh and caused damage not only to agriculture, housing, and livestock but also to public and private sector industries in the form of damage to buildings, raw materials, plant and machinery, etc. In addition, production in many types of industries such as steel industries, sugar and food industries, jute mills, textile mills and cottage industries were adversely affected due to disruption of work during the flood. Some industries incurred losses due to breakdown of transportation infrastructure due to flood. Production was also affected during post-flood rehabilitation work, which again affected the economy of the industries. The present study aimed at collection and analysis of data of all types of losses that occurred during the flood in different manufacturing and engineering industries in different parts of the country.

A.M.A. Huq, M. I. Hossain, A.K.M.S. Islam, M.A.H. Mamun, M. Rahman and Md. O. Gani


The major objectives of this study are: (1) to assess damage in manufacturing and engineering industries in public and private sectors; (2) to assess productivity loss in manufacturing and engineering industries; (3) to compare damage and loss in productivity in different manufacturing and engineering industries; and (4) to compare division-wise loss in different industries. METHODOLOGY The methodology involved collection of flood data, identification of industries for the purpose of data collection, and collection of data on damage and loss in productivity both for private sector and public sector industries. Due to limitation of time and the resources, the research team visited few industries for collection of data and major portion of the data were collected from secondary sources. Figure 1 shows different data collection sources. Table 1 shows the secondary sources of data collection.

All types of industries where technology is used to improve, service and manufacture different types of engineering parts and materials have been classified as engineering industries. Public sector engineering industries include dockyard and shipyard of Bangladesh Steel and Engineering Corporation, Directorate of Dredger of BWDB (Narayanganj), Bangladesh Inland Water Transport Authority and Corporation, etc. Private sector engineering industries include automobile servicing, light engineering industries, electronics service industry, machine parts manufacturing, industrial fittings manufacturing, metal and works, etc.

All types of product manufacturing industries are grouped in to this category. Manufacturing industries are classified into different types of industries, such as chemical industry, food industry, garments industry, jute industry, plastic industry, textile industry, small and cottage industry, pharmaceutical industry, leather industry. In the public sector, BCIC (Bangladesh Chemical Industries Corporation), BJMC (Bangladesh Jute Mills Corporation), BTMC (Bangladesh Textile Mills Corporation) represents chemical, jute and textile industries, respectively. But the major portion of Garments, Jute and Textile industries are in the private sector.

For the purpose of data collection, a questionnaire was prepared for obtaining detailed description of flood damage and productivity loss suffered by different industries. The research team visited a number of industries and filled the questionnaire. In some cases, the questionnaire was filled by the respective industries. In engineering industries, additional data were also collected. Different Corporations and Associations also supplied data to the research team.

The data on damage have been presented according to different fields to

Damage and Productivity Loss in Industries


which damages were incurred, such as structural damage in building walls, damages in land, plant and machinery, raw materials, finished goods and others. For manufacturing industries, a fall in production performance from the average rate during the flood was considered as productivity loss. It level of average rate was set by the respective industries. For engineering and service industries, the operation loss was considered as productivity loss.

Project

Manufacturing Industries

Engineering Industries

Chemical Industry Garments Jute Plastic Textile Small and Cottage Food Pharmaceutical Leather Miscellaneous

(1) Steel and Engineering Corporation: (a) Khulna shipyard Ltd. (b) Narayangonj dockyard

(2) Dredging: Dredger unit of BWDB (3) Inland Water Transport:

(a) BIWTA (b) BIWTC

Flood related information from FFWC, BWDB

Figure 1: Diagrammatic representation of different data collecting sources



DAMAGE AND PRODUCTIVITY LOSS IN ENGINEERING INDUSTRIES

Public Sector Engineering Industry

Bangladesh Steel and Engineering Corporation Table 2 shows damage in Khulna Shipyard Ltd. during the 1998 flood. The table shows that the highest loss in revenue is due to discontinuation of production and sale. Plant and machineries accounted for the second highest loss. Table 3 shows damage in Dockyard and Engineering Works Ltd. It shows that maximum loss occurred in revenue followed by that in land.

Bangladesh Water Development Board (Dredging) Table 4 and 5 show damage and production loss, respectively, in Directorate of Dredger of Bangladesh Water Development Board (BWDB) at Narayanganj. Table 6 shows approximate production /revenue loss due to flood in Directorate of Dredger of BWDB at Narayanganj.

Inland Water Transport The amount of damage in different divisions of Bangladesh Inland Water Transport Authority (BIWTA) was evaluated (Huq et al., 1999). It was found that highest damage was incurred in Narayanganj division followed by that in Dhaka division. Damages in Barisal and Khulna were also high. Aricha and Chandpur suffered damage worth about Tk. 100 lac. Damage in Patuakhali, Sirajganj and Mawa were comparatively lower.

Damage suffered by different field units of BIWTA was also evaluated (Huq et al., 1999). It was found that the dredge unit suffered the maximum amount of loss amounting over Tk. 900 lac. Conservancy also suffered heavy damage, exceeding Tk. 600 lac. Hydrography and launch stations incurred comparatively lower damages, less than Tk. 200 lac. In the field of structure and equipment of BIWTA, machinery and equipment suffered the maximum damage, amounting to about Tk. 80 lac. Structural damage in Narayanganj was around Tk. 60 lac. and in Khulna about Tk. 34 lac. The amount of loss in Dhaka division was higher. DAMAGE AND PRODUCTIVITY LOSS IN MANUEACTURING INDUSTRIES

Public Sector Chemical Industries

Loss incurred due to the 1998 flood by the Zia Fertilizer Factory, North Bengal Paper Mills and Sylhet Pulp and Paper Mills were evaluated. Loss suffered by the Sylhet Paper and Pulp Mills was very high, close to Tk. 120 lac. In comparison, the other two industries suffered negligible loss. This is primarily because Sylhet was highly affected by the flood.



Table 1: Secondary sources of data collection

Sl. No. Name of Sources 1 Ministry of Industries 2 Bangladesh Steel and Engineering Corporation 3 Bangladesh Sugar and Food Industries Corporation 4 Bangladesh Small and Cottage Industries Corporation 5 Bangladesh Textile Mills Corporation 6 Bangladesh Jute Mills Corporation 7 Federation of Bangladesh Chamber of Commerce and Industries 8 Dhaka Chamber of Commerce and Industry 9 Narayangonj Chamber of Commerce and Industry

10 Bangladesh Textile Mills Association 11 Bangladesh Specialized Textile Mills and Power loom Industries

Association 12 Bangladesh Jute Mills Association 13 Bangladesh Jute Spinners Association 14 Bangladesh Jute Association

Table 2: Damage in Khulna Shipyard Ltd. during 1998 flood

Sl. No.

Field of Damage

Financial Estimate

(Lac Taka)

Land (a) Land: Silt deposition on slipway 5.00

1

(b) Co-way wall: Collapse of co-way wall due to strong current

20.0

Building and Factories 2 Factories, administrative and residential buildings have been submerged and damaged

20.0

Plant and Machineries 3 Damage of machineries due to inundation 30.0

4 Raw Materials 0.00 5 Loss in Revenue Due to discontinuation of production and sale 100.0

Total 175.0 Source: MIS, Bangladesh Steel and Engineering Corporation



Table 3: Damage in dockyard and engineering works ltd.

Sl. No.

Field of Damage

Financial Estimate

(Lac Taka) Land 1 Ineffectiveness of slipway due to inundation and siltation 50.00 Building and Factories 2 Damage of roads 5.00 Plant and Machineries Damage of machineries and electric motor 5.00

3

Damage of electric line 26.5 4 Raw Materials 0.00 5 Loss in Revenue Due to discontinuation of production and sale 62.5

Total 149.0 Source: MIS, Bangladesh Steel and Engineering Corporation Table 4: Damage in Directorate of Dredger of BWDB at Narayanganj during 1998 flood (structural damage)

Sl. No.

Field of Damage Financial Estimate (Lac Taka)

1 Road near dockyard 0.50 2 Road of barrack 2.00 3 Road from riverside to main gate 7.00 4 Storage house 2.00 5 Security road with break soling 10.0 6 Boundary wall 0.50 7 Road between two workshops 2.00 8 Sewerage line 2.50 9 School building 0.90

10 Bachelor barrack 0.70 Total 28.10



Table 5: Production loss of Directorate of Dredger of BWDB, Narayangonj (Repair and Production Division)

Sl. No.

