Embed Size (px)
Citation preview
APPLICATION OF GEOINFORMATICS FOR FLOOD STUDYAT TARAPUR UNION OF GAIBANDHA
S. M. IMRUL HASAN
INSTITUTE OF WATER AND FLOOD MANAGl.~MENT
BANGLADESH UNIVERSITY OF ENGINRF.RING AND TECHNOLOGY
NOVEMBER, 2006
-•••
Application ofGeoinfonnatics for Flood Study at Tarapur Union ofGaibandha
A thesis by
S, M. Imrul Hasan
In partial fulfillment of the requirement tor the degree of
MASTER OF SCIENCE IN WATER RESOURCES DEVELOPMENT
INSTITUTE OF WATER AND fLOOD MANAGEMENT
BANGLADESH UNIVERSITY Of ENGINEERING AND "1ECHNOLOGY
NOVEMBER, 2006
•
BANGLADESH UNIVERSITY OF ENGINEERING AND n:CHNOLOGY
INSTITUTE OF WATER AND FLOOD MANAGEMENT
The thesis titled 'Application of Geoinformatics for Flood Stud)' at TarapurUnion of Gaibandha' submitted by S.MJmrul Has.an, Roll No. MF 0428003,Session: April 2004, has been accepted as satisfactory in partial fullillment of [herequirements for the degree of M. Se. in Water Resources Development inNovcmber13,2006.
BOARD OF EXAMINERS
Dr. G.M.Tarekuilslarn(Supervisor)Associate ProfessorInstitute of Water and Flood ManagementBangladesh University of Engineering and TechnologyDm."
~~h~Assistant ProfessorInstitute or Water and Flood ManagementBangladesh University of Engineering and TechnologyDhaka
....~ ....Dr. A.K.M. SaifullslamAssistant ProfessorInstitute of Water and Flood ManagementBangladesh University of Engineering and Technologymill"
rj,k.']Q.L0c.'12,J,,~DirectorInstitute of Water and Flood ManagementBangladesh University of Engineering and TechnologyDhili-"".'"""""',-....--tS.'"r ~->;)v>rI r--.................................................Dr. K.M. MuniruzzamanProlessorDepartment of Urban and Regional PlanningBangladesh University of Engineering and TechnologyDhili
Chairman
Member
Member
Member(Ex-officio)
Member(External)
It"-_,.,_'
CANIlLDATE'S I)ECLARA nON
It is hereby declared that this thesis or any part of it has not been submitted elsewhere
for the award of any degree.
S. M. Imrul Hasan
Roll No. MI' 0428003
Session: April, 2004
Table of Contents
Table of ContentsList of TablesLtd of FiguresList of AbbreviationsGlossaryAcknowledgementAbdraet
CHAPTER ONE: II'TRonUCTION
1.1 Background onhe Study1,2 Rationale of the Study1.3 Objectives of the Study
CHAPTER TWO: LITERATURE REVIEW
2.11ntroduction2.2 Brief History of GIS and RS in Bangladesh2.3 Use orGlS and R<; ill Flood Study
2.3.1 Application ofOIS and RS in flood Study in Bangladesh2.3.1.\ RS and GIS in flood Mapping and Monitoring2.3.1.2 RS and GIS in Flood Risk and Ha;,ard Mapping2.3.1.3 Application orOIS and RS in Flood Siudy in Meghna-
Dhonagoda Polder2.3.1.4 Flood Study on Sundarganj Thana
2.4 Vulnerability and Ri.,k2.4, I Vulnerability
2.4.1.1 Element-at-Risk Indicators2.4.1.2 Exposure Indicators2.4.1.3 Su>ccptlbility Indicators
2.4.2 flood Ri,k Analysis2.4.2.1 Risk Perception2.4.2.2 Risk A"e>sment
2.4,2.2.1 Specil1e Risk (Rs)2.4.2,2.2'1 otal Risk (Rt)
2.4,3 Relationship among Flood Damage, Vulnerability and RiskPerception2.4.3, I Relationship between t lewd Damage and
Vulnerability2.4.3.2 Relationship between Risk Percepti'ln and
PageN,.,Vlll
",I'"xiii'"
55
667W1012
131616171718181919191920
20
20
Vulnerability2.4.4 Flood Dam~ge Estimation
CHAPTER THREE: MICTHOOOLOGY
20
"3.1 R~eonnaiss.anee Survey 243.2 Sel~clion of Study Area 243.3 Primary Data Collection 253.4 Secondary DataColl~ction 253.5 Flood Frequency Analysi. 25
3.5.1 Goodness-of-fitAnaly<is 263.5. I Determination offlood Level 27
3.6 D~velopment of Flood lnundation and Flood Depth Maps 273.7 Multispectral Image Classification 273.8 Screen Digilizing 283.9 Development Flood Hazard Vulnerability Function for Agriculture 283,10 Development Flood Hazard Vulnerability l\1aps of Agricult\lre 293.11 Overall View of the Metbodolog)' 29
CHAPTER 1<'OUR:OVERVIEW OF THE STUDY AREA
4.1 Study Area4,2 Climatic Condition4,3 Soil Condition4.4 Topography and Hydrology4.5 Maior River System of the Study Area
4.5.1 Teesta River4.5.2 Brahmaputra River
4.6 Naturall)isa<ter4.7 Agriculture4.8 General Intormation of the Study Area
CHAPTER FIVE: DEVELOPMENT OF VULNERABILITYFUNCTION
30323232333333353637
5,1 Introduction 395.2 Vulnerability F\lnction for Agricullure of Tarapur Union 39
5.2.1 ldentification of Cropping Area and Pattern 405.2.2 Classification of Land based on Elevation 425.3.2 ldentificarion of Flooded and NOll-floodedAgricultural Land 43
5.3 Development of vulnerability Function of Agrie\llture through Depth-Damage Relation 45
CHAPTER SIX: ANALYSIS AND DISCUSSION
6.1 Introduction6.2 Deterl1lination of Waler Level of Tarapur
4848
_____1
6.3 FI<Xld Frequency Analy.,is 506.4 Digital Elevation Model 516.5 Flood Inundation Map 526.6 Comparison "fFlood Inundation Map Based on DEM and SateHite
Image 566.7 Floodu:vel Map 586.8 Flood Depth Map 616.9 ldentification of Agricultural Land and [lome,tead Area and
lnfrastructure 646.10 ldentification "f Agricultural Land Regarding Different
Return Period 676.11 Development of Vulnerability Map of Agriculture Using
Vulnerability Function 70
CHAPTER SEVEN: COJli:CLUSIONS ANDRECOMMENDA nONS
•
7.1 Conclusions7,2 Roxoffimendations
7677
References 79
Appendix A 84Appendix B 89
Appendix C 101
vii
LIST OF TABLES
Table No. Title PageNo.
51
16
5864
43
of 45
for 47
Overall view afthc flooded and non-flooded area
Ovemll view afthe variation offlood depth
Crop calendar of the study area
Flood ,lali,lies for flood level 00.00 m and 3.50m (PWO)
Homestead and cropped area of:'.undarganj thana (Area inacres) 36Area under different cmps of Tarapur union (Acre) 37Ba,ic characteristics of agricultural landusc practices of 37:'.undarganj thanaLand classification of Sundarganj 'i hana 40
42Land classification regarding flood inundation of Sundarganj 43thanaInundaled agricultural land of Tarapur union based OnOEM
flood and other disasters affeded cropping "reaSundarganj thana (hectare)Ovemll view of depth-damage relation of agricultureTarapur unionWater lcvels from flood frequenc} analysis
Table 5,4
Table 5.5
Table 5.6
Table 5.1Table 5.2
Table 5.3
Table 6.1
Table 6.2
Table 6.3
Table 4.2Table 4.3
Table 2.1
Table 4.1
viii
FigureNo.
Figure 2.1
Figure 2.2
Figure 23
Figure 3.1
Figure 4.1
Figure 4.2
Figure 4,3
Figure 4.4
Figure 5.1
Figure 5.2
Figure 5.3
Figure 6.1
Figure 6.2
Figure 6.3
Pigure 6.4
Figure 6.5
figure 6.6
Figure 6.7
Figure 6,S
Figure 6.9
Figure6.10
Figure 6, 11
Figure 6.12
-,-- -LIST OF )<'IGURES
Tille
Estimated flooded area, using OEM high water level of3.00m
SPOT image classified as flooded and non-flooded areas.u,ing density slicing.
Comparison between flooded areas on image and DEM (fluodlevel 3 m)
Overall view of the melhodology
Study area map
Study area map of Tara pur union
The river net-work syslem of the study area.
Types of flood and extent map
Agricullllmliand of lhe sllIdy area
Flooded und non-Flooded Area
Vulnerability function of agriculture for Tarapur union
River network sy,lem of the area
Maximum waler level of Tara pur through inlerpolation (July10 Sept.)Digital elevation model of Tara pur area
Plood inundation map for 2-year return period
Flood inundation map for 5-year retllrn period
flood inundation map for lO-year return period
Plood inundation map for 20-year return period
Flood inundation map tor 50-year return period
Flood inundation map based on DEM and water level data onthe date of satellite image acquisition (23Id Jul>, 2004).
Land water elassified map of based on screen digitizing ofRadar image on 23m July. 2004,
FI",}d level map for 2-year return period
Flood ievel map for 5-year relurn period
PageNo.
14
15
15293031
3435
41
"4749
50
5153
5454
55
55
56
57
5859
FigureNo.Figure 6.13
figure 6, 14
Figure 6.15
figure 6.16
Figure 6.17
Figure 6.1 R
Figure 6.19
Figure 6.20
Figure 6.21
Figure 6.22
Figurc 6.23
Figure 6.24
Figurc 6.25
Figure 6.26
Figure 6.27
Figure 6.2R
Figure 6.29
Figure 6,30
Figure 6.31
l'igure 6.32
Figure 6.33
Flood level map for 1O-year return period
Flood level map for 20-year rerum period
Flood level map for 50-year return period
flood depth map for 2- year return period
Flood depth map for 5- year return period
Flood depth map for 10- ycar return period
flood depth map for lO-year return period
Flood depth map for 50- year return period
Map conmining agricultLLralland homestead and infrastructure
Home,tead and infrastructure only
Map containing agricultural land only
Map containing agricultLLralland for 2-year return period
Map containing agricultural land for 5-year return period
Map containing agricullllralland fm 1O-year return period
Map containing agricultural land for 20-year relllrn period
Map containing agricLLlturalland for 50-year rerum period
Vulncrability map of agricultural land for 2-year return period
Vulnerability map of agricultural land for 5-year return period
Vulnerability map of agricLLlluralland for 1O-year returnperiod
Vulnerability map of agricultural land for 20-year returnperiod
Vulnerability map of agricultural land for 50-year returnperiod
,
PR~C
No.
"606061
62
62
63
63
64
656667
"6869
6970
71
72
73
74
,
ADROBBSBWDBCEGISomDEMEMSSFAPFI'WSFIAGISHDHYVlIEC-RASICID
ILWISIWFMIWMNOAAPPCCPWMKgKmRAPIDSRMSDSARSPARRSOWARPOmSOBSq. kmUFDSMVRSAP
A88REVlA nONS
ApplicatIons Development Resean:h OpportunityBangladesh Bureau of StatisticsBangladesh Water Development BoardCentre for Envirorunental and Geographic Information ServicesCentimeterDigital Elevation ModelEmergency Medical Support SpaceFlood Action PlanFlood Forecasting and Warning SystemFederal Insurance AdministrationGeographic Information SystemHydrodynamicHIgh Yielding VarietiesHydraulic Engineering Center", River Anaiysis SystemInternationai Commission on Irrigation and Drainage, BangladeshNational CommiTIl:eIntegrated Land and Water and information SystemInstitute of Water and Flood ManagementInstitulc onVater ModelingNational Oceanographic and Atmospheric Adm inistrationProbability Plot Correlation CoefficientProbabilily Wcightcd MomentKilogramKilometerReal-time Acqui,ilion and Processing _ Integrated Dala SystemRoot Mean Square DeviationSynthetic Aperlure RadarSpace Research and Remote Sensing OrganizationWater Resources Planning OrganizationMeterSurvey of BangladeshSquare KilometerUrban Flood Dynamic Simulation M"de!Vietnam River Systems lLI1dPlains
•••-..
AoreAmlUl
Ails13.Aman
RAllSBoraDouble cropDry seasonGross cropped
"""Homestead area
Household
IntermediateholdingIntensity ofCropping
Kharif
Operated area
Permanent croparea
Pre_KharifRabiRabi cmps
Single cropSmall holding
T. ArnanTemporary crop"reo
ThanaUnionTriple crop
GLOSSARY
Unit ofland measure; 0.41 hectareRice grown in Kharif-2 ~eusonRice grown in Kharif- 1 seasonDeep water rice so\\'n during Kharif-) and harve~led illKharif-2 seasonBroadcast variety of rice grown in Kharif-l seasonRice grown during (he winter seasonOnly two crops cultivated on the same land in one year6 months; December-May inclusive'] he aggregate area of various crops raised in the samefarmland during the census yearHomeslead area means the area residence of the holder's withall its ,>tmclUres, courtyard, and Ihe land occupied by thep"ss"ge enter and exitA tamily unit who shan: common resources for cooking andeatingA farm holding having operated area 2.50-7.49 acr'"
It represents the ratio in which total arca of temporary cmp'>10the nct tcmporary cropped area.