Month of Production

Actual Production (Lac Taka)

Average Production (Lac Taka)

Loss of Production (Lac Taka)

1 July, 1998 10.37 12.00 1.63 2 August, 1998 7.97 12.00 4.03 3 September, 1998 5.42 12.00 6.58 4 May, 1998 12.93 12.00 -- 5 June, 1998 12.85 12.00 --

Total 12.24 Source: Respective divisions of Director of Dredger, BWDB, Narayangonj Private Sector Chemical Industries

Loss suffered by the private sector chemical industries on different accounts (e.g., land, building raw materials, plant and machinery, production loss, etc.) was evaluated. It was found that majority of loss was due to production loss (about Tk. 800 lac). This was followed by loss in plant and machinery (about Tk. 250 lac), finished goods (over Tk. 200 lac) and raw materials (about Tk. 100 lac) and buildings (about Tk. 100 lac). Loss attributed to land was minimum. Ready Made Garments and Knitwear Industries

Loss suffered by the garments and knitwear industries on different accounts (e.g., land, building raw materials, plant and machinery, production loss, etc.) was also evaluated. Loss due to production stoppage was the highest for this sector and amounted to about Tk. 9000 lac. For many industries, land, building, plant and machinery were under water during the flood, causing loss in production for prolonged periods of time. Hosiery Industries

For hosiery industries flood damage of equipment was small. Damage in circular knitting machine was slightly higher. Flood damage in dying and processing was the maximum, amounting to about Tk. 1500 lac. Public Sector Jute Industries

Public sector jute industries suffered mostly due to disruption in export of finished products. During the 1998 flood very few finished goods could be exported to foreign countries that caused a large amount of loss in this industry. Flood damage and productivity loss in public sector jute industries located in



different zones (Adamjee, Chittagong, Dhaka and Khulna) were also evaluated. Among these zones, Dhaka suffered the most loss, about Tk. 160 crore, followed by Adamjee (about 100 crore) and Khulna (about 40 crore). This is because maximum jute mills are located in Dhaka zone. Jute mills in Chittagong remained unaffected. Table 6: Production loss/revenue loss (approximate values) due to flood

Name and type of Dredge

Name of Project/Position

During Flood

Days of Dredging Stoppage

Loss of Dredging

Hours

Loss of Production

(cu. m.)

Loss of Revenue

(Lac Taka)

S.D.Testa-18” M/s Siddique Textile, Kanchpur

75 450 9000 6.75

S.D.Kasalong-18” 210 MW, TPS, Siddirgonj, Narayangonj

75 450 9000 6.75

S.D.Kumar-18” Kirtinasha River Dredging, Sariatpur

45 270 54000 40.5

S.D.Dudkumar-18” Basundhara Papar Mills

70 420 84000 63

S.D.Dhaleswari-18” Akij Cement Factory, Chatak, Sunamgonj

80 480 96000 72

S.D.Bangshi-12” Bancharampur Degree College yard filling work

45 270 10800 8.1

S.D.Karnafully-12” Dumudda Channel Dredging, Sariatpur

60 360 14400 10.8

S.D.Surma-12” BSIC Industrial Area, sariatpur

60 360 14400 10.8

S.D.Dharla-12” Sinha Textile Ltd., Kanchpur, Narayangonj

75 450 18000 13.5

Total 232.2 Loss of revenue due to idle hour of the attending plants (30%) 69.66

Grand Total 301.86 Private Sector Jute Industries

Evaluation of damage and productivity loss in private sector jute industries showed that production loss and loss of raw materials (primarily jute) were very high, amounting to about Tk. 2500 lac. Since jute is biodegradable, large amount of jute rotted during the flood, which resulted in high loss in production.



Plastic Industry

Analysis of flood damage and productivity loss in private sector plastic industries showed that production loss accounted for majority of loss, which stood at about Tk. 450 lac. In plastic industries plant and machinery are very expensive, so minor damage in plant and machinery resulted in relatively high loss. Public and Private Sector Textile Industries

Analysis of flood damage and productivity loss in public sector textile industries showed that production loss accounted for the highest loss, amounting to about Tk. 400 lac.

Loss suffered by the private sector chemical industries on different accounts (e.g., land, building, plant and machinery, raw material, production loss, etc.) was evaluated. It was found that loss of production accounted for the maximum loss (around Tk. 8500 lac) in this sector; on the other hand, loss due to damage of land was minimum. Bangladesh Small and Cottage Industries Corporation

Flood damage and production loss in small and cottage industries within BSCIC industries estates were evaluated. The maximum flood damage occurred in equipment (around Tk. 850 lac) because the equipments were under water during the prolonged flood of 1998. Discontinuity of production also accounted for significant loss. This was due to unavailability of workers (whose houses became inundated during flood) and inundation of equipment. All these caused loss of production (amounting to about Tk. 400 lac). The damage of raw material and building were around Tk.200 lac and Tk. 300 lac, respectively.

Division wise flood damage and productivity loss in the small and cottage industries were also evaluated. It was found that Dhaka division suffered the maximum amount of loss in this sector (around Tk. 1800 lac). This was followed by Chittagong (close to Tk. 400 lac), Rajshahi (over Tk. 200 lac), Khulna, and Barisal. Sylhet division basically remained unaffected. The maximum flood damage occurred in Dhaka division primarily because of the large number of small and cottage industries in this division. Public Sector Food and Sugar Industries

In public sector food and sugar industries, there was no damage or production loss, as the flood period did not coincide with the sugar production season. But due to flood and water logging, sugar cane plantation within mill areas was severely affected.



Private Sector Food Industries

In private sector food industries, the major loss was due to loss of production and damage of raw materials. Production loss was estimated at Tk. 9000 lac and raw materials over Tk. 3000 lac. Minor losses were also caused by damages in buildings, plant and machinery, finished goods, etc. Pharmaceuticals Industry

Damage and productivity loss in private sector pharmaceutical industries due to the 1998 flood were evaluated. It was found that the maximum loss occurred in the area of finished goods of these industries amounting to about Tk. 2500 lac. Loss in raw materials was about Tk. 700 lac, and that in plant and machinery was close to Tk. 500 lac. Land and buildings of these industries were not very much affected by the flood. Leather Industry

Analysis of flood damage and productivity loss in private sector leather industries showed that the only loss suffered by these industries was due to disruption in production during the flood. It was found that losses due to damages in land, building, finished goods, plant and machinery, etc. were negligible.

ANALYSIS OF FLOOD DAMAGE IN INDUSTRIAL SECTORS

Damage Loss

Analysis of damage in private sector industries showed that loss suffered by the textile industries was the maximum (around Tk. 8500 lac). This was followed by food industries (over Tk. 7500 lac) and engineering industries (about Tk. 7000 lac). Most of the textile, food and engineering industries are located in the suburb area of flood affected districts, most of which were affected by the prolonged flood of 19998. So the loss in the textile, food and engineering industries were higher. Other industries suffering significant damages include (in descending order of damage loss) jute, pharmaceuticals, garments, leather, chemical, and plastic industries.

Productivity Loss

Productivity loss in private sector industries was also evaluated. It was found that in terms of productivity loss, leather sector suffered the most and the loss amounted to about Tk. 22000 lac. This was followed by textile and food industry (around Tk. 1000 lac). Engineering and garment industries also suffered significant productivity loss. Productivity loss suffered by jute and pharmaceuticals industry was moderate.



Damage loss and loss in productivity in private sector industries on different accounts were evaluated. It was found that loss suffered due to loss of production was the maximum (around Tk. 65000 lac). This analysis indicates prolonged production disruption due to prolonged flood of 1998 causes maximum production loss. The damage loss on account of raw materials, plant and machinery and finished goods were more or less similar (around Tk. 5000 lac). The damage loss due to building and land was relatively small. Public Sector Industry

In the public sector, jute industry incurred the maximum amount of loss. Loss in engineering industry was lower, around Tk. 5000 crore. Textile industries incurred comparatively lower loss, of the order of Tk. 2000 crore. Loss in chemical industries was low. Private Sector Industry

In private sector, leather industry incurred maximum amount of loss. It was followed by food, textile, engineering and garment industries. Loss in pharmaceutical, miscellaneous and chemical industries were lower. Loss in jute and plastic industry was comparatively very low.

CONCLUSIONS AND RECOMMENDATIONS From the analysis and discussion the following conclusions can be made: (1) Maximum loss in engineering and manufacturing industries during the 1998 flood has been incurred in and around Dhaka division; (2) Productivity loss was higher than damage loss for all types of manufacturing industries; (3) Among private sector industries, maximum loss was suffered by the leather sector; (4) Among public sector industries maximum loss was suffered by the jute sector, (5) For industries within BSCIC estates, maximum loss was incurred in the field of machinery and equipment for industries located in and around Dhaka division.

From the above analysis the following recommendation can be made: (1) In Dhaka division major industrial zone is Narayangonj and Tongi. Narayangonj can be protected from flood by construction of dam like the DND dam; (2) Industrial units should be located at proper elevation (above flood level). Entrepreneurs should be conscious in this regard to avoid massive loss during future events of flood; (3) To assure continuation of production during any disaster like flood following measures could be taken: (i) continuous supply and preservation of raw materials, (ii) Proper storage, distribution and export of finished goods (iii) proper transportation of worker to the industry; (4) Owners may motivate the worker by constricting worker's staff quarter inside the industry



above flood level or by giving extra payment for transportation; (5) Every industry should take proper measure for protection against flood; (6) Proper measures should be taken for transportation of finished goods, raw materials, etc. during flood. REFERENCES Huq, A.M.A., Hossain, M.I., Islam, A.K.M., Mamun, M.A.H., Rahman, M. and

Gani, M.O. (1999) Damage and Productivity Loss in Industries During the 1998 Flood, BUET.


Solar Powered Lantern for Flood Affected Areas

Md. Quamrul Ahsan and M. Alam

Department of Electrical and Electronic Engineering Bangladesh University of Engineering and Technology, Dhaka-1000, Bangladesh Abstract Lighting is an essential element of human civilization. It is quite difficult to provide electricity to people living in all the parts of the country due to economic and technical reasons. The situation usually worsens at the aftermath of natural calamities like flood, cyclone, etc. A solar powered lantern as a lighting system has been proposed in this paper, in an effort to minimize the sufferings of flood-affected people living in isolated parts of the country. The design and construction principle of this lantern is presented in the paper. It investigates the performance characteristics of the proposed lantern. The paper also presents the comparison of the proposed lantern with the conventional lighting system, hurricane lantern and candle, in terms of cost and performance. INTRODUCTION Bangladesh being a low-lying country is a flood prone area. Due to incessant shower in the rainy season or due to the effectd of En-Nino and Tsunami, the country often experiences deluge during the rainy season and many areas become inundated under floodwater. The power supply in the flood-affected areas is also disrupted. An alternative source of electricity, thus, may be used for an isolated rural home in such conditions. The operation of the alternative source should be less sophisticated so that rural people can easily operate it.




Photovoltaic (PV) cells may be an alternative source for an isolated home lighting since it does not require a complicated technical system for operation. Moreover, its input is available at every place, as long as sunlight reaches there. As the people are becoming more concerned about environmental pollution, the researchers are putting renewed emphasis on the use of PV cells. Over the last two decades, researchers have developed a large number of techniques (Bishop, 1989; Molenbrock et al., 1991; Pellegrini, 1991) to improve the performance of PV cells. The investigation on the use of PV cells for the isolated energy sector is also getting increasing importance (Chakma et al., 1997; Alam et al., 1998).

This paper presents an application of PV cells for an isolated home lighting. It proposes a solar powered lantern to meet the lighting system of a house. The basic components of this type of lantern are: (i) charging controller, (ii) one rechargeable battery, (iii) low voltage protection circuit, (iv) an inverter and (v) compact fluorescent lamp (CFL). The inverter circuit is properly designed so that its output ac voltage is maintained at an appropriate level. To control the charging of the battery the output of the solar panel is fed to the battery through c control circuit.

This paper investigates the real life performance of the proposed lantern. Accordingly, it estimates the number of lanterns required for a standard rural home. The paper also presents a comparison between the proposed lantern and other alternatives, conventional hurricane lantern and candle, in terms of performance and cost.

SOLAR POWERED LANTERN The source of energy of the proposed lantern is the electricity produced by a solar panel. The output illumination of this lantern is produced by its compact fluorescent lamp. The output of the solar panel is a dc voltage, while the required input for the lantern is ac. The energy from the panel is stored in a rechargeable battery. A voltage control circuit controls the charging of the battery. An inverter circuit is used to convert the dc voltage into ac and the ac voltage is fed as an input to the lantern. A low voltage protection circuit is incorporated to prevent the battery from deep discharging.

The rechargeable battery is placed inside the casing of the lantern such that it can be taken out of the casing for charging or can be placed inside the casing easily. The schematic of a solar powered lantern along with the solar voltage regulator is given in Fig. 1.

The locally available solar panels are mostly of two types: one having output voltage of 10V (6.5W) and the other 20V (10W and 43W). The commonly available rechargeable battery is of 6F and compact fluorescent lamp of 5W and



9W. Therefore, in a 10V panel a single battery may be charged while in a 20V panel two batteries may be charged simultaneously.

Figure 1: Schematic of a solar powered lantern Construction

The principle that is followed in the construction of a solar powered lantern is that only those components are selected which are locally available. For the solar powered lantern only the voltage controller, low voltage protection circuit and inverter are designed and fabricated. In the following sections, the constructional details of voltage control circuit, low voltage protection circuit and inverter are presented. Charging Controller

Figure 2 presents the connection diagram of a voltage control circuit. The main function of this unit is to charge the battery at an appropriate voltage and to ensure that the charging is stopped as soon as the battery attains the required voltage. In Fig.2, the relay operates when zener diode (ZD2) starts conduction in the reverse direction. This situation occurs when each battery is charged with a pre-defined voltage, Vbat. The operation of the relay causes the disconnection of the battery from the supply source, the solar panel. That is, the charging process is stopped. The LED is incorporated in the circuit only to indicate the on/off mode of the charging process. The battery and the parallel branch containing LED get disconnected from the source simultaneously. That means when the LED is “off” the battery is not in the charging mode. Low Voltage Protection Circuit

Figure 3 shows the connection diagram of a low voltage protection circuit. The main function of this unit is to monitor the battery voltage under loaded condition and to ensure that the discharging is stopped as soon as the battery voltage drops to a preset low voltage level and, thus, prevents the battery from deep discharging.



Figure 2: Voltage control circuit The circuit shown in Fig. 3 consists of timer, switching device and low

voltage sensor. IC 555 is an integrated circuit timer. Here IC 555 is connected in the monostable mode. When a negative pulse is applied to pin 2, the output goes high and terminal 7 removes a short circuit from capacitor C4. The output remains high for a time given by thigh = 1.1 R2C4

Figure 3: Low voltage protection circuit



The high output in pin 3 is inverted by the transistor logic inverter comprising R6, R7 and T3. The zener diode along with its series resistance forms the low voltage sensing part. The zener voltage is chosen in such a way that its zener breakdown voltage, Vz is equal to 80% of Vin, where Vin is the input voltage. If Vin is less than Vz, the transistor T3 remains off and T1 remains on; so trigger input pin 2 is shorted to ground. Thus total input voltage appears at terminal 3, which is logically inverted by transistor T3. The corresponding low output at the collector terminal of transistor T3 isolates the externally connected inverter circuit from the battery and thus prevents battery from deep discharging.

The inverter circuit is a standard one. It converts 6.7-volt dc to 215-volt (Peak to peak) ac. The main components of an inverter are a transformer, a H1061 transistor and a capacitor. The transformer has a turns ratio of 18/ 300 (for 9W inverter) and 18/160 (for 5W inverter) with a 2:1 tapping in the primary. A view of control circuit, inverter circuit along with low voltage protection circuit and solar powered lantern used in experiments are given in Figs 4(a), 4(b) and 4(c), recpectively.

Figure 4(a): Control circuit Figure 4(b): Inverter and low voltage protection circuit Performance Characteristic of a Solar Powered Lantern

In this investigation, three different solar panels of rated output powers 43W, l0W and 6.5W have been considered. The particulars of these PV panels are



presented in Appendix (Table A1-A3). The daily output of the considered PV panels is measured. The output power, the open circuit voltage Voc and the short circuit current Isc of a typical sunny day for each of the panels are shown in Figs. 5(a), 5(b) and 5(c), respectively. In the region under study, the sky remains cloudy for a significant period of a year. To compare the output of a PV panel for a cloudy day the output parameters of a cloudy day are also shown in Fig. 5(a).