Cropping ,>eaSOnbetween March Octoher, allen divided intoKharif-l (March-Junc) and Kharif-2 (July-October)The area owned by the holder plus the area rented from othersminus the owned given to others for operation.The part or net sown area with permanent crops or plantedwith fruit trees which occupy the land for a long period oftime ami do nol need for many years "Iler each harvest.March-April periodCropping season between November and FebruaryCrops grown ill the Rabi season such as pulses, oilseeds,spices, etcOnly one crop cultivated on the same land in one yearA fmm holding having operated area 0.05-2.49 acres, withminimum of 0.05 acre as cultivated area.Tran_'>plantedvariety of rice gmwn in Kharif-2 seasonThe part of net sown area where crop" whose growillg cycleor lenL>thoflife is less than one year
Administrative unit comprising a number orUnionsLowest administrative unit comprising a few villageMore than t,vo crops cultivated on the same land in one year
'"
ACKNOWLEDGEMENT
First of all, I would like to thank almighty Allah for giving me the ability tocomplete tbe work.
i would like to expre~~ my sincere and hcartiest gratitude to my ,upervisor Dr.G.M.Tarekul Islam, A"sociate Profes,or, institl.lte of Water and floodManagement, BUET, for hi" constant gl.lidanec, valuable advice, generoLlS hclpand constructive discus~ion to carry Ol.ltthis research. I consider my.<,elfto beproud to have worked with him.
1 am deeply grateful 10Md, Mizanur Rahman, Executive Engineer of BangladeshWater Devclopment Roard for his valuable suggestion~ and for providing data andrelated information regarding the study.
I also wish to express my deepest thanks to Mr, Zahidur Rahman, SeniorScientiilc Officer, SPi\RRSO and Mr. Raju Ahmed, Rcmote Sensing Specialist,CEGlS. Dhaka for their valuable commcnts and coo,tructive suggestions.
[ expre," my profound respcct and deepest sense of gmtill.lde to all of myrespeded tcachers ofIWFM. BUET for their fruitful adviccs in different times andvaluable teachings under diffcrent c()urses that have helped to reach my study atthi" stagc.
I am vcry grateful to Thana Statistical and Agriculture officer, Sundarganj ofGaibandha District and all the villagers of Tarapur Union participated in thequestionnaire survcy. I would like to acknowledge my best gratitude toBangladesh Water Devel<:>pmentBoard for their iruitful advices in differenl timesand for pmviding data.
The author acknowledges Library Assistant and other staffs of IWFM for theirkind cooperation,
S.M.lmrul HasanN(lVemhcr,2006
Xll'
,
•ABSTRACT
Tills study has been taken 10develop vulnerability maps for agriculture using RemoteSensing (R:'.) and Geographic Information System (GIS) regarding different returnperiod for flood hazard vulnerability asse."rnenl in a flood prone area, farapur llnio~()f Sundargal1i thana under Gaibandha district of Ba~gladesh, The study mainlyfocuses on the development (If a vulnerabilily function for preparing vulnerabilitymaps for agriculture.
For tlte development of vulnerability function depth-damage relation has beenfollowed. For establishing relationship between flood depth and agricultural damage,an extensive field _,urvcyhas been carried (lut in the study area. Satellite images andother relaled data are also widely analyzed for establishing the relation morefruitfully. A Landsat and a Radarsat image of the sllldy area nave been u,ed 10identify agrieullllral land and for identification of flooded and non-flooded area.
Although the study area is located in a flO<ldplainlocality but the terrain is not somuch flat. Due to lhis, depth of flood level varies in different location correspondingto a fixed level rise of water leveL In this regard for the development of a depth-damage function, the ,tudy area has bocn divided into four categories based on lheelevation. In monsoon (July to September), one meter rise of water level leads 100%damage of all the crops in the very low lying river associated land. Wherea, about 2meter rise of water level causes 100% damages of ali the crop" in low iand aml 2.5meter and 3 meter ri,c of water level; eauses 100% damages in all medium and highland crops of the study area respectively.
For the deveiopment of flood inundation and floo<! depth map, Digital ElevationModel (DEM) and water level data of the study area have been collected formBangiadesh Waler Development Board (AWDB). Water levels are used based onflood frequency analysis for 2-, 5-, 10-, 20-and 50-year return periods,
The agricuitural land of the study area has been found in between the clevation>22,9to <25.9i meter, Vuinerability fun~tion was ba..ed upon these land areas. Themagnitude of vulnerability for agriculture of Tarapur union is found to increaselinearly up to three meter dcpth of water and then it becomes horizontal.
From the vulnerability maps it has been observed that for ali the return periods, thestudy area is very much vllinerable, And area cio~e to river bank is relalively morevulnerable compared to the area tur away from the river bank. Flood levei dlle to 20-and 50-year return periods damages 100% damages of crops.
xiv
__ 1'--
CHAPTER ONE
INTRODUCTION
1.1 Background "fthe Study
Flood is the major disaster affecting many countries in the worid year after year. It is
an inevitable natural phenomenon occurring frnm lime to lime in ali rivers and natural
drainage systems. ]( causes damage to lives, natural resources and environment as
well a< the ioss of economy and health. The impaels of floods have been increased
due to a number of factors such as in~reascd development a~tivities on floodplain.
Fioods cau;e widespread suffering' and severe stresse; on sustainable and equitable
development. Such disasters exacerbate poverty through disruption of Iivelihood<,
environmental degradation, destruction of vital infrastructure and services and loss of
ceonomic inveslment in agriculture, aquacuiture and personai property.
'Flood Disaster' derives most of its meaning from 'disaster', which is the term often
given 10the harmful event that occasionally ari",., a, a consequence of huard, Thus
according to Parker and Cannon (2002), a hazard is a perceived natural event which
threatens both life and property and a disaster is the realiLation ofthi, hazard. In other
words, natuml hazard is the probabiiity of occurrence of a potentially damaging
phenomenon within a 'pcelllcd period of time and within a given area,
Flood disasters are e"traordinary events that cauSCgrcat destruction of property and
may re,uit in death, physical injury and human suffering, Flood disa,ter is an cvent
which afflicts a communily, thc consequences of which are beyond thc immediate
financial, materiai or emotional resourccs of the C[lmmLLnity.Disaster has also been
defined as usually [lverwnelming cvents and circumstances that test the adaptlllional
responses of community or individual bcyond their capability, and lead at least
temporariiy to massive disruption or function for community or individuai (Parker
and Cannon, 2002). This definition impiies that communilies and individuals have
•
2
differential resilience or capacity ta cope with disasters, and this again highlights the
importance af the concept of vulnerabilit}.
Ollila and Hearth (2004) presented an analysis of flood disaster trends in Asian
countrie, during the last 30 years, The analysis was carried oul to understand the
changing characteristics of flood disaster; in Asian countries ill the past three decades
with the rapid change in socio-cconomlc conditions. 1l show~ lila! tlood disasters in
most of the Asian countries are increasing significantly, especially in the last few
years A,ia witnessed a rapid increase of flood disasters,
With (he increase of population, industrial growth, more and mOre settlements and
development in floodplains, tlood haUlrd:. in Bangladesh have become ever-
increasing natural disa,ters relulting in highest economic damage among all kinds of
natural lIisasters. In order to fonnulale Ilood management and control policies, a
thorough understanding of Ilood problems and how it can be solved is required. Thi"
can be donc through data collection and analysis along with flood impac! assessment
(A~i~ et aI., 2002).
The magnitude and duration of [loods have changed during lhe lasl few decades. The
durati,," of 1998 flood was over 70 days. In i999 a prolonged flooding condition
prevailed lhroughout much ofthe mOllW,,"season.
The lOlalarca covered by major flo()ds ha, been steadily in,rea8ing sincc 1974, with
un exception of 1984 t1oods. The land area a/reeted by major Ilooll, has increased
from 35% in 1974 to 68% in 1998,
Riverine tloods occur when the amounl of runoff originating in a walcrshed exeeed8
the carrying capacity of natural alld ,onstructcd drainage system. Flooding can occur
due 10river overflow or surface runoff. There are two types of floods which occur III
Bangladesh: annuall100ds (barsha) that inundate up to 20% of the land area; and low
frequency lloods of high magnitude that inundalCmore than 35% of the area (bonno).
While the annual floods are essential alld desirable for overall growth of the
3
Ba~g:ladcsh delta and the economy, the major floods that ()Ceurred 1"1954, 1955,
1974,1984,1987,1988,1993,1998, 1999, and 2000 were very much destructive.
The rivers in the extreme northern area (Tccsta, Dudkumar and Dharla) have ;leeper
gradients (1 in 2000) than elsewhere and mosl of their catchments lie in India and
Bhutan. The"e river, frequently cause flash floods. The part of the region along
Brahmaputra suffers severely from river flooding caused by breaches, mainly in the
main Brahmaputra River and to limited scale in the Teesla River. The severity of
flooding may be exacerbated by rainfall within the regieln.(lClD, 1997).
Riverbed aggradation is so pronounced in Bangladesh that changes in riverbed level
can be observed during one's litdlme. Riverbed aggradation rcduccs the water
carrying capacity ofriyers, causing floods due to bank overflow. This recent increase
in riverbed levels has undoubtedly contributed to the increased tlooding propensity in
Banglade~h.
Flood control measures in Bangladesh are mainly limited to buildIng of earthen
embankments, polders. and drainage. Embankments and polders have reduced
floodplain storage capacity during floods, leading lO an increase in water le~e1and
di,charge, in many rivers. Earthen embankment, can easily breach and can be
damaged by riverbank erosion. Most of the embankmenl< in Bangladesh have
experienced breaching and erosion more than once since their completion.
Embankments have created a false <;enseof security among residents living within
embanked areas. The effeetivene,>, (}f embankmcnts is being questioned in other
countries as well.
Agriculture lS the essential source of income in the developing countries like
Bangladesh. Especially, the poor people sutrer the most from flood disasters because
they tend to live in marginal area.' and depend on high-risk, low return livelihoo<!
systems such as rainfed agriculture and face many sourees of vulnerability. The
preferred areas for agriculture are river banks and floodplains due to their natural high
productivity and availabillty of water, but during flood events high percentages of the
harve,! orten gelSdamaged.
4
The study area Tarapur union of Sundarganj thana under Gaibandha district is
subjected to flood almost every year from both the boundary rivers of the Teesta and
the Brahmaputra. with roth beneficial and adverse c/lccts, In the year 2004 a heavy
rain fall for six to seven consecutive days in July and upstream river flow caused a
devastating flood in the locality. The river Teesta is responsible for bolh lypes of flash
flood and monsoon flQod in the study area. The river Brahmaputra alw has a
significant impact on both types of flood in Tarapur union. Riverine monsoon flood is
very much destructive and occurs vcry Irequently in the study area. Fla,h flood is also
destructive but does not occur so frequently. Nonnally monsoon flood occur in July
to September.
The economy of the sludy area is mainly agrobased and alm'''l 82% arCaof Tarapur
union is now being used as agricultural land. Among the crops, paddy occupies aboul
74% of lhc gross cropped area and normally grown in thc low-laying areas (BBS,
1996). Monsoon flood leads a huge damagc to agriculture almost all the year.
Considering tbat damage to agriculture in Tarapur union is most ,ignificant of thc
total damage, empbasis is given ill this 'ludy to asscss flood hazard V\llnerability of
agriculture for mOnSOOnflood.
For an effective managdncnt of flood in low lying fl(}()d-pronc areas, Geographic
Infonnation System (GIS) and Remote Sensing (RS) technology i, proving to be a
useful and efficient tool (Wagner, 1989; Wu and Xia. 1990; Rahman, 1992). The
existing inadequate communication system and high acce"ibility of remote sensing in
case of extraction of in/ormation make it a useful tool for flood study. Satellite images
can be "'I'd for preparing !lood inundation map, monitoring of flood and estimation of
flood damage.
For tlood disaster risk management, it is important to analyze and assess vulnerability
of different elements at risk. The flood inundation maps are very much useful f\1f an
effective management of flood while the flood hazard vulnerability maps for
agriculture are also u,cful for flood risk management. Vulnerability function relates
hazard magnitudes wilh vulncrability and is the basic ingredienl of flood hazard
".,,-
5
vulnerability maps. Hence for beller management of flood and for selection of
location for agricultural and other activities, flood hazard vulnerability maps of could
be used.
In (his sludy gcoinformatics is applied to assess flood hazard vulnerability in a flood
pmne area Tampur union. Flood hazard vulnerability of agriculture arc as>csscd for
different flood magnitudes through development of flood inundalion maps, flood
depth maps and vulnerability function.
1.2 Rationale of the Study
Flooding cannot be completely avoided, but damages from severe flooding can he
reduced if effective flood mitigation scheme is implemented. This can be achieved if
the ,ufficienl infomlalion regarding flood can acquired both in time and in quality.
Ilydrologic applications of GIS and R~ll1ote Sensing rangc from synthcsis and
characterization of hydrologic tendencles to the prediction of response to hydrologic
~venlS, Th~ payoff comcs from lhe mulliplc ways in which thc d~ta c~n be used once
it is made to be digitally accessible in a GIS format. Use of GIS and R~mol~ Sensing
will provide supplementary data in hydrology for such analysis and will lead to easier
interpr~lation and und~,"landing of r100d phenomena and characteristics.
1.3Ohj~..,tives"f the Stud).
The major objectives of the study arc as follows:
1. To dev~lop flood inundation map' of [he study area.
2. To as"e>s flooding (ar~al ~xtenl, d~plh) for differenl flood magnitudcs.
3. To develop flood hazard vulnerahilily funclion of agriculture.
4. To as",ss flood hazard vulnerability of agricullure for differenl flood
magnitudes.
CHAPTER TWO
LITERATURE REVIEW
2.1 Introduction
The people of I:langlade,h are repealedly confronted by natura! and humulHnade
catastrophes such as flooding, While some of lhese environmcntal degradations and
calamitie, are nOldirectly related to human activities and landuse practices (such as
earthquakes. tornadoes. and cyclones), other, are relalcd 10 human Interactions with
the nature. Flooding is OnCwaler-related environmental problem that is direelly
related to human activitics. The magnitude and intensity of flooding are very much
dependent on landuse practices in the watershed of each rivers or streams. Finding
solutions to the flooding problem i, imperative. It is important to understand the
processes that influence t100ding in [Iungladesh, before any solulion 10 lhe flooding
problem is prescribed (Khalcquzzaman, 2000).