It is observed from Figs .5(a), 5(b) and 5(c) that Voc and Isc increase as the sun goes up (from 6:30 am) and Isc starts to decrease from 12:30 pm. and Voc from 3:00 pm with the declining sun. The variation of Voc from 9:30 a.m. to 4:30 pm is insignificant. The maximum output power and Isc for a cloudy day have been found to be slightly less compared to those of a sunny day, as expected.

Figure 4(c) Solar powered lantern Charging Characteristics of a Battery

While using 43W panel or l0W panel, two rechargeable batteries of 6.7 volts were connected in series to the output bus of the solar voltage regulator to study the charging characteristic of the battery. On the other hand, for 6.5W panel one single battery has been used. The increase in the battery voltages along with charging current and power with time for different panels is presented in Tables 1(a), 1(b) and 1(c). Table 1(a) presents the gain of the battery voltages for both cloudy and a sunny day. The power consumed by the two batteries, the corresponding short-circuit current Isc, battery current and the pane output power with time are shown in Figs. 6(a) 6(b) and 6(c).



Figure 5: Output characteristics of (a) 43W panel, (b) 10W panel, and (c) 6.5W panel



Figure 6: Charging characteristics of battery connected to (a) 43W panel, (b) 10W panel, and (c) 6.5W panel



Table 1(a): Development of charges in a battery connected to a 43W panel

In a Cloudy day In a Sunny day Battery voltage

(volt) Battery voltage

(volt)

Time

No. 1 No. 2

Total battery current (amp) No. 1 No. 2

Total battery current (amp)

06:30 am 4.57 4.50 0.04 3.98 3.02 0.06 07:00 am 5.70 4.25 0.06 5.67 3.59 0.07 07:30 am 5.85 4.86 0.12 5.83 3.93 0.08 08:00 am 5.79 5.69 0.13 5.84 4.69 0.14 08:30 am 5.85 5.75 0.16 5.89 5.14 0.17 09:00 am 5.90 5.89 0.20 5.93 5.65 0.20 09:30 am 5.93 5.94 0.24 6.19 5.85 0.26 10:00 am 5.97 5.98 0.24 6.20 5.86 0.28 10:30 am 6.01 6.00 0.24 6.18 5.92 0.28 11:00 am 6.04 6.03 0.24 6.19 5.98 0.29 11:30 am 6.04 6.03 0.22 6.21 6.05 0.30 12:00 am 6.05 6.03 0.20 6.23 6.09 0.30 12:30 pm 6.06 6.04 0.16 6.23 6.12 0.30 01:00 pm 6.08 6.07 0.22 6.24 6.14 0.29 01:30 pm 6.09 6.07 0.20 6.25 6.15 0.29 02:00 pm 6.09 6.08 0.14 6.26 6.16 0.24 02:30 pm 6.11 6.10 0.17 6.27 6.18 0.19 03:00 pm 6.11 6.11 0.15 6.27 6.21 0.18 03:30 pm 6.13 6.13 0.13 6.27 6.23 0.17 04:00 pm 6.15 6.15 0.15 6.27 6.25 0.11 04:30 pm 6.15 6.15 0.01 6.34 6.25 0.03

Table 1(a) and Fig. 6(a) show that 43W solar panel requires about nine and a half-hour to charge a battery in a cloudy day. Table 1(c) and Fig. 6(c) show that 6.5W panel can successfully charge a single battery in a day, while l0W panel cannot charge two batteries in a day, which is evident from Table 1(b) and Fig. 6(b). Figure 6(a) shows that a battery attains the similar voltage in a sunny day in three to seven hours depending on the initial charge of the battery. It is observed from Fig.6 that the panel output power is much higher than the power consumed by the battery. This conclusion is further intensified by the following analysis based on energy consideration.



Table 1(b): Development of charges in a battery connected to a 10W panel

Time VB1(V) VB2(V) IB(A) PB1(W) PB2(W) 06:00 am 5.1 5.14 0.000 0.0000 0.0000 06:30 am 5.1 5.14 0.012 0.0612 0.0617 07:00 am 5.2 5.25 0.063 0.3276 0.3308 07:30 am 5.29 5.36 0.099 0.5237 0.5306 08:00 am 5.32 5.41 0.183 0.9740 0.9900 08:30 am 5.39 5.48 0.275 1.4823 1.5070 09:00 am 5.43 5.60 0.310 1.6833 1.7360 09:30 am 5.51 5.69 0.380 2.0938 2.1622 10:00 am 5.63 5.76 0.400 2.2520 2.3040 10:30 am 5.71 5.80 0.420 2.3982 2.4360 11:00 am 5.78 5.89 0.480 2.7744 2.8272 11:30 am 5.84 5.90 0.500 2.9200 2.9500 12:00 am 5.91 5.99 0.490 2.8959 2.9351 12:30 pm 5.98 6.05 0.470 2.8106 2.8435 01:00 pm 6.06 6.12 0.440 2.6660 2.6928 01:30 pm 6.09 6.16 0.450 2.7405 2.7720 02:00 pm 6.13 6.21 0.410 2.5133 2.5461 02:30 pm 6.19 6.28 0.390 2.4141 2.4492 03:00 pm 6.22 6.31 0.360 2.2392 2.2716 03:30 pm 6.22 6.32 0.255 1.5861 1.6116 04:00 pm 6.23 6.33 0.125 0.7788 0.7913 04:30 pm 6.16 6.20 0 0 0

Table 1(c): Development of charges in a battery connected to a 6.5W panel

Time VB(V) IB(A) PB(W) Time VB(V) IB(A) PB(W) 06:00 am 4.80 0.000 0.0000 11:30 am 6.16 0.610 3.7576 06:30 am 4.80 0.004 0.0192 12:00 am 6.29 0.630 3.9627 07:00 am 4.90 0.050 0.2450 12:30 pm 6.30 0.600 3.7800 07:30 am 5.20 0.0910 0.4732 01:00 pm 6.36 0.580 3.6900 08:00 am 5.50 0.169 0.9295 01:30 pm 6.41 0.580 3.7178 08:30 am 5.62 0.249 1.3993 02:00 pm 6.44 0.480 3.0912 09:00 am 5.79 0.350 2.0265 02:30 pm 6.48 0.390 2.5272 09:30 am 5.91 0.430 2.5413 03:00 pm 6.51 0.370 2.4087 10:00 am 5.98 0.510 3.0498 03:30 pm 6.51 0.2550 1.6600 10:30 am 6.03 0.550 3.3165 04:00 pm 6.52 0.140 0.9128 11:00 am 6.08 0.600 3.6480 04:30 pm -- -- --



Comparison of Panel Output Power and Charging Performance

For crucial comparison of the panels, the output power and charging performance of the panels have been considered. The schematic view of the comparison is shown in Fig.7. The area under each curve i.e., the total energy delivered by solar panels or consumed by batteries, as found by trapezoidal rule, is tabulated in Table 2.

Figure 7: Comparison of panel output power and charging characteristics

Table 2: Power delivered by panels and power consumed by batteries

Energy delivered by panels (Watt-hr) Energy consumed by battery (Watt-hr) 43 W Panel 43 W Panel

Cloudy Sunny 10W Panel

6.5 W Panel Cloudy Sunny

10W Panel

6.5 W Panel

1412 1531 149 80 41 52 77 47

From Table 2 it can be estimated that a 43W panel can charge a thirty sets of two batteries simultaneously. However, from realistic point of view a conservative calculation may he adopted which allows twenty sets of batteries to be charged by a 43W panel. That is, in a day a 43W solar panel may be used to charge 40 batteries on an average and these 40 batteries can energize 40 solar powered lanterns, one battery for a lantern. It has been found that one l0W panel cannot charge one set of two batteries simultaneously in a day, but one 6.5W panel can charge a single battery in a day.