The literature review is primarily directed to an overview of application "f Remote
Sensing (RS) and Geographic Information Sy,lem (GIS) in present and past flood
study in Bangladesh as well as in other parts of the wmld, The aim of this research is
to develop a methodology using geo-infonnatics (RS and GIS) for vulnerability
assessment for agricullure. For developing vulnerability function it is needed to
realize damages caused by flood. Again for flood hazard vulnerahility asses,ment an
understanding about thc risk analysis and the concept of hanrd factor, and
vulnerability factors should be clear. So the review of literature includeB here ail the
factors along with Remote Scnsing and GIS.
2.2. BriefHistor~' of GIS and RS in Bangladesh
The complex and dynamic environmenl of I:langlade,h, ilS susccplibilily to nalural
hazards and hcavy dependency on natura! resources makes spatial infonllation a
crucial ingredient for the development process (Huque, 2000). The existing
communications infrastructure is inadequate to support efficient c"ilection and
dissemination of information regarding management of natural calamities like flood.
7
Therefore, aetivilie~ related In photograrnmetry, Remote Sensing and spatial
informal;,," science have been developing over the past three and a half decades
(Huque. 2000). Regarding this. in this chapler. especial emphasis is given On the
application of Geographic and Information Syslem (GIS) and Remote Sensing in
flood management.
Photogrammelry and Remote Scn,ing in Bangladesh has been growing al a moderate
pace but was limited to specialized national agencies until recently. During the
nineties, increasing numbers of development activities within the country and
tremendous technological advances worldwide have caused a major increase in the
use of these technologies. Practical applications of Remote Sen,ing pr{lducts started
filling the gaps in available up-to-date information. Huge amounts of spatial and
attribute information are being digitized in order to make best usc of GIS teehnoiogy
for enhancement of planning and management capabilities in the count')'. The nineties
have aiso seen a iatge increase in iocal professional; capabie of dealing with these
technoiogies. However, the activities are still mostly limited to government
organization, and dOllor fUlldedprojects. The major agencies that have been working
in tbese fieids are Space Research and Remote Sensing Organization (SPARRSO),
Survey of Bangladesh (SOB) and Centre for Environmental and Geographic
Infonnation Services (CEGIS).
2.3 Use of GIS and RS in Flood Study
Khatun (2005) conducted a study to develop a melhodoiog)' for soeio-eeonomic
vulnerability assessment and to create a 1,,55 map using GIS and developed
vulnerability indices for a scenario flood event for urban flood disaster risk
management in Hanoi and Bangkok. the capitals of Vietnam and Thailand
respectively.
Sbrestha et aL (2004) mapped flood risk and vulnerability using GIS. Hydrologic
analysis was combined with socioeconomic reSOUrCeSand constraints with tbe help of
8
GIS data processing software Arclnf" and ArcVicw 3.1. Landsat imagery at 30 rn
resolution and lime series areal photograph. were used to detect change in river
morphology and 10 identifY flood plains. DEM wa, incorporaled inlo Hydraulic
Engineering Center's River Analysis System (HEC-RAS) software for flood elevation
estimates. Primary data on past hazards, socio-ce<.momic condition. vulnerability and
response capallility in the recent past were collected III develop vuln~rab;lily and risk
map. For vulnerability four sets of paramelers were chosen: Houses, Built-up area,
Landusc and land cover and Road Infra,lruc!urc, To develop the vulnerability map,
parameler maps were transformed into weight map' b} assigning a welghtage to each
class of parameter. Risk was obtained by the multiplication of cos!, vulnerabiiity and
recurrence interval of the natural phenomenon.
Gupta et al. (2004) reviewed flood damage assessment in the Mekong della, Vietnam.
The Vietnam River Systems and Plain< (VRSAP) model was calibrated with the year
2000 flood data to predict depth, duration and extent of inundation at ditl"erent return
periods.
Dulla and Tingsanchaii (2003) developed loss functions in <,orneseiected regions of
Bangkok for urban flood risk analysis. Questionnaire surveys were carried Ollt in two
selected districts 011residential and non-residential buiidings for gathering data and
information tor development of loss functions. Based on the estimated damage to
residential and non-re,idential buildings due to hypothetical flood" of different
durations, hypolhetical Joss functions were established for the foilowing four
categories.
• Categoryi: Damage to building struclures,
• Calegory2: Damage to building contents! stocks,
• Category3, Damage (0 outside properties,
• Category4: Emergency and cleaning cost, due to floods.
9
In this study it was observed that damages to different categories significanlly varied
with the duration of floods and due to that flood duration is considered as one variable
in loss function develaprnent.
J)utla el al (2003) presented an integrated mathematical model for di,lriburcd
modeling of flood inundation and damage. Based on the landuse patterns and nature
of damage they developed mathematical flood damage models for three types of
landuse as Urban, Rural and Infrastructure. The hydrologic model consider< major
processes oflh" waler cycle through physically based governing equations. The 10%
",lirnation model is formulated based on stage-damage relationships between
differenl flood inundation parameter, and landuse features, For urban and rural
damag"', the damage estimation models are formulated as grid ba<;edmodel and are
developed for each catcgory. This model was applied in Ichinomiya River Ba,in,
Chiba, Japan.
Dulla (1999) prcsented the methodology for flood damage assessment using GIS and
Distributed Hydrology Model and a case study in lchinomiya River Basin, Chiba,
Japan. The Hydrologic Model which includes tour major componcnts such as
overland flow, river flow, unsaturated zone flow and saturated zone flow wcrc uscd in
hi, 'ludy for flood modeling.
Hutchison and Watt (1978) presented a ,ystcmic approach to the flood risk mapping.
In thai ,tudy it was concluded thaI:, flood risk mapping was the corner,tonc of the
new, comprehensive approach to the reduction llfflood damage both in Canada and in
the United states. Ngoc (2003) prcsented the implementation of GIS technology in
Urban Flood Dynamic Simulation Model (UFDSM). The main charaderistics of
UPD:'>Mare ,imulation of tile detailed flood process. therefore, the result, can be
applied in many aspeds ,ueh a, flood loss calculation, flood hazard map, regillnal
planning, flood insurance, flood control planning etc.
10
2.3.1 Application afGIS and RS in Flood Stud~' in Bnngladesh
2.3.1.1 GIS lind RS in Flood Mapping and MODitorin!!:
In Bangladesh, during the last decade several initiative, have been taken up by the
Flood Action Plan (FAP-19) in oroer to monitor the flood process by using
geographical information.
In 1993 the use Df lime-series of ERS-l irnage~ led to c1assificat;on of flooded and
non-flooded area, The results have been published in internal reports (fAP-19, 1995)
and at several symposia and in the European Space Agency (ESA) quarterly journal.
In 1996 Rada~at images have been used in the Applications Development Research
Opportunity (ADRO) "",earoh programme for a feasibility study on flood monitoring.
In the ;Iudy detailed measurement, with Radarsat Fine beam images (8 m re~oluli(}n)
and Radarsat \Vide beam images (25 m resolution) have been generated. Under
European Space Agency Data User Porgramme (ESADUP) funding a transportable
ground receiving station was temporarily installed at SPARRSO in 1999. During the
monsoon ,everal SAR images "ere acquired and processed in near real time for flood
applications with help of Centre lor Environmental and Geographic Information
Services (CEGIS). The data was processed near real time into flood information
products and ha, 'erved as an eye-opener ba,ed on which a number of other
promising applications were investigated briefly (Valkengoed et al., 200 I).
Floods inundate 30 percent of the country in normal lllonSOOnyears. and occas;onal
excessive floods inundate lllOre than 60 percent of the country (BWDH, 1998).
Bec"u~e of dense cloud-cover during the monsoon t100ding period, (SynlhetieAperture
Radar) SAR has proven to be the most reliable tool tor mapping and monitoring the
dynamics of the flooding process, EGIS carried out research in 1993 (FAP19, 1995),
1996 (EGIS, 1997) and 1997 (EGIS, 1998) using the European ERS-l and Canadian
RADARSAT-l SAR images (under ADRO funding) and develop a melhodology to
map open water flooding. Results from RADARSAT SAR show over 90 percent
agreement witb t1cld observation, for the open water cia,s derived from digital
II
cia,sification. During the catastrophic floods of 199R,RADARSAT ScanSAR Wide
und ScanSAR Narrow image" covering almost the entire country. were used in near
real time m",le. Flood maps sllOwing "pen water tlooded areas and percentage of area
flooded were produced for disaster management and relief distribution.
In 1998, SPARRSO conducted a flood mapping and monitoring project with
RADARSAT images under an Applications Development Research Opportunity
(ADRO) project or the Canadian Space Agency. Following the fiood a project was
conducled, again using RADARSAT data to ~tudy the progress in post-flood
agrlcultuml rehabilimtlOn,
In 1998, a pilo! study area in north central Bangladesh wa, selected and
methodologies have been develo~d for computing flood extent, depth and duration
using a lime series often RADARSAT F3 SAR images, digital elevation model, <TIS
data and hydrological information. Land use/land cover pattems were generated from
the SAR images with the support of ground truth information, The output of this study
can be used to assist in the planning and management of floodplain resources in
Banglade,h.
Real-time Acquisition and Proccssing-Integrated Dala Systcm (RAPIDS), a low-cost,
small ,ized, ea,ily transportable, PC based ground receiving station was set up tor the
monsoon of 1999 for ninc months to receive ERS SAR images for demon,trating ncar
real time flood mapping and monitoring, A largc archive of mulli-temporal ERS-2
images both in asccnding and descending mode were collected for a major part of the
country and proccssed in near real time. The images were applied in flood extent-
tlood depth, crop and shrimp fann mapping. Thc station was set up at SPARRSO and
the projecl was carried out in collaboration with SPARRSO, National Aerospace
Labmatory (NLR) and SYNOPTlCS of lhe Netherlands with funding from the
Europcan Space Agency.
12
From 2002 onwards EGIS is aiming to be an independent organization for
environmental and geographical services in the ncar furure. The centre is continuing
to provide services for several applications in coastal and river areas in the field of
tlood and agricultural monitoring. Co-opcratlon with several partners (Resource
Analysis, Synoptic>, NLR) in internationally funded project, has resulted in
mcthodologie~ to map flood extent, depth, duration, flooding direClions and crop
classification on the bm;i, of space borne SAR and optical imagery_ During the
monSo(m season, typically from July-November. radar (SAR) satellites provide
guaranteed imagery as opposed to optical satellites, The acqui,ition of SAR images is
nOlhampered by cloud coverage and ground receiving station can guarantee Frequent
data availability. The immediate handling of image,>, covering the whole of
Bangladesh in times of emergency, is essential For Fast response mea,urCS by local
aUlhorities and validation in the field.
Estimating the extent of floods In Bangladesh a study was conducted by Blasco et al.
(f992). In that study SPOT Sa!ellltc images were u,>edForcontinuous monitoring of
flood through measuring the area inundated rcgarding different time period. Sado and
Islam (f997) conductcd a study on satellite Remote Sensing data analysis for tlooded
area identification of Dhaka city. Bangladesh. Rahman (1992) carried out a research
work by using GIS, Remote Sen,ing and Models for flood ,tudics in Bangladesh. This
was an analytical study in a flood prone polder in Bangladesh. The srndy mainly
focused on monitoring: of flood. Rasid and Pramanik (1990) conduded a study on
visual interpretation of ,atcllltc imagery for monitoring floods in Bangladesh.
2.3.1.2 GIS and RS in Flood Risk and Haznrd Mapping
Bamn and Nagasawa (2003) reviewcd risk assessment and empowering community
heahhcare facilities in coastal citie, of Bangladesh. Thi. study focused on the di,a,ter
risk caused by flood and cyclone, its post-disaster diseases and emergency medical
support system that needed within thc existing shelter-cum-community healthearc for
sustainable development. The asse<>mentrcsult shows that tbe community healtbeare
13
facilities with Emergency Medical Support Space (EM~S) can pia) an effective role
"" a part "f long-term disaster mitigations with the other r~overy activiti",_
Islam and SOOo(2000 3, b) ,[udied flood hazard in Bangladesh using I\OAA AYHRR
data with GIS. In that study hanrd a"e~_,rnenl was made considering two major flood
hazard parameters (depth of flooding and duration offlO<lding). Satellite images were
used for the prediction of tlood depth and duration. ~il1ally, hazard map was
developed considering combined effect of both parameters. The flood hv.ard map
provides intormalion for the developmenl of counter measures and preparation of
high ri,k areas, on a priority basis, against flood damage. Phy,iographic divisions,
geological divisions, land cover categories and drainage netwmk dala were used as
GIS components.
Islam and Sado (2000 c) also carried out a research work on Satellite Remote Sensing
data analysi, for nood damaged zoning with GIS for flood management Pramanik
(1994) conducted a study on Remote Sen,ing Appliealions in Disasters Monitoring in
Bangladesh. Thi, ,ludy mainly discusses the applications of Space Sciences, Remote
Sensing and GIS by variO\lS agencie, in Bangladesh
Ochi el aL (1991) conducted a study on flood ri,;k evaluation in Bangladesh usIng
Renl()te Sen,ing and GIS. This study deals by using NOAA-lOA YHRR imagery with
other thematic maps by Geographic Information System (GIS) to estimate flood
damage area in 1988 and also calculate ,Orne thematic information inundated by
floods.
2.3.1.3 Application ofGlS aud RS in I<'lood Study in Meghna - Dhonagoda Polder
Brouder (1994) of International Institute for Aero'pace Survey and Earth Sciences,
Netherlands conducted a research on flood Study in lhe Meghna - Dhonagoda Polder.
Bangladesh. In this study spatial data in the digilal data ba,e of the GIS, such as
digital elevalion model (DEM) were used In predict the effects offulllre event,. SPOT
Multispectral CCT of 10"' October 1988, also was taken during the period of flooding
provIde an effective means of mapping the extent oflhe flooded areas.