Illumination Produced by a Solar Powered Lantern One single rechargeable battery is the design requirement of the proposed lantern for both 9W lamp and 5W lamp. The battery is placed inside the lantern. The illumination produced by the lantern is measured at different distances from the lantern. The performance of both the lamps has been investigated in order to find out the efficient output voltage of the inverter circuit for each lamp. The variation of illumination level with different battery voltage at different distance from the lantern is presented in Tables 3 and 4. The graphical presentations are shown in Figs. 8 and 9. To determine the duration of acceptable light intensity the variation of lux with time and at a distance of 3 feet is tabulated in Table 5. Table 3: Illumination produced by a 9W solar powered lantern at different distances

Voltage → Distance (ft) ↓

7.0 v

6.5 v

6.0 v

5.5 v

5.0 v

4.5 v

4.0 v

3.5 v

3.0 v

2.5 v

2.0 v

1.5 v

1 230 320 280 260 260 240 220 200 180 150 110 80 2 60 100 90 90 90 80 80 50 40 35 26 19 3 45 58 55 50 48 43 40 30 22 18 13 10 4 32 40 41 38 36 30 28 25 18 15 11 8 5 24 28 26 25 24 22 20 18 17 14 8 7 6 19 20 20 18 19 18 17 16 15 12 7 6 7 13 13 12 13 13 12 11 10 10 8 6 5 8 11 12 12 12 12 11 9 8 8 7 5 5 9 8 9 8 8 9 8 8 6 7 6 5 5

10 7 7 7 6 7 6 7 5 5 5 5 5 11 6 6 5 5 5 6 6 5 5 5 5 5

It is observed from Tables 3, 4 and 5 and Figs.8 and 9 that near the lantern it

is possible to conduct all activities of the house for up to four and half-hours. The standard illumination required for different places of a residential house is presented in Appendix (Table A4). All activities including reading of a hand written material is possible up to 3 hours after switching the lantern at a distance of 5ft from the lantern. It has been observed that clear visibility exists for up to four and half hours in all places of a room of 5.4 x 5.4 meters if the lantern is placed at the center of the room.

Estimation of Number of Lanterns

The study considers that a typical rural home usually consists of two bedrooms, one kitchen, one courtyard and a bathroom, located a little away from the house.



The activities of the rural people continue up to three to four hours after the sunset. The study period of the children is usually two to three hours in the evening. Therefore, two lanterns may be required for two bedrooms, one for the courtyard and one for the kitchen/bathroom. That is, a maximum of four lanterns may be required simultaneously in a house. However, a conservative plan may reduce the requirement to one lantern during flood.

Table 4: Illumination produced by a 5W solar powered lantern at different distances

Voltage → Distance (ft) ↓

6.0 v

5.5 v

5.0 v

4.5 v

4.3 v

4.0 v

3.5 v

3.0 v

2.5 v

2.0 v

1.5 v

1.0 v

1 80 80 88 170 150 150 148 130 111 82 50 2 28 25 28 55 54 53 54 48 37 30 20 3 17 18 18 28 24 25 23 19 19 17 12 4 11 12 13 18 16 16 17 15 14 12 9 5 11 11 11 14 11 12 12 10 11 11 8 6 10 10 9 11 10 9 10 8 9 10 6 7 8 7 6 8 9 8 8 7 8 9 6 8 6 5 4 7 7 7 7 6 6 8 5 9 4 4 4 6 6 6 6 5 6 6 4 10 4 4 4 6 5 5 6 5 5 6 4 11

Light turns on but not

stable

4 4 4 4 4 4 5 4 4 4 4

Table 5: Variation of lux with time and distance

Solar powered lantern (9W) Solar powered lantern (5W) Time Battery

voltage (v) Lux Battery

voltage (v) Lux

0.00 7.00 45 7.00 18 0.30 6.80 49 6.85 18 1.00 6.45 57 6.60 28 1.30 6.20 56 6.40 24 2.00 5.85 52 6.01 25 2.30 5.55 50 5.75 23 3.00 5.09 48 5.43 19 3.30 4.43 42 5.07 19 4.00 4.21 41 4.78 17 4.30 4.16 40 4.53 12



Figure 8: Variation of illumination level for different battery voltage at different distances (9W lamp)

Figure 9: Variation of illumination level for different battery voltage at different distances (5W lamp)

Estimation of Number of Lanterns

The study considers that a typical rural home usually consists of two bedrooms, one kitchen, one courtyard and a bathroom, located a little away from the house. The activities of the rural people continue up to three to four hours after the sunset. The study period of the children is usually two to three hours in the



evening. Therefore, two lanterns may be required for two bedrooms, one for the courtyard and one for the kitchen/bathroom. That is, a maximum of four lanterns may be required simultaneously in a house. However, a conservative plan may reduce the requirement to one lantern during flood. Cost of a Solar Powered Lantern The cost and life of each unit of a solar powered lantern along with the solar voltage regulator are presented in Table 6. The cost of a solar powered lantern is evaluated by considering a 10% interest. In this evaluation, it is also considered that a 43W solar panel is capable of charging forty batteries in a day, l0W panel charges two batteries and 6.5W panel charges one battery in a day. Considering the appropriate present worth factor, the annual repayment cost of each unit of a solar powered lantern is evaluated and is presented in Table 7. Table 6: Price and life of different units of a solar powered lantern scheme

Description of Unit Total cost in Taka Life in years

PV Panel 1,700.00 (6.5W Panel) 7,800.00 (10W Panel)

18,700.00 (43W Panel)

20

Voltage Control Unit 135.00 20 Low Voltage Protection Circuit 51.00 20 Inverter 58.00 20 Casing 100.00 20 Rechargeable Battery 310.00 2 Compact Fluorescent Lamp 120.00 10

Table 7: Annual repayment cost of each unit of a solar powered lantern

Annual Repayment Cost in Tk. Unit 43W Panel 10W Panel 6.5W Panel

Solar Panel 54.91 458.00 199.68 Battery 178.62 178.62 178.62 Lamp 19.53 19.53 19.53 Voltage Control Circuit and Low Voltage Protection Circuit

21.58 21.58 21.58

Inverter and Casing 18.56 18.56 18.56 Total 293.20 696.29 437.97



CONVENTIONAL SOURCES OF LIGHTING IN A RURAL HOME The conventional sources of lighting in a rural home are usually two types: (i) Hurricane lantern and (ii) Candle. The photographic view of Hurricane lantern and Candle used in our experiment are shown in Figs. 10 and 11, respectively. The hurricane lantern is made of steel. It has a reservoir/tank for fuel. The usual fuel is kerosene. The flame is produced by firing a cotton feather, which absorbs kerosene from the fuel tank. A tubular glass covers the flame.

A candle is made of wax. It comes in different sizes. For this study a candle of 24.5 cm height and 4.8 cm diameter is considered. It provides light for 20 hours for its complete burn. The illumination produced by a hurricane lantern with low and high flame and a candle has been compared with that of the solar powered lantern in Table 8. This table gives the illumination level at different distances from the source.

Figure 10: Hurricane-lantern Figure 11: Candle In evaluating the cost of a hurricane lantern, it is considered that its life is 5

years and price is Tk. 100. The consumption of kerosene by a hurricane per hour is 41 ml and the price of kerosene per litter is Tk 18.00. It is also considered that hurricane lantern is used for 4 hours per day for illumination. Therefore, the annual repayment cost of a hurricane lantern including the fuel is Tk. 1121.38. The cost of a candle is also evaluated. Considering 4 hours of illumination in each day the annual expenditure becomes Tk.2190.00.



Table 8: Illumination produced by a hurricane lantern and a candle at different distances

Light Intensity (Lux) Solar Powered

Lantern

Distance

(ft.) Hurricane Lantern

(Low flame)

Hurricane Lantern

(High flame)

Candle

9W 5W 1 55 60 65 320 170 2 17 20 19 100 55 3 9 11 10 58 28 4 8 9 8 40 18 5 5 7 6 28 14 6 3 5 4 20 11 7 2 3 2 13 8 8 2 2 2 12 7 9 1 2 1 9 6

10 1 1 1 7 6 11 1 1 1 6 4

COMPARISON OF SOLAR POWERED LANTERN WITH THE CONVENTIONAL SOURCES The variation of illumination with distance of a solar powered lantern is compared with those of conventional sources in Fig. 12. It is clearly observed that the solar powered lantern produces higher illumination level at all distances.

From Table 8 the average illumination level of each source may be evaluated. The average illumination of solar powered lantern (9W and 5W), hurricane lantern and a candle are 55.73, 29.73, 11 and 10.82 lux, respectively. Note that the average illumination of a hurricane lantern with the low flame is 9.45 lux. Therefore, a 9W solar powered lantern is equivalent to two 5W solar-powered lanterns, 5 hurricane lanterns and 5 candles, as far as brightness is concerned. Considering this illumination equivalence the annual cost of the illumination of an isolated rural home by a solar powered lantern is compared with those by the conventional sources in Table 9. It is clearly observed from Table 9 that a solar powered lantern is much cheaper than the conventional sources. Moreover, it produces a higher illumination than a conventional source. Also it is hazard free from the operational point of view.



Figure 12: Comparison of the illumination level of solar powered lantern, hurricane lantern and candle

Table 9: Comparison of illumination cost of a rural home with different types of sources

Sources Annual Expenditure (Tk.) With 43W Panel 879.60 With 10W Panel 2088.87

9W Solar Powdered Lantern

With 6.5W Panel 1313.91 With 43W Panel 1759.20 With 10W Panel 4177.74

5W Solar Powdered Lantern

With 6.5W Panel 2627.82 Hurricane lantern 16820.71 Candle 32850.00

CONCLUSIONS This paper proposes the use of solar powered lantern for the lighting system of an isolated flood affected home. It presents the design, construction and the performance characteristics of a solar powered lantern. The lantern is much cheaper than the conventional sources of illumination. Moreover, it produces higher illumination without any operational hazard.



REFERENCES Bishop J.W. (1989), Microplasma Breakdown and Hot-Spots in Silicon Solar

Cells, Solar Cells, pp.335 – 349. Molenbrock, E., Waddington D.W. and Emery K.A. (1991), Hot Spot

Susceptibility and Testing of PV modules, Proc. of 22nd IEEE PV Specialists conference, pp.547 - 552, Las Vegas, Nevada.

Pellegrini, B. (l99l), Reverse Current- Voltage Characteristic of Almost Ideal Silicon p-n Junctions, J. Appl. Physics, pp.1071 - 1080.

Chakina, B., Saha, U.K., Khisa, J.K. and Ahsan, Q. (1997), Economic Benefits: Use of PV Cell for an Office lighting, Proceedings of ISAAE, pp.542-549, Johor Babru, Malaysia.

Alam, M., Karim, R. and Rahman, H. (1998), Solar Powered Lantern, B.Sc Engineering Thesis, BUET, Dhaka.

Kaufrnan, J.E. and Haynes, F.H. (1981), JES Lighting Handbook, Reference Volume, Illuminating Engineering Society of North America.

APPENDIX Table A1: Description of 43W PV panel Manufacturing company: Arco Solar Inc. Model: M65 Solar Irradiants and Cell Temperature as indicated Made in USA Rated Power at 20oC = 43W Maximum amp at 47oC(sc) = 3.68 A Maximum volts at 0oC = 20 V dc Size: 48 x 42 sq. in.

Table A2: Description of 10W PV panel Manufacturing company: Webel Solar Model: SQR49L Solar Irradiants and Cell Temperature as indicated Made in India Rated Power at 20oC = 10W Maximum amp at 47oC(sc) = 1.23 A Maximum volts at 0oC = 20 V dc Size: 15 x 14.75 sq. in.



Table A3: Description of 6.5W PV panel Manufacturing company: Siemens Model: GL418TF Solar Irradiants and Cell Temperature as indicated Made in Japan Rated Power at 20oC = 6.5W Maximum amp at 47oC(sc) = 1.5 A Maximum volts at 0oC = 10 V dc Size: 12 x 9 sq. in.

Table A4: Standard Illumination (Kaufman and Haynes, 1981)

Sites Standard Illumination (Lumen/m2) Living room/dining room/hall 2.0 Kitchen/laundry 3.0 Bathroom/toilet 3.0 Corridors 1.0 – 1.5 Working sites 5.0 Hand writing places 7.5


Effect of the 1998 Flood on Non-Engineered Structures

Salek M. Seraj and Md. Rezaul Karim

Department of Civil Engineering Bangladesh University of Engineering and Technology, Dhaka-1000, Bangladesh Abstract

In 1998, Bangladesh experienced the worst ever floods in terms of severity, destructiveness and duration. The duration of the flood exceeded all previous records. Besides loss of human lives and livestock and damages to roads and bridges, there have been extensive damages to buildings in the flood-affected areas. The conditions of the majority of houses in the flood-affected areas are poor and therefore susceptible to damages. In this paper, an attempt has been made to have a quantitative estimation of the damages to houses caused by the 1998 flood. Study has also been made on the conventional techniques used in building the vitte and bera of houses by villagers and some conclusions have been drawn about the vulnerability of houses to flood damage. It appeared that the villagers do not use any special technique for the construction of flood-resistant houses. The tentative and preliminary estimates made in this study suggests that over Taka two thousand five hundred crore would be required for the reconstruction and rehabilitation of the flood-damaged non-engineered rural houses of Bnaglsdesh.

INTRODUCTION In 1998, Bangladesh experienced the worst ever floods of the country. This flood exceeded all other floods in the living memory of the people of the country in terms of severity, destructiveness and duration. During the floods, apart from the heavy loss of human lives, livestock and standing crops, large-scale damage




of the roads, infrastructure and rural houses took place. The havoc of this flood that ran into several months can hardly be ever estimated. The housing condition of the majority of households of Bangladesh is simply poor and, thus, the losses to houses were of considerable magnitude. In this paper, an attempt has been made to have a quantitative estimation, albeit tentatively, of the damages caused by the floods of 1998 to rural houses. Here, efforts have been made to conduct a comprehensive study on the conventional techniques used by the villagers in preparing the Vitte (plinth) and Bera (wall) of the rural houses and houses that were seriously affected due to particular conventional technique used by the builders have been detected. Relevant information and data on flood 1998 have been collected from various available sources, as well as through a questionnaire survey carried out among 139 rural people, who were directly affected by the flood of 1998. A very approximate estimate has suggested that about 0.05 million families were most severely affected by the flood of 1998. The flood either washed away or severely damaged their houses and household belongings. From the study it has been found that to reconstruct all the affected houses (with an average size of 18 x 11 x 7 feet), about Tk. Twenty five hundred Crore (Tk. 2528,00,62,500) would have been required. Use of appropriate flood-resistant house building techniques by the villagers could have helped them to cope with this flood with lesser economic losses.

Usually when an area goes under water and remains under that condition for sometime it is called inundation. On the other hand, when this inundation causes damage to property and life, disrupts communication and causes harm to people as well as fauna and flora, it is called flood. Flood, apart from causing damages to economy and dislocation to public health, renders thousands of people homeless, causes damage to crops in the field, disrupts normal agricultural cycle, causes death to livestock, poses serious threats of nutrition, and may results into hunger and famines. The poor and downtrodden groups of any society usually suffer the most during and at the aftermath of any flood. The amount of damage from an event of flood usually depends on various factors. The location (urban/rural) of the area subjected to flooding as well as the depth, duration, and extent of flooding usually determines the potential amount of damage an incidence of flood may cause. Bangladesh experiences moderate to heavy flooding almost on a regular basis. Official statistics of last few decades indicate that each year on an average 20% of the land area of Bangladesh goes under floodwater. In some of the previous devastating floods, the effects on human life as well as on land were quite unbearable.

In reality, floods are natural phenomena in Bangladesh as it is a flood plain country. The floods in Bangladesh are mainly caused due to the fact that: (a) Bangladesh is a deltaic land situated just about fifteen feet above the sea level; (b) Bangladesh has a small geographical area; (c) Bangladesh accepts the huge

Effect of Flood on Non-engineered Structures

volume of the Himalayan ice-melt water flowing through the rivers up-stream in the neighboring countries; (d) The rivers of Bangladesh are not deep enough, (e) Tidal bore and, of course (6) Excessive rainfall.

The flood of 1998 made the people of Bangladesh to be faced with a disaster with catastrophic dimensions. This worst flood in the history of Bangladesh lasted from July to September 1998. During this flood, about 100,000 sq. km out of 148,393 sq. km, i.e., 69% of the total landmass of the country was literally inundated for about two and a half months. Floodwater engulfed about 52 out of 64 districts of the country directly affecting nearly 30.1 million people of the country. During the floods, more than one million people took refuge in about 3000 makeshift flood shelters (schools, colleges, community centers etc.) throughout the country. The timing, duration and magnitude of flood 1998 in Bangladesh have proven it to be unlike any other in the country’s history. This flood was unique in the sense that the floodwaters receded at a much slower pace than the other floods that Bangladesh experienced prior to 1998.