"The main objective of thaI Slud~' Wll$ to 1lSSt'SSGIS and RS 11510015 for mannging
flood - prone 1l~IlS in Illlngilldcsh. In thaI study II comp:lTison of the 1988 flood maps
of the study area obtllincd b)' SllICllilC imagery and the DEM was shown. And Il
calculation of the acreage of the orcn floodc'd and the depth Ilnd volume orthe flood
Wlltcr of the stud)' Ilrell for Ihe 1985 flood Well: shown. The effeet of flooding (area
eXlent, dcplh nnd mlumc) 115swning difTt1ent high wllter lcvclllnd different IOClllions
of dike failun: WIlSIl[SOidentified in !hal study.
As the flood level on HI!!>Octcber 1988 WllSbetween 3.00 lind 3.50 m a.m.s.1. twO
flood maps were created in thlll stud)', One for II WfileTlevel of 3.00 m and one for 11
waler lCI'cl of3.5llm. Figun: 2.1 shows the flood map l'l:sulling from the o~mlion on
the J)EM due 10 3.00m waler level. Ala wlIter level 3.00 m 83% or the lll'l:1l""'IIS
estimated 10 be f101X1ed.
51 neetS"
!" ,.•.----
Figure 2.1 Estimated flooded area. using II DEM (High II'liter level of3.00m)[Sour«: Iirouder (l994JI
By using the simple technique ofdcnsily slieing (SPOT image of 10l!lOetober 1988)
flooded and non_flooded a= II'~ differentiating in lhat sludy. Figure 2.2 shows the
flooded and non-flooded areas obtained by sli,ing the near infmred b:md image of 10
Oelober 1988. It appeared tllat 83% oftlle area was flooded.
"
tH
• Km
l'igu~ 2.2: SPOT im~se classified 11~flooded and non-flooded areas, using densityslicing. (Source: Bmuder (1994)]
In order to compare the flood ml1~obtained b>' processing oflhe SPOT image!)' and
the OEM, these lwo products were overlayed (Figure: 2.3) Dnd the diffcrcn<:cs and
similarities Wtll' l1nal)'7w.
1"iN
Figure: 2.3: Comp:lrison between flooded Ilreas on image lind DEM (flood kvd 3 m)(Source: Broudcr(I994)]
16
Flood statisti" i, essential for the management oftlood prone arCaS and to assess the
expected consequences of possible flooding events. Flood statistic, for a flood level
on.oo m and 3.50 m found in that >Iudy is given below.
Table 2.1: Flood statistics for a flood level oD.OO m and 3.50 m found in that study is
given below.
Flood lenl (m) Arell Flooded (million hal Volume of flood wllters (million m )
3.00 2.89 1634799
3.50 3.16 2443454
Source: Srouder (1 <)<)4)
2.3.1.4 Flood Study on Sundarganj Thana
Aziz ct al. (2002) conducted a research on Suodarganj Upa-Zilla ofGaibandha district
to develop an integrated approach to disaster mitigation in Bangladesh through
dynamic flood warning system. In [hat study Hydrodynamic model MIKE l1-FF
(NAM. HD lind FF) was successfully integrated with GIS in the Arc view GiS
environment. flood inundation m~ps were being created for Sundarganj (h~n~ by
overlaying landusc features and infrastructures to delineate flooded areas u,ing
dynamic spati~i modeling.
2.4 Vulnerability alld Risk
Vuincrabiiity depends on the m~gnilUdes of hazard like depth of flood water. An
under<;tanding about the relationship among vulnerability, risk pcrception and
damages due to ha£ard and an appropriate estimation of dalnage i~ essential tor
a,sessment of vulnerabi lity of an arCathrough.vulnerabi lity function. Risk perception,
hazard magnirudes and slatu, of damages largely detcnnine the nature or
chamdcri,tics of vulnerability fundion, Flood damage, vulnerability and ri;k
perception eonter speeifLcinformation about hazardous threats to the various elements
at ri<;k.(e,g., Alexander 1993; Heyman et .11.1991).
17
2.4.1Vuln"rabili~'
Vulnerability is a set of prevailing or consequential conditions composed of physical,
socio--economic or political factors which increase a community'.' ~u>ccptibility to
calamity or which adversely affect it, ability to respond to evenlS. 10 general view
vulnerability indicates the degree of loss to a given elements \lr set of elements at risk.
Vulnerability can be con,idcrcd in terms o[five components.
1. Initial well-being
2. SelfprOlection
3. Social protection
4. Livelihood resilience
5. Social capita!
The actual amount of flood damage of a specific flood event depends on the
vulnerability of the affected socio.cconomic and ecological systems, i.e., broadly
defined, 011 their potcnliallO be harmed by a hazardous event (Cutter 1996; Mitchell
1989). Generally speaking. an clement at risk of being hanned is more vulnerable, the
more it is exposed to a hazard and the more it is susceptible to its forces and impacts.
Therefore, any f1(}od vulnerability analy,is requires information regarding these
factors, which can be specified in terms of element-at risk indicator" exposure
indicator, and susceptibility indicator,. In this regard, natural and social science
indicators are highly significant.
2.4.1.1 Element-at-Ri,k Indicators
"I he subject mallcr of any flood vulnerability analysis i8 the group of eJement8 "hich
are at risk of being hanned by flood events. Element-al-ri,k indicators specify the
amount of social, economic Orecological units or 'y,tems which are at ri,k of being
affected regarding all kinds of hantrds in a specifiC area, e.g. persons, hou,eholds,
agrl-flrm~, economic production, pri~ate and public buildings, public infrastructure,
cultural asset', ecological species and landscapes located in a hazardous area.
The elements arc at risk of being affected by flood evenIS, thc magnitude of damagc
can be e<timaled in monetary and non,mo'nctary units, which arc rcfl~ted in total the
maximum possible fload damage. This i, aiso eailed damage pOlential. And becau'le
every element at ri,k i, more or less exp",ed to tlood event, and more or less
susceptibie to them, exposure and susceptibility imlicators are always reiated to
e1ement-at-ri,k indicators and contribute significantly t(l lhe analysis of flood
vulnerabiiity.
2.4.1.2 }~~posure Indicators
Indicators supply infmmation about the location of the various element, at risk, their
elevation. their proximily to thc river. their do,encss to inundati(ln ar"aS. rdum
period, of difTercnt types of flClOdsin lhc floodplain. 'faken together, tllese indicalors
inform us of the frequency of flood< in floodplains and of the threat to the various
clements at risk of being inundated. Indicalors of the second category focus on
general flood characteristics likc duration, vel"city, scdimentation load and
inundation depth.
2.4.1.3 Susceptibility Indicators
Susceplibiiily indicators measure how sensitively an e1em"nt at risk behaves when it
is confronted wilh some kind of hanrd. Regarding social and economic systems, an
important group of indicalors refer to susceptibility in a narrow sen<e. measuring the
absolute or relative impact offloads on individuai clements at risk. For example, the
impact of inundation depth and flood duration on crops is frequently a major issue of
damage analysis and re>earch, attempting t(l id~ntify crops types that feature similar
,u,ccptibiiities. And thi, makes sense, because paddy is much more susceptibic to
floods than jute. Vulnerabilily can be detined in terms of functional relationships
belWeen expected damages regarding ail elements al risk and the susceptibility and
exposure ,haracteristics of the affected 'y~lcm. referring to the whoic range of
possibie tlood hazards.
19
2.4.2 Flood Risk Analysis
2.4.2.1 Risk Perception
The notion of risk perception refers to the intuitive risk judgments of individuals and
social groups in the context of limited and uncertain information (Slcv;c, 1987).
These judgments vary between individual> due to ditlerent levels of information and
uncertainty, due to different intuitive behaviour, and also due to specific power
con,lellalions and positions of interest. As a consequence, the individuals of a
community may as.'''SSthe risk orbc!ng tlooded very differently. Meausc they do not
have the same information about the probability of flood hazard events in their region,
about f1o",1 mitigation measures and their effectiveness, and they perhaps have a
diftereD! historical backgmund regarding lhc experience of living in a floodplain and
of being flooded. In case of the very diverse risk perceptions within society, a
communicalion process on Hood nsk and flood risk perception ,hould be encouraged
as a basi, for policy, [f prevailing perceptions and value concepls become lransparent
and open to public debate, a commOn perception of communities may evolve and
conlribute to an increased acceptance (If fl,md protcclion policies and for development
a vulnerability function for any lield.
2.4.2.2 Risk Assessment
Risk lIS,cssmenl means the integration of specific ri,k of a ha£ardous system of
activity and its signit1cance in an appropriate conlext United Nalions (1991) outlined
a methodology of risk assessment for mitigating nalural disaslcrs in a manual for
policy makers and planners. As per tllis manual, the risk is expressed by the product
of specific risk and elements al ri,k.
2.4.2.2.1 Spcciflc Risk (Rs)
The specific risk (Rs) mcans the expected degree of loss due to a particular natural
phen(lmenOna, a runction ofbotll natural hazard and vulnerahility.
20
2.4.2.2.2 Total Risk (IU)
Total ri,k indicates the expected number of lives los!, persons injured, damage to
property, or disruption of «anomie activity due 10a particular natural phenomenon
and e,,-pressedas the multiplication of clements at risk and ,pccific risk. The intensity
of natural hazard depends on flood dcptll while vulnerabi lity depends on population
density. '] his methodology was used f(lr risk assessment to prepare comprehensive
disaster management program under relief Ministry of the Government of
Bangladesh.
2.4.3 Relationship among Flood Damage, Vulnerability and Ri,k Perception
2.4.3.1 Relationship bem'cen Flood Damage and Vulnerability
Hood damage analysis alms at quantifying flood damage, for specific future
scenarios with clifT"r"nt !lood events and flood policies in order to quantify the
bencfits of tlood protection measures and, thereby, ,upporl policy decisions. In thi,
context the concept of damage potenlial is crucial. ['he damage potential of a specific
area repre"",ls lhe maximum possible amount of damage which may occur if the area
become, inundated. In these analy,e, vulnerability aspects must be considered in
ordcr to estimate the proportion of the damage potential which will finally
materialize, i.c. to detennine expeded damage,. In somc vulnerability analyses, such
a factor i, dcrived trom expert knowledge and empirical data on flood damages and
then expressed on a scale between 0 (no loss at all) and I (lotal loss) in order to
quantif)' the expccted damage reduction for several categories of elements at risk
(e.g" Elsner et al" 2003; Glade, 2003). The most important vulnerabi lity indicalor for
estimating damages in currenl flood damage analyses is [he cxposure indicator
inundation deplh,
2.4.3.2 Relationship between Ri~k Perception and Vulllcrahility
With regard to the ,ocial and economic fealures of vulnerability, lhe notion of risk
perception is crucial, too. If (average) flood risk pcrception is low in a region there
21
would probably be taken any action 10 decrease the risk Or to prepare for ll(}()d
damage, vulnerability and risk of flooding. Conver,ei}, if people arc well aware of a
flood risk - perhap' because they experience a flood with varying severity lime and
again _ they tend to be beller informed and prepared (Messner and Meyer, 2005). As
a rule of Ihumb it can be stated that regions with low levels of tlood risk perception
and a [ow degree of preparedness for coping will1 flood events tend to experience
flood damage levels above average and vulnerabil ity to flood events is usually high.
2.4,4 Flood Damage Estimation
TIle loss e,timation duc to any natural disaster is a hypothetical, quantitative
description of the errect, of a future disaster (,) upon geographical areU - a city,
region, state or nation, Therc arc many flood variables which aftect damages caused
to agricullural lands, for example: water depth, duration of inundation, time of the
year. The relative importance of those factors will vary between different crops with
varying growing and harvesting c}c1es (Ngoe, 2003).
With different lypes of research works several types of flood damage are generally
recognized (e.g, Higgins 198i; Parker et al. 1987). First it is customary to distinguish
between tangibie and intangible damages based on whether or not monetary values
can be assigned to the consequences of flooding,
Tangible damages arc those, which can be evaluated in terllls of monetary value. But,
intangible damages are difficult to express in monetary values such a$ ill health,
anxiety, inconvenience and disnLption of social activitie,. Tangible damages are of
two t}pes, direct and indirect damages. Direct damages are caused by physical contact
of floodwater su~h as damage to crop,. Indirect damages result from the interruption
of community activities, including traffic flows, trade and industrial production
(Khatun, 2005)
Tangible damage~, hoth direct and indirect, are measured in terms ofm,metary values
in two ways, One is by de~eloping stage-damage functions, It relates tlood parameters
(ex. duration oft1ood, depth of inundation, etc,) with damage extent to different types
22
of properties (ex. crops, houses etc). Thc>c stage-damage funclions are derived either
from past flood dala analysis or through analytical descriptions of flood damage to
various properties.
Today digital landuse data like the digital landscape Illodel is frequently used for
flood damage estimation which allows its spalial implementation by meanS of
Geographic Intormation System (GIS). By intersecting maps of inundation area and
damage potential in a GIS and relating them accmding!y, the amount of valuables or
people afTectcd can be detennined. The vulnerability factor of lhe valuables, i.e, the
share that is expedcd to be damaged, i, in most cases exclusively related to
inundation depth. Hence, relative depth-damage functions are used to calculate the
expected damages
Gupta et al (2004) reviewed llood damage assessment in the Mekong ddta, Vietnam,
The Vietnam River Sy,tems and Plains (VRSAP) model has calibrated here wilh lhe
year 2000 flood data to predict depth, duration and extent of inundation at different
return periods_ Direct damages, indirect damages such as economic los>8s duc to
inlerruption of eCOnOnllC aClivilies, intangible damages .sueh as anxiety,
inconvenience, ill health and loss of cultural significance were con.\idered. Based on
these resuits, damage-frequcncy functions were devdopcd; functions were of lincar
fmm for residential and commercial sectors and are of power form for the agriculture
and infraslruClufCsectors,
Kharun (2005) presented a procedure for measuring tangible damage through field
survey. In surveying procedufC questionnaire ,urveys were conducted in nood
affected area among afTecled people to estimate 10" to properties_ Extensive rcsearch
works was conducted for deriving ,tage-damage lunctions from flood damage
analysis.
In case of the great varicty of methods of damagc analysis, the choice of an
appropriate method does not only depend on the size of the area undcr consideration,
but also on other factors like the availability of necessary data. time. manpower and/or
23
m\mey resources and not least 0" the goal of the respective -'IUllyand the management
level for which it should provide decision-making support.
'J he major considerations far estimating the flood damages in Bangladesh by return
period (HeA, 2003) are as follows:
Y The people of Bangladesh are accustomed to protect thelr properties in their
long lite experience to minimize the damages. They do not always consider
the flood to bring the damages to them a~ the negative benefits bllt rather they
a,sumcd to have utilized the po;itivc effects af f1(lod such as natural water
.<upplyby flood tor irrigation, new >oilflew twm upstream to fertile their farm
lands, fishing in the flooded Held or river and ~o On,
Y The crops are planted in the proper season to minimize the flood damages by
taking account of the characteristics of crop,;,
}> There are very slight /lood damages io !h~ flood of short term return p~riod
sueh as 2-and 5-year. Then it was assumed that olll} th~ indoor movables and
erop~ would be damaged in flood prone area of2-year of return period.}> The flood damages in catastrophic conditions, like the tlood that happened in
1998 of 50-}ear return ~riod, tend to be drastically increased.
•• The flood damages for infra,lruclure in the short return period are negligibly
smaller than those of longer return period.
y '1he flood damages are related te>the years or return period and flood prone
area by return period.
CHAPTER THREE
MF.THOI)OLOGY
3.1 ReCOnnllisslIDceSurvey
Reconnaissancc survcy is essential for conducting a research, for understanding lhe
flood statu,; of the localily and its impaels on different sectors like agriculture. It is
also essential for setting a outline for the ovemll 'lmly. For lhis sludy rcconnaissance
survey was undertaken with individual respundenl.\. allhough a number of
respondenls occasionally might be interviewed at lhe same time. The objective of the
survey was to quickly obtain basic infonnalion in developing an undcrstanding of the
area and hclped ensuring that the formal questionnaire was designed in a manner
under~landablc and rc1cvant to agrlcullUral damage due to flood, fanner<;'
circumstances and sensitive local i>sues relaled to flood.
ln addition to lhi" informal discussions werc conducted witll officers of Bangladesh
Water Development Board (BWOB), Institute of Water Modeling (lWM),
Bangladesh Space Researeh and Remote Sensing Organization (SPARRSO), the
agricultural officers and some farmers abour the exi51ingsituarion of the area.
3,2 Selection "fStudy Area
The sludy area Tarapur union of Sundarganj thana under Gaibandha district i; one of
the worst flood affcctcd rcgion of Bangladesh. From the reconnaissance survey and
Iilerature review, it Is found that the characrerisric.' of flood of Tarapur union are
almost samc with the floods of other parts of flangladesh. So a flood study on Tarapur
union using G1S and Remote Sensing will be a representative for the wh"le
Bangladesh.
Moreover as one of the basic requirements of the stud) area is a Digital Elevalion
Model (DEM) and a prepared DEM of lhe study area Tarapur union by BWDB was
helpful in laking the decision of sel~rioll of the stud} area. In selection of the study
area, availability of satellite image of the area for a required time period is also
essential. ln this aspect Tarapur union waS suilable. Moreover in ease of other study
25
related infomlation, existing communication system and smooth law enforcement
Tarapur un;,ln is more suitable for conducting the research work.
3.3 Primary nata Collection
A semi-structured questionnaire "as prepared to collect the primary data (Appendix-
A). The questionnaire was prepared on the basis of the objectives of the study mainly
for the development of vulnerability function curve for agriculture. In this regard for
selection of the interviewer>, framers were given more priority_ And the age of the
interviews was not less than 40 years b«ause the questionnaire was used to obtain
information on the situation of the sllIdy area in the pa,l. And wider range lime frame
infonnation about flooo of the study area will be more helpful in developing a more
precise vulnerability ftlnctlon for agriculture.
The total number of sample was selected randomly and 48 No. of sample was
distributed randomly among tbe villages of Tarapur \lnion,
3.4 Secondary Data Collection
Digital Elevation Model (DEM) of the st\ldy area was collected from flood
Forecasting and Warning Center of Bangladesb Water Development Board (BWDB),
Tbe resolution of DEM Is 50 meter. DI:M was \lsed to develop flood inundation maps.
Satellite images were collected from Space Research and Remote Sensing
Organization to devclop flood inundation map and landuse map. In tbis regard,
lLW1S software was used (ILWIS 3.0 Academic, 2001). Water level data of the year
1988 to 2004 ofTccsta and Brahmaputra River al dlfferenl station (Dalia, Noonkaoa,
Bahadurabad ;talion) (Appendix B-1) was collected from lWFM library and from
BWDB for flood fre'lllen~y analysis. For flood frequency analysis year!} maximum
levels of water were used.
3.5 Flood Frequency Analysis
Frequency anal}sis is merel} a procedure for estimating the frequency of OCC\lrrenCe
or probability of occurrence of past or fllture events. To determine the water level of
Tarapur in respon,e to 2-,5-,10-20- and 50-year relurn peri"d, flood frequency
26
analysis was cond\lcted ba>ed on the water le~e1data of the year 1988 to 2004, To
,elect the appropriate distribution a total of five distributions were compared in this
study. These are Log Normal Distribution, 3-Pammcter Log Nonnal Distribution,
Pearson Type ]JJ Di,tribution, Lug Pearson Type 1Il Distribution and Gumble
Distribution. The comparison is based upon goodnes~-of-nt analysis. Based on 2, 5,
10,20 and 50- year return peri(>t!s,depth of flood were used for preparation of flood
inundation and flood depth maps. Detail calculation for estimating the flood flow
using the above five di,triblllions are given in Appendix B-ll.
3.5.1. Goodness-or-Fit Analysis
Good ness-of-fit study based upon probability piol corrc1ation coefficient (PPCC) is
useful for assessing whether a proposed distribution is consistent with the at-site data
sample (Stedinger et ai.. 1993). Another goodness-<lf~tit analysis based upon
probability piot (plotting position) is the root mean square deviation (RMSD) in fit.
Goodness-of-Iit analysis based upon PPCC is empioyed in this study.
The adequacy of a fitted distribution can be evaiuated by the PPCC which is
esscntialiy a measure of the linearity of probability plot. It gives the correlation
between the ordered observation, and the corresponding fitted quantiles determined
by a plotting positiol1. If Y denotes the observed ;-th now, II' the computed flow at
the i-th plotting position, Y the average value of observations and Wthe average
value of computed fl()"s, then lhe PPCC of the fitled tiistribution for n sample size at
a site is given by
" """p= L:(y, -f)C~-W)/[L:(Y,-F)'L:CW, -W)'],_, 1=' ,_,
The maximllm vaiue of PPCC can be equai to one, Among filted distributions to a
sample of data, the one \\iith the largest PPCC has the best fit to the data. For sc1e.;tion
of the be,t fitted method of flood frequency analysis for Tarapur point, the above
procedure was used (Appendix 8-1Il).
27
3.5.2. Iletcrmination of Fl"od Level
The Pearson Type III distribution was found to be the best distributiDn that describes
the flood frequenc} at Tararur (Appendix H.IV). Hence for determination of flood
levels regarding different return periods Pearson Type 1[I distrihution wa, ",ed.
3.6 De.-elorme"l of Flood Inundation and Flood Depth Maps
Flood Inundalion Or Flood Extent Map and Flood depth map was prepared fonn
Digital Elevatioll Model (DEM). A Radarsal image On 23,d July, 2004 was used for
Screen Digitizing in order to differentiate betwee" flooded and non-flooded areas.
The water level on the date of acquisition of satellite imagery was used to develop
flood inundation map using DEM so that, it can be compared with flClOdinundalion
map oblained rrom satellite imagery, All the above operations were done by using
ILWIS software (fLW1S 3.0 Academic, 2001),
3.7 Multispectral Image Classification
Multi-spectral imuge clu88ificution is u..ed to extrael thematic information from
satellite images in a semi automatic way. Different method8 for image c!a"ificalion
exist; some of them are based on the theory about probabilitie8_ In order to make the
elassifier work wilh themalic classes, some "knowledge" about the relationship
between cia,.e8 and feature vectors must be given. Theoretically. this could be done
from a database in which thc rciati(}mhip' bctween (thematic) classes and feature
vectors are stored.
At a certain image pixel in M bands simultaneously, M values are observed at the
same time. Using multi-spectral (SPOn images, where M=3, three reflection values
per pixel are found. For instances, (34, 25, 117) in one pixel, in anolher (34, 24, 119)
and in a third (I I, 77, 51), The values found for I pixel in several bands are called
feature vectors. It can be recognized that the first lVv'O sets of values are quite similar
and that the third is difterent from other two. The fir<t two probably belong to the
same (land cover) class and third one belongs to another elus8.
•
28
In this study Multi-spectral image classification through ILWIS 3.0 software was used
to classify kmd cover. ]n this regard for identif)'ing agricultural and non-agricultural
land a Landsat image on 24"' February, 2005 of/he study area was used.
3.8 SCrffR Digitizing
In ILWIS it i, furthennore possible to digItize a satellite image'" scanned pholograph
with the mouse pointer on computer screen, Screen digitizIng is the process of
creating or editing a segment (line) or point map while an existing raster map is
displayed as a background in a map window. The raster map can be for instance a
band of satellite Image, a color composite. a scanned map, Or a scanned photograph.
The background map or photo needs to have a georeference of type: Tie-points: for
satellite image; Direct Lin~ar: when the photograph is taken with a nonnal camera;
Orthophoto: when the aerial photogrdph i~laken with a professional photClgramm~lrie
camera and has fiducial marks. far the direct [in~ar orthopholO gcoreferences aDEM
should be available. An orthopholO is a rectified (north- oriented raster map with
square pixels) se~nned photogrammetric u~rial pholograph with corrections for tilt
and relief di'p!aeemcnt.
In this study ,cre~n JigitiLing was used to identifY flooded and non-flooded area of
the sUldy area thrClugh visual observation of lh~ Radarsat image in mder to CClmpure
flood inundation map based on DEM and Sutellit~ image. Map through screen
digitizing "ias also u8~d for giving an overview of flood damages Clfth~ study area.
3.9 Denlopment of Flood Haurd Vulnerability Fnnction for Agriculture:
Vulnerability function expresses the degree of 10"' On a scale bcl\veen 0 for no
damage and I for complete damage. Vllln~rability function for agriculture was
dev~loped from the field survey and existing reports alluiysi,. A depth-dam~gc
re!atiCln was foJ lowed for lhe development of vulnerability function.
"3.1(l Development of ~'ood lhu..prd VulnCl1lhlllf)'Mnf'!i of Agrkultu rC
Flood hl1lJJ.,d~ulrn:rnbil;ly maps for ngriculture was developed from LonduSCor
specified ngl'kullul'lll land use mop (ba~ on DEM), flood depth mop and
\'Ulncmbility function by lIsing II.WIS WftwlIlT.
J.1l Ovcrnll Vit"'"of the Mdhodot~.
Figure. 3.1: Overall vicw ofllle methodology
.I
CHAPn:R Io"OUR
OVERVIEW OF"THE STIfflY ARIo:A
4.1 Study AlUI
Tarnpur union of Sundarganj thana Under Gnibandha district in Bangladesh has been
selected a'I study an:tl. It islOCll1cd in !he northern part of!he coulltr). (Figure 4.1 and
Figure 4.2). Two IIUIjorriveTSbound Sund:uganj thana. IJrnhmaputrn (JnmWlll) in tJK,
C35tCmside and Tccst.II in the oonhem part of Bnngladc:$h. Tarnpur union is situlltcd
between 25~4'and 25~~north latitudes and hel\\=n 8900:29'and 89~31ClIst
longit\.ll:lc."l.
••-~.
Gaibandha District
Figure 4.1 : Study IlI'CIlIIUIp
•."3•
!~•••eg~
i]
I!~
• '"I •II ~
32
The sile of urban area of Sundarganj thana is 5,0 Sq. kIn and population is 9, 940,
Total area of Tarapur union i, about 30 Sq. km. Homestead area of Tarapur union lS
6.42 Sq. km. and population is 25,796 (BBS, 1994).
4.2 Climatic Condition
The climatic condition of the district i,mild with equable temperature, high humidity
and plenty of rainfall. The summer season commences in April and lasts up to early
June. The monsoon uS\lall} ,els in carly Jul} and continues till the end of September.
The winter SeaSOn sets in early November and Ia:;t, uplD the end of February. In the
months of January and February relatively severe cold is experienced in the districl
due to its proximity to the Himalayas and (Jaro hills. The average maximum
temperature recorded in the district during summer is 34.2oC while the average
minimum temperature recorded in the locality in winter is 10.1oC. During the hot
months humidity is high (about 90%) and remains high until December when it falls
sharply to reach a low in April. The maximum rainfall is generally observed during
the months of July and August. The annual rainfall in the locality is 2872 millimeters.
4.3 Soil Condition
The soil of the district is mainly divided into two groups i.e. Khiar (Brand tract) and
Poll (alluvium). The Barind tract ill the northern region of the district contains brown
and gray sHty clay loam soils of the Older Teesta Meander Floodplain. Along the
Teesta river valley, active and Young Teesb Meander Floodplain soil is also
prevailing in the region. In the eastern part of the district, soil along the Jamuna river
valley is active young floodplains with grey ,illy loam,
4.4 Topography and Hydrology
The topographical condition of lhe atea Is low gradient of the terrain and subjected to
floods almOSl every year. As the river courses change continuously, the terrain is not
much stable and is basically alluvium flood plain. This area is classified as shallow
depressions and valleys of moribund river channels formed by long morphological
hi,tory of changes in the river courses. Flash flood i.<,caused by the river Teesta and
the river flooding by the river Tee,ta and Brahmaputra. Excessive precipibtion in the
."