During this flood, beside others, houses in the affected areas suffered heavy losses. Some houses were completely washed away by riverbank erosion and houses that were not washed away suffered extensive damages. However, the total losses in the housing sector have not been estimated so far. The havoc of this long flood can hardly be ever estimated. But its longer-term impact required a great deal of careful study.

Whereas complete prevention of flood is not possible, suitable measures can be adopted to alleviate the distress of the people. At the same time novel strategies concerning future planning and development may be formulated so that people learns to live with such devastating natural disasters with dignity and pride. The lessons learnt from the flood of 1998 may very well guide the decision-makers to plan a better future for Bangladeshi homeowners in the flood prone areas of the country. It is again very much important that suitable affordable techniques of building flood-resistant non-engineered rural (poor) housing are developed so that in the event of future floods losses to houses may be minimized as far as practically feasible.

Soon after the flood of 1998, various agencies carried out surveys on the people directly affected by it. A number of pockets, which were badly affected by the flood, in various parts of the country were identified. These vulnerable pockets have been given in Table 1 as well as shown in Fig. 1.

METHODOLOGY

Some of the households affected by the flood of 1998 were carefully selected so that they act as representative samples. Only those rural households that were directly affected by flood were chosen.




Whereas the flood of 1998 affected a total of 52 districts of Bangladesh, three districts out of these 52 were selected for conducting necessary surveys. The selected districts were, of course, identified as the most severely flood affected areas and were also easily accessible for the field survey. For each of the districts selected for flood survey, one Thana was selected randomly. From each Thana, one or two unions (depending on the degree of damage) was/were selected and from each union, one to three-village(s) were selected for the study. The study area is given in Table 2. During the field survey conducted across several villages, a questionnaire was prepared in Bengali and the survey was conducted among the flood affected peoples. The outputs of the survey were then processed in the computer. Again, relevant information available from various sources was gathered and efforts were made to tentatively determine the total number of affected households throughout the country due to this flood.

As the degree of damage sustained by the rural houses varied from house to house, as well as from village to village, an indicator was assumes to quantify the degree of damage. For the houses whose Vitte was completely damaged and Bera was also washed away, the indicator has been set at 100% damage of the house. Similarly for the houses whose Vitte was not washed away but fully damaged and Bera was also fully damaged, the indicator has been taken as 50-100% damage of the house. If either the Vitte or the Bera showed no significant damage, then the indicator has been reported as 0-50%. For other combinations, the extent of damage has been taken as 50% or has been left at the discretion of the surveyor.

Table 1: List of areas affected by the 1998 flood

District Thana Union Villages

Pauch Gachhi Char Shitaijhar Kadamtala Madhay Kusumpur Noabosh Ghughudanga Char Vagabotipur

Noanipara Kurigram Char Zattrapur Nayerhat Kawerchar Fulerchar Kurichar Kurigram Ashtimirchar Tapurchar

Hakanda Jadurchar Ulipur Jatia Jatia



All Kurigram (contd.) Erandabari Erandabari

Rajibpur Fazlupur Fazlupur Kanchipara Kanchipara Gozaria Gozaria Uria Uria Shaghata (Char Area) (Char Area) Gaibandha All Gaibandha Shadullahpur All Idulpur Idulpur Palashbari Barisal Rampur

Dewanganj Char Gelabari Motherganj Char Nandanerpara Bakshiganj Bir Nandanerpara Islampur Chenadoli Kutubullah Char Jamalpur Shingvana Dewanganj Chenadoli Char Gelabari Bahadurabad Sardarpara Jamalpur Ghoradhap Ghoradhap Dewqanganj Merur Char Kolkihara Sarishbari Satpowa All Jhagurara All Tupkarchar All Sarishbari Mahmudpur All Chineytola All Sudhibari All Haripur All Bhaluka Berunia Chandar Hat Nandail Rajgati Rajgati

Mymensingh Kishoreganj Latibabad Durail Charpara Latifabad Itna Badla Taleshwar Raituli Pangdalang Sherpur Sherpur Sherpur Char Sherpur Netrakona Barhatta Singba Arshira Barhatt Baushi Sahipura





Pabna Shujanagar Shujanagar Tarabari

Bogra Kahalu Bir Kedar Bholta

Barisal Bobnadi

Gaurnadi Gaurnadi Kolejdi

Pirojpur Najirpur Sriramkathi Buichakathi

Gajipur Kapashia Shammania Shammania

Narsingdi Raipur Moheshpur Joynagar

Faridpur Faridpur Faridpur Pourashava Charmadhabia

Shovarampur Modhukhali Baghat Gamara Shariatpur Gosairhat Marachfali Gosairhat Lalbagan Lakshmipur Ramghar Taraganj Char Mathia Natore Natore Sadar Khajura Khajura

Table 2: Areas covered under this study

Zone District Thana Union Villages Total No. of houses

Pauch Gachhi Char Shitaijhar 200 Kadamtala 100 Noabosh 40 Zattrapur Char Vagabotipur 30 Noanipara 50

Study area-1

Kurigram

Kurigram

Sadar

Char Zattrapur 150 Jamalpur Islampur Chenadoli Shingvana

300Faridpur Pourashava

Charmadhabia 50

Study area-2

Faridpur Faridpur

Shovarampur 100


Figure 1: Extent of the 1998 flood

DESCRIPTION OF THE HOUSES

Various types of houses that exists in the study areas were critically inspected to quantify the state of the houses. In all study areas, most of the houses were Kancha (Figure 2). Some semi-Pacca (Figure 3) and Pacca (Figure 4) houses were also found. Kancha is the type of house in which the building materials are mainly bamboo, G. I. sheet, wood, G. I. Wire, rope etc. The Vitte of the Kancha houses is made of compacted earth. The pillars are made of bamboo. Again, semi-Pacca is the type of house in which the building materials are mainly bamboo, G. I. sheet, wood, G. I. Wire, rope etc. The Vitte of this type of house is made of compacted earth or sand, with a 10 inch brick wall boundary around it.




This brick boundary of the plinth starts from a certain depth below the ground level and extends up to the floor level. Pacca is the type of house in which the building materials are steel rod, cement, sand, brick chips, bricks, G. I. sheet etc. The structure is built as RCC structure except the roof is made of G. I. Sheet. Again, Chhawn house is a type of Kancha house with the only difference that the roof is made of grass sticks (matured, dried rice, wheat trees etc.) instead of G. I. Sheet. The overall condition of the Chhawn house is, in general, not so good. It is made by the lowest income people just to take shelter. The full portion of this house is made of Catkin sticks including with bamboo sticks (sometimes, the roof may be made of G. I. Sheet). No established Vitte has been found for this type of house.

Figure 2: A kancha house

According to the present study, it has been found that about 68% of the houses

of the study area Kancha, about 10% of the houses are semi-Pacca, only 4% of the houses are Pacca and 18% includes other types of houses such as Chhawn’s house, Kancha shed etc.

Questionnaire survey was conducted on a total of 139 people randomly selected from all the three study areas. About 57% of the people of the study area were farmers, 34% of the people were day labourers, 2% of the people were businessmen, 3% of the people were service holders, and 4% were engaged in other types of occupation such as boatman, fisherman, etc. Interview was also taken on females; all of them were day labours. The schematic representation of the occupation of the people is shown in Fig. 5. All the people interviewed were


permanent residents of the study area since birth and were present in the area during the 1998 flood. In 1988, average depth of inundation of floodwater was 4 feet for all the study areas. In contrast during 1998, average depth of inundation of floodwater was three feet from the ground level of the houses with an average period of inundation of about 76 days.