33
upstream, localized drainage congestions and insufficient capacity of hydraulic
structures an: the causes of flooding ill this area.
4.5 Major River System of the Study Area
Then: are a number of rivers and rivulet, in Gaibandha Zilla. The"" rivers are the little
Jamuna, the Gaigat, the Manas, the Teesta, the Ilengali, the Karutoya, the Akhira and
Brahmaputra flows through eastern border of the study area, All these rivers arc non-
tidal and are not navigable during all season" (BBS, 1996). Among these rivers Tee;ta
and Brahmaputra play significant role in maintaining the flood level and
characteristics of flood of the study area,
4.5.1 Teesta River
The River Tcesta plays a significant role in maintaining the hydrologic characteristics
of North West region of Bangladesh; carving out verdant Himalayan temperate and
tropical river valleys. 'I he emerald c(}l(}redriver then forms the border between
Sikkim and West Bengal beJore joining the Brahmaputra as a tributary in Bangladesh.
The river originates in the Cho Lbamu Lake at an elevation or 5,330 m (17.500 fcet)
aoove MSL i~ the Himalayas.
4.5.2 Brahmaputra River
The Brahmaputra is a major international river covering a drainage area of 580 000
sq. km. of which 50 percent is in China, 34 percent in India, 8 percent in Bhutan and 8
percent in Bangladesh (G(}swami, 1998), The gradient of the river in tlle garge section
ranges from ab(}ut 4,3 m to 16.8 mlkm, On entering India, the Tsangpo, now called
Dihang, traverses 226 km. ora mounlainous cOurSebefore it emerges onlOthe A,>arn
plain ncar Pasighat (elevation 155 m). Near Kobo, 52 km. soulh ofPasighal, the Lobit
River joins the Dihang, and lhe combined flow, called tbe Brahmapulra, move,
west"ard through Assam for no km. until near Dhubri,
The Brahmaputra originates in the grcat glacier mass of the Kailas range In southern
Tibet at an elevation of 5300 m. In ribet, where it is called T,al1gpo, the Brahmaputra
flows eastward for 1100 km, At the extreme eastern end of its course in Tibet, tile
34
Tsangpo enters a deep narrOW gorge (elevation 3500 m) and continues s()uthwards
across the Himalayan range where it swer,cs to the south and enlers Bangladesh. The
Brahmaputra has a gradlent of 0.09 to 0.17 mlkm near Dibrugarh at the head of the
valley and is further reduced 10about 10 cmJkm ncar Pandu (Goswami, 1985). The
slope near Bahadurabad reduces to 7.5 em/km and the average width of the river
within Bangladesh is about 11 km with a depth of about 5 meIer,. The river carries
>ediment of about 590 million tons per year (EG1S 1997 and \VARPO, 2002).
The Brahmaputra is the main source of fresh water in the dry sea,on and also causeS
flooding during the monsoon. The river starts rising in March/April due to snowmelt
in the Himalayas and reaches a peak in June. It rises again and reaches the annual
peak in lale July unlil mid August The river nel-\lIork system of the study area is
shown in Figure 4.3.
BuriTeesta
TeeSla Rill
.; i rahmaputra River,
ld Brahmaputra River
Figure 4.3: The river net-work system of the study area.
"Although floods and fe'o'o' otht'r diSMlers occur frequtnlty in !he locality but flood is
mon: dCVllStD.lingand occurs vny freqUC1ltl)', Both IlllSh flood lind monsoon flood
occur in the locality. Flash flood oecur normally in lhe mOllth May-June when: liS
monsoon flood OCCUrin the month July-September. Although Teem is responsible
for both types of floods bUI the nvcr Bmhmllputnl also has II signifICant influence in
flood chm!claiSlies of the locality. Excessive pm:ipimlion. 10000Jiud drninagc
congestions Dod insuffICient ClIpacity ofhydmulic structures Iln: the causes of flooding
in this 1I.m!. Flood ~ hasa:l on NOAA Satellite imagery for all over Bangladesh
lire given in the I'igun: 4.4.
=----ClI--ClI--.-c:I--ClI-__ ....._=_ ...• .,.."
figure 4.4: Types offlood and extent map
FLOOOTYPES'-- .•_- ':-.....
,. ,"
36
4.7 Agriculture
The economy of the study area Tarapur is mainly agrobascd. About 62.97% holdings
are fanns that produce varieties of crops namely local and ]-IYY rice, wheat, jule,
mustard, potato, sugarcane, different kinds of vegetables and others (BBS, 1996).
High soil fcrtilil}, climatic condition and ,ilt deposition make the study area
agricultural dependent.
Gross cropped area uflhe study area is 7974 aCre and net temporary croppeJ area is
3972 acre, single crop 12.29%, double crop 74.29% and triple crop land 13.42% and
cultivable land under irrigation is 70%. Among (he pea~allts38% are landless, 29%
small, 23% intermediate and 10% rich and cultivable land per head i, 0.10 hectare
(BBS, 1996). Area under difTerent landuse practices and croP" in the study area arc
shown in Tables 4.1 and 4.2.
Table 4.1: Homestead and cropped area ofSundarganj thana (Area in acres)
Classifiea All Operated Homestead Temporary Permanent ColS!tinn holdings lIrea area crop anll crop area ColJ
in %I 2 3 4 5 6 7
AllHoldings 77376 67649 5258 55779 1809 82.5Non-Farm 28911 2039 1112 25 52 1.2Honschol
"TotalFarm 48465 65610 4146 55754 1757 85.0Holdings
Source: BBS (1996)
37
Table 4.2: Area under different crops of Tara pur union (Area in acres)
Name of No< Gross H" Local Wheat Jute Vege-Villages Temporary Cropped Arnan Paddy tables
C" rt PaddyChachiua Mir .. 840 1738 386 955 97 146 143Gbagoa 779 1660 315 918 117 167 134Khurda 230 452 55 191 47 91 65Latshala 414 778 18 299 113 192 154Nachanigoa 238 597 16" 219 53 86 58Nohal; ClInchi .. 377 700 lJ3 347 67 73 72Nizam Khan 464 984 162 316 109 114 164Tatapur 630 1065 36 534 137 228 122Source: BBS (1996)
Table 4.3 represents the basic characteristics of agricultural Inoduse practices of
Sundarganj Thana.
All hotdin5258100
5825886.118095578095,74126470,8
ItemsHomestead area% offann holdinNet cu!tinted area% of" crated areaArea under crmancnl en'Net tern Orllr ero ed area% of net cultivation areaNet irri ted area% of cultivated areaSource: BBS (1996)
Table 4.3: Basic characteristics of agricultomllanduse practice, ofSlIndarganj thana(Area in acres)
Total farm holdin s4146 '78.855817088.717585575595.84124970.9
4.8 General Information of tbe Study Area
Among the total population of the study area males are 51.84% and female 48.16%;
population d"",ity per ;q km is 1988 (BBS, 2001). Administration of Sundarganj
thana was establi,hed ill 1875. It became an Upa-Zillil 1983. Average literacy rate Is
24.1%; male 31.9% and female 16.5'%. Agricullure is [he milin occupations for
l8
48.57% people, agricuUurallaborer 29,64%, wage laborer 2.06%, commerce 6.63%,
service 2.59%, fishing 1.25% and others 9.26% (RBS, 200 I).
Communication facililic\ include roads: 46 km. pLLcca roads semi pucca 3 km. and
mud road 650 kIn; railways 8 kill; waterways 8 nautical mile.
CHAPTER FIVE
DEVELOPMENT OF VULNERABILITY FUNCTION
5.1 Introduction
Vulnerability functions relate hazard intensity to vulnerability (Maathuis and Schier!;
2004). Vulnerabilily function detennines lhe status of vulnerabi lily or ri,k of an aTea
which normally uses to prepare vulnerability map. Vulnerability function depends on
the status of damages corresponding to hazard magnitude like flood depth, dumtion
and number or amount of elements al risk.
For the development of vulnerability funclion of agriculture for Tarapur union 10
a%ess the flood hazard vulnerability the relalion between flood depth and agricultural
damages has been followed in this study. The function has been developed on the
basis of damage to agriculture with the flood level. For establishing the relationship
between flood depth a~d agricultural damage 3J1exte~sive ,ocial field survey has been
carried out in the stlldy area. Based on this survey data and existing secondary data a
depth-damage relation,hip between flood level and agriculture has been established in
this study.
5.2 Vulnerability Function for Agriculture of Tara pur Union
From the above study, it is found that development of vulnerability function for
agriculture due to flood depends 1mthe appropriate damage estimation corresponding
to rise of water level. So previou, data of damages to agriculture due to flood is
necessary for development vlllnerability funclion of an area. Due to insufficient data
of !lood damages to agriculture of the sllldy area, differcnt procedures wcre applied in
this study to estimate previous t100d damage. Following steps were followed for
estimatlng flood damage to agriculture with respect to water level.
First an identification of cropping area, types and paUern was carried out through
existing secondary data analysis and multispectral classification of Land,at image.
40
Then classification of land of the study area based un elevation was carried OUI.By
comparing Landsat image with land water c1a>sified map of Radar image (through
screen digitizing) and on (he ba.,is of existing data of flood affeded area, an overall
view of flood affected agricultural land waS found. Finally all of these results were
compared "ilh the collected field survey data,
5.2.1 Identificati(}n of Cropping Area and Pattern
The vulnerability function to agricultural product, due to flood is closely related to
different ~ariables like the cropping pattern, locatlem of agricultural land, flood depth,
land class lind duration of standing "ater. Identification of agricultural land is very
essential tor assessing the flood affected cropping area and for measuring its
su,ceptibility to ha£ard. In few cases limited extent of floods become neee,;sary or
beneficial for ag.ricuUure. In general flood depth and damage to agriculture is
considered as directly proportional. For identification of cropping area existing
secondary dala were used. 'fable 5.1 shows thai almo,1 82% of Ihe study area is now
using a, agricuUuralland,
Table 5.1: L3nd c1a,sification of Sundarganj thana
Source: Sundarganj Upa-Zilla Agricultural Dcpartment
SI. No. Types of land Area (hectarc)
1 A riculturalland CUJTentlyin use) 27,052.00
2 Tem orarv fallow land 1950.00
3 Permanent f3110w land 5370,00
4 Homestead urea 3831
5 Road, Infrastructure and others 3630.14
6 Total a iculturalland 34316
7 Total area 41833,14, % of agricu lturalland ,,[ total are3 82
For idcntification of 3gricultllral land of tile study area a Landsat image on 24u,
February, 2005 of the study area was used. Figure 5.1 shows vegetation or <lgricultural
land of the study area ba,ed on muUispectral elassific3tion "f that image.
41
Figure 5.1: AgriCOJltuTIlI I.nd of the study llfell (Multispedflll clusifical;on of L.eoosatimage)
Visual interpretation orthe classified image ShowSlhat about two third portion of the
study IlTell is covered "';111green vegetation lind this portion is very much closely
lmOCiatedwith the bank of the Tcesta River. Aceording to BBS (1996), the
agricultunll land ortlle study lll"MSundarganj thana is about 8~A.(Table: 5.1) which
.Imost supports the visual interpretation of the image regarding the \olal cropping
42
Vulnerability to agriculture largely depends On type, of crops and cultivation period.
High water tolerance specie< is less ,usceptible compared 10 low water tolerance
species. For this reason crop calendar is very important for determination agricultural
vulnerability due to flood. Table 5,2 shows the crop calendar "rthe study area. More
infonn3tion of cropping area and pattern of the study area and a c(}rnparison of
land use practices ofSundarganj thana are given in Appendix-C.
Table 5,2: Crop calendar of the study area
Source: BRR!. Bangladesh
N.me "fCrop
T, Amon -f{h.rlfllurI! Hort>- RQM
Wk•• i_Rubl
Mu,teTed _Rub!
Pulse-RaM
SugllFcane _R"b;
UInl", V'/t"lubl.
Rob!
HYVT. Au>-
KharlfJ
HYVT. Au>-
Kllarifl
July .1""
5.2.2 Classification orLand based (IIInevatioD
Slmdarganj Upa-Zilla Agricultural Department prepared a land classilicalion
regarding l100d inundation (Table 5.3), TIlis table represents the land cia"ification for
the whole Sundarganj thana, Tarapur union is the worst flood aft~ted area C<Jrnpared
to other, fifteen uniom of Sundarganj thana. In field survey, maximum portion of
very low and low lands were ob,ervcd in thrcc unions along with Tarapur.
43
Table 5,3: Land classification regarding flood inundation of Sundarganj thana
Sl. No. Types of Land Area (Heetare)
I High l.and (Not inundated in normal monsoon flood) 8485.00
2 Medium High Land (Inundated maximum 90 em in 14285,()()
normal monsoon flood)
3 Medium lA,w Land (lnllndated 90-180 em in normal 7978,00
mDnsoon flood)
4 Low I ,and (Inundated 180 275 om i, normal 3535.00
monsoon flood)
5 Very Low Land (Inundated more than 275 em in 932,00
normal monsoon tlood)
6 Total 35,215.00
Source: Sundarganj Ura-Zilla Agri~ullural Department (2004)
or about 1350 hectares of currently uses Kharif-ll crops area of Tumpur union about
25% area becomes inundated with Ilood water over 250 em depth due to monsOon
flood (fahle 5.3 ). Moreover ba,cd on OEM of Tarapur the land classification i, given
in Table 5.4.