Figure 3: A semi-pacca house

Figure 4: A pacca house under flood water




EXTENT OF DAMAGE TO THE HOUSES

Extent of damage of the houses basically depends on the condition of the house, i.e. whether the house is Kancha, Semi Pacca or Pacca. The 1998 flood had caused serious damages to the houses of the affected area, especially the Kancha and Semi Pacca houses of the locality. However, all types of the houses were not affected by the same degree. Again in some houses, the Vitte was completely washed away but the Bera was not completely damaged pointing to the fact that different segments of the houses were affected by different degrees. Thus, an indicator for the extent of damage was developed to quantify the amount of damage due to flood. For the house whose Vitte was completely damaged and Bera was also washed away, the indicator has been taken as 100% damage of the house. Similarly for the house whose Vitte was not washed away but fully damaged and Bera was fully damaged, the indicator has been taken as 50-100% damage of the house. If either the Vitte or the Bera underwent no significant damage, then the indicator has been reported as 0-50%. For other combinations, the indicator for the extent of damage is taken as 50% and in some cases engineering judgement has been applied.

From the field data, it is found that about 16% of the houses of the study area was completely (100%) damaged, about 25% houses were 50-100% damaged, about 42% houses sustained 50% damage and only 17% houses were damaged by 0-50%. According to estimates published in the daily newspapers, the total houses damaged (assumed as 100% damaged) due to 1998 flood was 5,50,000 numbers, and using this information the total number of houses damaged under various categories have been estimated and as shown in Table-3. The schematic representations of the extent of damage of the houses of the study area and the total number of houses affected are shown in Figures 6 and 7, respectively.

Table 3: Number of houses affected by flood

Description of damage

of houses (%) Percentage of total

damage (%) Total number of houses

affected (Nos.)

100 16 5,50,000 50-100 25 8,59,375 50 42 14,43,750 0-50 17 5,84,375

Total 34,37,500


57%

2%

34%

3%

0%

4%FarmerBusiness manDay labourerService holderHousewifeOthers

Figure 5: Description of the occupation of the people of the study area

10016%

50-10025%

5042%

0-5017%

10050-100500-50

Figure 6: Extent of damages to houses of the study area

REPAIR COST OF HOUSES DAMAGED

Repair cost depends on the size of the house and the condition of the house. Field observation revealed that the average amount required to make a house of average quality (in between Kancha & Semi Pacca) and of average size (18 x 11 x 7) is Tk. 12,257. About 31% of the total cost is spent in preparing Bera and 69% of the total cost is spent in making Vitte of houses. Repair cost of such houses under various levels of damages was estimated and is given in Figure 8. The total repair and reconstruction cost of all the partialy and completely damaged houses has been tabulated in Table 4. From the table, it appears that the 1998 flood resulted in a total rural house reconstruction cost of about 25280 million Taka.




3437500

550000859375

1443750

584375

0

500000

1000000

1500000

2000000

2500000

3000000

3500000

4000000

No.

of h

ouse

s

Totalaffected

100 50-100 50 0-50

Extent of damages of houses

Figure 7: Number of houses affected by the flood

Table 4: Repair and reconstruction cost of damaged houses

Description of damage of houses (%)

Total amount required to

make a house (Taka)

Total number of houses affected

(Nos.)

Repair cost (Taka)

100 5,50,000 674,13,50,000 50-100 (75) 8,59,375 790,00,19,531

50 14,43,750 884,80,21,875 0-50 (25)

12,257

5,84,375 179,06,71,094

Total 34,37,500 2528,00,62,500


12257

9192.75

6128.50

3064.25

0

2000

4000

6000

8000

10000

12000

14000

Rep

air c

ost i

n Ta

ka

100 50-100 50 0-50

Extent of damage of houses

Figure 8: Repair cost of a house of average size 18 x 11 x 7 feet

RECONSTRUCTION TECHNIQUES OF THE HOUSES DAMAGED

Villagers usually use no special technique to build their houses. In making the Vitte of the house, most of them use only soil by filling it up to the desired height. The desired height of Vitte depends on the ground level of the area. On an average, height of Vitte is 3.36 feet. Some people plant trees around the Vitte as a preventive measure against natural forces. In making the Vitte, villagers use water to mix with soil, so that the filled soil can be compacted well. No other materials such as rice husk, rice husk ash etc. are generally used by them. They basically do not take any special care for protecting Vitte from flood damage. During reconstruction of Vitte, house builders first put a perimeter mud wall around the proposed Vitte, then put soil inside and compact it (Figure 9). Lateral support using bamboo and trees are sometimes used (Figure 10). Usually people do not take any measures for protecting Bera from damage. Only a small percentage of people use plastic, tar for this purpose (Figure 11). For strengthening the pole of the house, pole of cement is used by some solvent people. It has been gathered that about 19% people uses bamboo for giving lateral support to Vitte, 39% plants trees around Vitte and the rest 42% take no




special measures (Figure 12). Again, whereas both male and female workers are usually engaged in the reconstruction works, women workers are paid fewer wage than their male counterparts (Figure 13). Average time needed for getting into the house after making Vitte is about 14 days. It has been revealed that people would like to use high technology, as long as it is affordable.

Figure 9: Rebuilding of Vitte washed away by flood

Figure 10: Mud-plinth (Vitte) protection using bamboo


Figure 11: Bera protection using polythene sheet

39%

19%

42%

By planting trees around the Vitte Using bamboo/trees for lateral supportNone Figure 12: Special techniques used for protecting Vitte from damage




62%

38%

Male Female

Figure 13: Average labour charge

In this regard, some recent works of Roy, et al. (2000) and Saha (2002) can be

effectively used to minimize losses due to floodwater, which usually washes away the mud-plinth (Vitte) or mud-wall resulting in the total collapse of the poorly built rural houses. From these works it has been observed, quite conclusively, that by mixing 4%-6% cement by weight, depending upon soil type, with the soil of plinth towards the later stages of Vitte construction, mud-plinth can be made floodwater resistant. Figures 14 and 15 show mud-cement block casting and testing under submerged water, respectively.

Figure 14: Mud-cement block preparation


Figure 15: Mud-cement block testing under water CONCLUSION

The present study is based on a limited survey conducted over a small fraction of the total area that went under water during the flood of 1998. The study relied on information that was available in the mass media as well. The following is the summary of the findings: • It has been found that about 57% of the people of the study area (on whom

observations were made) are farmers, 34% of the people are day laborer, 2% of the people are businessman, 3% of the people are service holder, and 4% includes other types occupation such as boatsman, fisherman etc.

• In most part of the country, the scale of severity in 1998 was much higher than that of the flood 1988 in terms of duration and overall damage to properties.

• About 68% of all the houses of the study area were Kancha, about 10% Semi Pacca, about 4% Pacca, and about 18% include other types of houses.

• Average depth of inundation of floodwater in 1988 was four feet in the study area. In contrast, in 1998, the floodwater remained three feet from the ground level, with an average period of inundation of about 76 days.

• Severe damage of the Kancha house of the people has been reported. About 16% of the houses of the study area were completely (100%) damaged, about 25% houses were 50-100% damaged, about 42% houses sustained 50% damage and only 17% houses were damaged by 0-50%. About 34,37,500 number houses over the whole country have been estimated as affected by the devastating flood of 1998.




• The total amount of money required to reconstruct all the affected houses of rural Bangladesh at the aftermath of 1998 flood is estimated to be about Taka two thousand five hundred Crores. This tentative estimate is, perhaps, on the higher side. However, this gives a clear indication of the large scale losses that took place in the non-engineered rural housing sector during 1998 flood.

• Currently, villagers do not adopt any special technological solution to make their houses floodwater resistant. Appropriate and affordable village building technologies may be adopted in the future to minimize losses. Again, further research, with the objective of being implemented immediately, may be initiated in an effort to build safer houses in rural Bangladesh. Such a study should look into alternative materials as well as construction techniques.

REFERENCES

Roy, U. K., Seraj, S. M., Roy, P. S. and Alam, M. S. (2000), "Some Aspects Towards Development of Hazard-Resistant Rural Homes in Bangladesh", In Affordable Village Building Technologies, Proceedings of the Second Housing and Hazards International Seminar held in Dhaka, Bangladesh, 6-8 February 1999, edited by Seraj, Hodgson and Choudhury, pp. 29-40.

Saha, S. (2002), “Development of Non-Engineered Rural Houses”, M. Engg (Civil) Thesis, Department of Civil Engineering, Bangladesh University of Engineering and Technology, Dhaka.

Documents

arc456.files.wordpress.com · Contents Authors and Title Page Preface v Foreward vi M. Feroze Ahmed and K. Nurul Ashfaque Sanitation and Solid Waste Management in Dhaka City During