Table 5.4: Inundated agriculturallUlld of Tarapur union based on DEM
81.No. Rise ofwater Overall depth of water level % of area
level inundation, 100 em 23.91(m) 202 200 em 24.91(m) 65
4 300cm 25.91(m) 90
5.2.3 Identification of Flooded and Non_tlooded Agricultural Land
The >ludy area Tarapur union experienced !he worst damage due to monsoon flood in
July, 2004, A Radarsat image on 23" July, 2004 or (hc study arca was also used to
identify flooded and nOll-floodedarea through sCreendigilizing (Figure 5.2).
44
By merging both the images (agricuUuml land classified Landsat image and noodcd
and non-nooded area c1a,sifieJ RaciaL,a! image) finod aITecled agricultural land can
be identified. From visual interpretation of both the images. it was found that almost
all the agricultural land was inundated.
1:1 Flooded Area 78 % iii l\'on-Flooded Area 22%
Figure 5.2: Flooded and non-Flooded Area. (Screen digitizing "[Radar,,,! image ofthe study area)
45
Data or infonnation regarding damages due to the mo"'o"" n"od was nOI sufficient
in the study area. Only an estimation of overall natural disaster affected agricultural
land of lhe year 2004-05 was carried out by Urn-zilla Agricultural DepartmenL The
ava; lable damage data for the year 2004-2005 to agriculture of the ,Iudy area is given
in table 5.5.
Table 5.5: Flood and olher disasters affected cropping area ofSundarganj thana(Hectare)
SI. No. Types of CroJlS Area Cultivated Area % of affected
Affected area
1 Kharif-JI 18800 15228 81
2 Jute 990 3'2 "3 Oil Seeds 1010 382 384 Wheal 800 75 9
5 Vegetables 280 99 36Source: SundarganJ Upa-btla Agr",ult\lral Department (2004 -2005)
According to the Upn-Zilla Agricullural Deparlment of Sundarganj me overall flood
affected agricultural land is 81% (Table 5.5) which is very close to the ovemll flood
affected area (78%) found in land water classificd map of Radarsat image. So based
on this data (both image and Agricultural Department) of flood affected area due to
monsoon flood and level of water (25.5Im PWD) on the datc of Radarsat image
acquisition 23'0 July, 2004 a depth-damage relation for agriculture of ,tudy area can
be established. For the development of depth-damage function of agriculture for
Tarapur union, all ofthese data are widely analyzed with field survey data to establish
the relationship more fruitfully and mOre logically.
5.3 Development of Vulnerability Function for Agriculture through Depth_
Damage Relation
Although the study area is located in a floodplain locality but the terrain is not so
mtlch flat. OtiC to this depth of flood Ivater level varies in different location
corresponding to a fixed level ri,e of water level. ]n this regard for the development of
46
a depth damage function, the ,Iudy area has been divided into four categories based
on the elevation. This land c1m;s;ficat;cm ha, be"" done on the basis of extensive field
ob""valions and existing data analysis (Table 5.4) of Tarapur union of Sundarganj
Upa-Zilla Agricultural Department.
The field survey was conducted on [he basis of a questionnaire through personal
inlerview with (he villagers of Tarapur union. About 4R villager, of differenl villages
of the study area, and a group of three otIlcials of Upa-Zilla Agricultural Departmen!
headed by Upa-Zilla Agricultural Officer werC interviewed in the survey.
From lhe field survey and existmg data analysis it is fQund that in monsoon (July 10
September), (lne meter rise of waler levellcads to 100% damage to all crops in the
very low lying river associated land. Where as aoout 2 meter rise ofwaler level causes
100 % damage to all crop, in low land crops and a 2.5 meter and 3 meter rise of water
level causes 100 % damages in all medium and high land crops of the study area,
respectively.
From the above data analysis it could be summarized that in monsoon (July to
September) about 30% of the total agricultural crops of the study area is damaged due
to I (one) meter or about 3.5 feet rise of mon.WOnflood water where as about 50%
erops is damaged if the level of water rises by 2 meter or about 7 feet. Again another
0.5 meter rise ofwaler level leads 10over all 80% damage of crops and finally overall
3 meter ri,e ofwaler level leads 10 100%damage of all kinds of crops (lfthe locality.
Table 5.6 shows overall view or depth-damage relation of agriculture for Tarapur.
I:lased on this table the function of agricultural damage due to monsoon flood for
Tarapur union has been developed.
47
Table 5.6, Overall view of depth-damage relation of agriculture for Tarapur unicm
SI.No. Rise offload water level (m) % of damage to agriculture
1 1 30
2 2 503 25 &II
4 J 100
Vulnerability function of agriculture for Tarapur union based on field survey and
existing data analysis is given in figure 5.3.
1
0.8
~ 0.6,i 0.4~
0.2
00 1 2 2.5 3 4
Water Oepth (Meter)
Figure 5.3, Vu[nerab! lity function for agriculture of Tara pur lInion
Vulnerability function shows more or less a linearly increase trel1d~up to three meter
depth and then it bc<:omcs horizontal as three meter rise of water level brings 100% or
complete damage to crop,. The increa,e of vulnerability from 0 to J meter
depth and 1 to 2 meter depth was sharper than that from I to 2 meter depth.
Thi, vulnerability function ha, been used to generate vulnerability maps from flood
inundations and nood depth maps for different retum period,_
CHAPTER SIX
ANALYSIS AND DISCUSSION
6.1 Introduction
Flood occurs in dinerent magnitudes in cach ycar in Bangladesh. The recent floGelof
2004 was abnormal in character and magnitude. Huge damages to crops, property and
infrastructurc and losscs of Iifc had occurred. Thc damagcs to agricultural products
are closely related to the cropping pattern and flood depth. The choice of crops to be
grown in an area is determined to a large degree by flooding characteristics. The
dcpth, timing. duration and ratc of risc and fali of waler levcis arc important factors
influencing whcn and what crops Can be grown. Judging from thc rciationship among
crops, iand class amI flood depth, flooding:doe, not always result in crop damagc and
there is a phenomenon that the floods and crops have been co-existing. Here it should
be noted that floods do not oniy resuit in damage but can also benellt crops and
farmers through all appropriate study. Hellce de,eloprnellt of flood hazard
vulnerability map is essential, and helps devise an appropriate management policy_
For better management of flood and for reducing the intensity of !lood damages and
for ",iectioll of location for agriculturai and other activities, flood vulnerability map
regarding different return periods eouid be important toois. In this section tlood
hazard vulnerability maps have been developed through flood inundation map', fl,,,,d
depth maps and vuinerabiiity function by using gcoinformatics to assess flood hazard
vulnerability for flood different magnitudes.
6.2 Determination of Water Level at Tnrapur
In order to develop t100d vulnerabiiity map due to monsoon !lood of the study area,
water level data for the monsoon period (July to September) of the river Teesta at
Tarapur is needed. As no water gauge station of the river Teesta at Tarapur i~
available, the water levei of Teesta at Tampur has been determined through
interpoiation based on distance.
49
To detenninc the water level at Tarapur, the river, Buri-Teesta (00 to 35.00 kIn),
Naotara. (00 to 10.00 km), Teesta (13.50 to 121,00 km), Dharala (21.5 to 48.00 km),
Ghagot (132.27 to 138.00 kIn), Jamuna (8.00 to 84.50 km) and Brahmaputra (DO to
31.00) have been considered. Actually Tarapur area lies on the bank of the river
Teesta at the chainage of 103.19 kIn from Dalia. The Tcesta faces influence on its
stage/inflow during flood period by the other rivers as considered in thi, network.
Figure 6.1 shows the basic network mentioning the name of rivers.
-- - .,--•.
• ••• _••• __H"_
•.
fjJl.U81i".,..,.,.. -----,.
.• , ...••• , .. : •••• , ••••••.•••••. i••.•. .
l.2:wi-Tee:- -- ".."" ..,....
: Dalia; , ,.._. .. ; ; , ., w'IS .,
- - -- -,.•••• « •• ,>,-, , ••• ",., •••••. , 0<.
;ree~ll'I__ RweJ;._
. . .•.,_••.0""''' ,_."" ..._".,. ..",.".....;--:------ ...... ,- ...-
•__u', ••• j , ..1 ):r.~pUI
-------------j.
•............ , "
•.
'Chiliii8i'i, .. a", t "'
Figure 6.1 : River network 'y,lcm of the area
Based on the waler level of Dalia and Chilrnari, the water level or Tampur has been
determined through interpolation as Tarapur is localed in between these two slations.
The interpolated value then compared with othen; ~ery closely associated water gauge
50
Slalions. Figure 6.2 show> maxImum yearly (July to September) water level of
different stations.
601~9 O-G0 OO-GO 90 0 9 050 -
I---e-- Dalia WL'0,~
,40 - 1-- Noonkhaoa WL
">"" --4- "farapur WL" 30~ ~;;;~
-- Chibnari WL
20 ,"~ BahadlJl1lbad WL
10 ,---,-- ",--,.-,--,-, ---, -,--,1988 1992 1996 2000 2004
Year
Figure 6.2: Maximum water level of Tara pur through interpolation (July to Sept.)
6.3 Nand Frequen<-'Y Aual)" .•;s
For preparing flood extent, depth and vulnerability map of the study arca, water level
are considered based on differenl rdum peTiOli>.To deleno;"e the water level of
Tarapur in response to 2-.5-,1O-,20-and 50-year return periods flood, frequency
analysis "as condllcted based eln the watcr level data oflhe year 1988 to 2004. Based
On the mo.'! appropriate Pearson Type JII distribution, water levels for different reUlm
periods are given in the Table 6.1.
Table 6.1. Water levels from flood frequency analysis
SI.No. ReturnPeriod~1 22 53 104 205 50
6.4 Digitlll ElevlltioD Model
Water Level Meter25.3125.7125.9126.0726.24
51
Digital Elevation Model (OEM) is one of the essential data fur flood study of an area
using ILWIS This study is very much dependent on DEM of the study area. The
DEM of Tarapur is given in the Figure 6.3, The resolution of the DEM is 50 m
l¥QkUlo ZUl_21.3l7
E3 21.387.22.17422.11Jl_ 22.lI6
1t;1122.96-Zl747=Zl747-XSM
"'" -"'"=25.32l _3'i.1Ol21.1111- 2I.IIM
• 3II1II- 21.l1li
0'""'"
Figure 6.3: Digital Elevation Model (DEM) ofTarapuf area
52
The DEM of Tarapur, prepared by BWDB, he' been used for this study for spatial
analysis to assess the tlood extent and noad depth map for five return periods. The
OEM is also used for identification of agricultural land uw, homestead area and for
preparation of vulnerability map for agri~ulture, The DEM can also be used for
detcnnination of homestead damages, planning of llood mitigation and economical
analysis for teasibility of the !load mitigation project.
6.5 Flood Inundation Map
Flood inundation or extent maps have been prepared based on the Tarapu. DEM
using the flood level 25.3lm, 15.71m. 2S,9\m, 26.07m and 26,24m cOITe'p{mdingto
2-, 5-, 10-,20- and 50-year return periods. In this pr(lCcdure, assuming onc breach
location at the outlet of the existing river system a map named ',tart' has fir,tly b(l8n
prepared. The flood inundation maps have been compukd using neighbourhood
functions where pixels assigned as llooded are at or below the flood levels and
connected to the breach location, to prevent area' that are lower than the ,pecified
flood levels but arc not connected to !he aITectedarea from laking in to consideration,
The output maps arc flood inundation map' indicating flooded and non-flooded areas
for each particular flood water level. flood inundation maps have been calculated by
iteration. Displaying the breach location map: start. It is found the Value '?' stands for
background (undefined); and Value '1' stands for breach location,
In the output flood inundation or flood extenl map for 2_year return \vater level has
been considered 25.31 meter above the mean sea level. The output flood inundation
map with a value ranging from 0 to I and precision of 1, is !he flood inundation map
defining the 2-year return period flood inundation map. The process in words: lfthe
altitude in the used Digital Elevation Model is more than 25.31 meter, then return the
pixel values of raster map Start (which are undefined, as is indicated by lhe ?)
otherwise, assign the maximum value of the neighbouring pi;>,e1sfound in raster map
Start (which is the value of 1). In the fir;l iteration there ;s only one pixel that ha;
value 1 (the starting pixel). In every iteralion, the neighbouring pixel> lhat satisfy the
conditions (altitude < 25,3] meter) will get the same value a, that starting pixel. This
•••
53
will continue until the next neighbouring pixels have an altitude of more than 25.31
meter. If there is nO connecting pixels <25.31 meter exists. Ihen the process stop' and
thc results are displayed. The flood inundation map for 2-year return period is shown
in Figure 6.4
'., -- , ',"-,-- '<,
fl!il,F1ooded.l\rea 87%
OnNon-F1ooded'Are.- 13%- . "., ,
,--eL1:;' - .,
Figure 6.4: ['IClOdinundation map for 2-year return period
Following the same procedure flood extent map for 5-, 10-, 20- and 50- }ear return
periods corresponding to 25.71 m, 25,91m, 26.07m. and 26,24m depth of water levels
arC shown in Figures 6.5, 6.6, 6,7 and 6.8, rc>pcctively.
:OGNoo-F1oadedArea 7.25%
Figure 6.5: Flood inundation map of5-year relurn period
54
,-+".,," ' .
... x._.' =__=_~o 2:S 11m
EJI Fl~od.dAr'. :~4.,
00 Non-FloodedArea
Figure 6,6: Flood inundalion map for 1O-year ret\lrn period
•'-t"', .-- -- ,--
2.5'km
55
•-+'.~~=-~=--0, 2,ikm
Figure 6.7: Flood inundation map for 20-year return period
'o~Non-F1ovdedArea
•-+""'.0.=__ = __o ,2.5m'
figure 6,8: I'lood inundation map,of50-year return period
6.6 Coinparisun of Flood Inundation Map bns~d on nEM and Satellite Imag~
Flood inundation maps based on OEM and watcr level 255m On the date of s.alellite
image acquisition On 23'" July, 2004 and land-wate' classified map oftbe sludy area
based on SCreendigitizing of Radarsat Image on the dale 23'" July, 2004 are shown in
the Figures 6.9 and 6.10, respectively.
IliIl Fib;6d"'d,jlj-e~'.," '88 %. .." ... ,..", ."., '
Figure 6.9: Flood inundation map bascd 00 OEM and water level data on lhe date ofsateliite image acquisition (23mJuly, 2004).
57
I!lII Flooded Area 78% (approx.) Em Non,Flooded Area 22% (approx.)
Figure 6.10: Land water classified map based on >ereen digitizing orRadar imageon 23'" July, 2004.,
The waler level al 'J arapur on 23" July, 2004 was found to be 25.5 meter on the ba,is
of interpolation of water level data of the boundary rivers of the study area. The
interpolated water level close!} represents original data. This is why a little bit of
dissimilarity bel,,~~n nood inundation map based on DEM (88% lloodcd) and land-
water classified map based on screen digitizing of Radar Image on 23rd July, 2004
(78% flooded) was found (Table 6.2). The dissimilarity may also be due to quality of
OEM. But from the satellite image it is found that the maximum land area was
inundated all 23"' July, 2004, Flood inundation map based On OEM represent the
same.
58
Table 6,2: Overall view of the flooded and non-flooded area
$1. No. Types of inundation Depth % of area % of area- non
m.p of water flooded flooded
I 2-year return period 25,31 87 13
2 lO-year return period 25.91 94 6
3 20 year return period 26,07 95 5
4 50-year return period 26,24 % 4
5 Satellite image - 78 22
6.7 Flood Level Map
As the undefined pixels in the calculated flood extent maps are not flooded the value
of 0 will be defined instead of using undefined pixel values. To eliminate the
lllldefined pixels preparation offload level map is needed. Flood level maps based on
different return periods are given below:
-, ,,_ '---.:;,;: :c-.,<-,' -.'<"-; '~_'
, .C
II Undefinedarea II DefinedareaFigure 6.11: Flood level map for 2-yea:r return period
59
II Undefinedarea II DefinedareaFigure 6.12: Flood level map for 5-year return period
•• Undefined area iii Defined area
Figure 6 13: Flood level map for lO-ycar return period
60
a Undefined area • Defined are
Figure 6, 14: Flood level map for 20-year return period
,
:~(k;~'0~?i:;f.'-~,,,,-,'
a Undefined area • Defined area
Figure 6,15. Flood level map for 50- year return period
61
6.8 Flood Depth Map
Flood depth map represents the depth of flood water in different location due to flood.
To calculate the flood depth map, the elevation (from the DEM) of each flooded raster
cell has been subtracted from the corresponding flood level from the different
scenarios. The flood depth maps have been prepared using map calculation command
from operation Jist ofILWIS software
""'"0=_-=-o 2.5"'"-16;:~-t~!z.~-l!!!!3.5-I!!!!!I ~.7
Flood Depth (Meter)
Figure 6 .16: Flood depth map for 2-year return period
N
+
.1°."• 1'-1 1.3"I-I 2.6"~ 3.8- 5.1
Depth (m)
,
+.~~.=~-=~-o 2,5 Ilm
62
Figure 6.l7: Flood depth map for 5-year return period
1iiI""1I:]1.3
I"...", 5.3
Depth (m)
,
+......,.~=-~=-.o 2,5km
Figure 6.18: Flood depth map for IO-year return period
63
~O'=~_=~_o 2.5 km
Depth (m)
•+
Depth (m)
.1"-1-11,4- 1~12,a- l1liU- 5,5
•+Figure 6.19: Hood depth map for 20-year return period
Figure 6.20: Flood depth map for 50-year return period
Table 6.3 represents overnll view of flood depth observes in flood depth maps
regarding different return periods.
64
Table 6.3: Overall view ortlle variation ofl100d depth
51. No. TyJ1f'5ornood depth mJlp Muimurn depth ofwftlfr
I 2.y=- mum period 4.7 (m)
2 5-ycar return period 5.1 (m), lo-year mum period 5.3 (m)
4 2G-year return period 5.S (m)
5 SO-year return period 5,6 (m)
6.9 IdC1l1ifiCilion of Agrieullunl ynd and ltomntC1ld Arfll .nd lnfraJtnl(1urt'
From the water level datllanalysis of the major boundary rivers of Tlll1lpur union of
about 17 years for the period July to September, it is found that on average the
minimum water level at Tarnpur location during this period is 22.91 meter. This water
level indicates dllli al the monsoon period all the crops are cultivated in Tarnpur union
above this land. So the agricultural land, homestead area and infl1lstrueture of the
study area Tarapur umon will be >22.91 meter land as shown in Figure 6,21.
'nn..~.,. I • 25.3
".•tt.'
EJe~lIlion (m), •,'. '" •••• ,,• •,••
,
+.~-:II=-':~-o 2.5"'"
Figure 6.21: Map containing agrirol!unllland homestead and infrastructure
65
As the agricultural land, homestead and infrastructure of the study area fuund >22.9\
meter land area, the land area of Tarapur unioO>22,91 meter gives an indication of
unspecified [anduse map of the study area based on DEM, Such types of landuse
maps have similarity with the land use map prepared based on multispectral
classificatioll of Landsat image on the date 24th February, 2005 of the study area
Tarapur union,
From field survey and existing data analysis, it is fOlllld that homestead area of
Tarapur union is located on an elevation considering flood depths corresponding to
IO-to 20-year return periods_ The homestead area of Tarapur union is found 10 be
greater than 25.91 meter (PWD), Figure 6.22 shows homestead area and infrastructure
only. Maps containing homestead area and infrastructure have been prepared using
map calculation command from operation list.
Elevation (m),
+.~.:.=-=-
II) '2:.5 I;m
I
I,
•••
.~...•on ..t'"• •,. ,I ' ., . ~ ,., ~',
Figure 6.22: Homestead area and infrastructure only
66
From the above data analysis, only the agricultUnllland orlhe study are!! is found in
between the elCVlllion21.9 10 25.91 meter. As this study imends to ddaminc the
vulnerability for agriculture, the flood depth below or equal to 25.9\ meter Of below
or equal In 10-year return period will only ddennine the vulnet1lbility for agriculrure,
flood depth greater than 25.91 meter will determine the vulnefllbility for .IIgrirotruTe
and other KCtor like homestead and infT'll.$lrucrure.. Figure 6.2J shows the map
contlining agricultural land only. Map containing agricultural land only hnve hem
prepared using map calculation command from operation list.
~
~.,~.•. ~'!.4"..".
•+ -=-,==---"=--o 2.51cm
Elevation (m)
Figure 6.23: Map containing agricultul1Illand only
•
67
6.10 Identification orVulnerable Agricultural LllJld Regarding Different ReturnPeriods
In order to develop flood vulnerability maps for different return period vulnerable
agricultural land regarding different flood depths for different return periods will have
to be identified Vulnerability function will then act upon these identified vulnerable
lands. Maps containing vulnerable agricultural land regarding different return periods
have been prepared using map calculation command ITom operation list.
I,,'
-' O,b;1IOi1.2,,.
".Depth (m)
,
+ ~~,.=~-~=_.o 2.$ flm
Figure 6,24: Map containing agricultural land for 2-year return period
8'.', 0,7
S"- 2,1,.Depth(m)
•+.~.,.=~-=~•..o 2.&km
Figure 6.25: Map containing agriculluml land for 5-year return period
80.'0.'
8.."'0Depth (m)
,
+0"
Figure 6.26: Map containing agricultural land for lO-year return period
2.6km
69
S'.,,.,U
02,51:13,2Depth (m)
•+• ••••••
Figure 6.27: Map containing agricul1l1f3lland for 2(}-year return period
-~O,4-1:]1,1
8"- 2,6..1:13.3Depth (m)
•+.__.-------o U.km
Figure 6.28: Map containing agricultural land for 50-year return period
70
6.11 DrvdopmtDI of VulntrabiJity M.lp (or Agnralturt' ll~iDgVulntrnbilll)'
FuncHon
Vulntfllbility maps repr~nl the response or susceptibility of an area to Tl!lUraJ
ha1.ard, or disaSler like flood regarding 0 for no damage or I for complete damage.
VulllCl'1lbilitymal" indicate the SlatuS of an area so that priority list for mitiglltion
mC3~ ClIObe laken. Vulnerability maps are also important for taking the decision
of what types of mitiglllion measures in what area will be taken or will be suitable.
VuloeTllbility maps are essential for delermination of specific risk, elements at risk,
total riiik ami for estimation or total dama~ of an ana due to flood or any types of
natUTllIdi5aSler. In short vulnerability map gives an ovemll overview of an area
regarding a natural di5llSlcr like flood, Flood vulnerability maps for agriculture of the
Sludy Mea regarding different flood rdurn periods arc given below:
0.000.17
. I '0,3"0.510.68
Vulnerobility
,
+
Figull: 6.29: Vulnerability ITIllpof agricultural land for 2-ycar return period
2.5 km
71
Vullln1lbility
N
+ """,.,"=""--==,,,,-o 2.5 Ionn
Figure 6.30: Vulnerability map offtgrieu1turalland for 5-ycar rcturn period
•
72
~
'.,. ~ 0.26. 0,50
'.n,.~Vulnernbilily
•+ '"'""'=""-==--"o 2.5 ••••
Figure 6.31: Vulnerability lI1lI.pofagriculturallllnd (or 100year return period
13
., ...-d.lI: .i:t ,•••••••• • I, '
~ ~ ' ~ •"•
•
•
6'.M
- I 0.32- I O.SS
•.n'.00
VulnC'l'lbilil)'
,
+ """"=""--==_ ••o ~.s'""
Figure 6.32: Vulnerability map ofa.griallturalllnd for 20-)'car rrtum period
74
e ••••
••. 8',
c1I " ,
no,"o,~o,~o,m',00
Vulnerability
N+ .._...~=--==--o 205km
Figure 6.33: Vulneroibility map of agricultural land for 50-year return period
• 75
From the flood vulnerability map it is found that the magnitude of vulnerability
increases corresponding to different (higher) return period and about three (3) metcr
rise of water level brings complete damages to crops of the ,Iudy area. Due to unCven
distribution of topography equal rise of water level does not lead to equal damages to
all the area.
Complete or 100% damages observe in ooth flood vulnerability map" of 20-and 50-
year return periods. The extent of higher magnitude of vulnerability is high in flood
vulnerability map of 50-year return period compare (0 20-year rCUlm period. And in
all the maps higher vulnerability was observed in the area close to river where as low
vulnerability was observed in the high land filr away Hum the river.
.''_a
CHAPTER SEVEN
CO"lCLUSIONS AND RECOMMENDATIONS
7.1 Conclnsions
Bangladesh is a high tlood prone country, The stud} area Tarapur is a union of
Sundarganj !hana nnder Gaibandha district. It is located in the northern part of the
country bounded by two major ri~ers-Brahmapntra (Jamuna) in the eastern side and
Tecs!a in the northern part, Geoinfonnaties provides a broad range of tools for
determining areas affected by J100dsor for forecasting areas likely to be flooded due
to high river water levels. This ,tudy attempts to apply geoinlormatics for flood study
in Tarapur with a view to develop J100d inundation maps, to assess fl{)()ding for
ditl'erent J100d magnitudes, to develop flood hazard vulnerability function of
agriculture and to assess flood hazard vulnerabilily of agriculture for different t100d
magnitudes.
Flood inundation maps identify flooded and non-flooded area, for cach particular
flood level. To assess flooding for dilferent flood magnitudes flood inundalion maps
have been prepared based on J)I'M and satcllite image. A gradual Increase of J100d
inundation has been observed in tlood inundation maps with higher return perIods.
For 2-,5-,lO-,20-and 50-year return period about 87%, 93%, 94%, 95% and 96% of
the area were found to be inundated, respcctively. The same trends were also
observed in flood depth maps. For 2_,5_,10_,20_and 50-}ear return periods the
maximum depths of !load "ater were observed 4.7m, 5.lm, 5.3m, 5.5m and 5.6m,
respectively.
In flood inundation map based OllOEM and water level on 2301 July'04 about 88%
area was found 10bc Inundated while the inundated area was found to 78% on flood
inundation map based on ..creen digitizing of Radar image on 23'" July'04.
Vulnerability functions relate haLard intensity to vulnerability, In order to develop a
vulnerability funetlon for agriculture of the study area a relationship bem-'een flood
depth and damage to agriculture was established. This was achieved through an
• ••
77
extensive field survey together with secondary data analy~is. The magnitude of
vulnerability for agriculture of Tarop\lr union is found to increases more or less
linearly up to three meter depth of water and then il becomes horizontal.
Vulnerability maps represent the response Of susceptihility of an area to natural
hazards or disa,ter like !lood regarding 0 for no damage or 1 for complete damage.
The vulncrabjJjty of Tarapur has been assessed with five tloods viz. 2-, So, 10-, 20-
and 50-year return periods, It i, found that Tarapur union is vulnerable for all the five
return periods as most of/he area are low lying nat lands. Vulnerability due to 10.,
20- and 50-year return periods arc vcry high comparee! to 2-or 5-year return period.
Area close to river bank is relatively bigh vulnerable compared to the area far away
from the river bank, Flood levels due to 20- and 50-year return period, brings almost
100% damages to crops. The magnitude of vulnerability of agriculture due to floods
varies in different locality of the study area due to Uneven distribution ofwpography.
During the monsoon period homestead area i, also found to be affected by flooding of
10-,20- and 50- year return periods.
This study is basically an integration of Ol:,Mand Remote Sensing .The basie element
of this study is Digital Elevation Motlel (OEM). This st\ldy ,hows that integration of
OEM and Remote Sensing can be an dlicient technique w mmlyLc flood inundation
and relatcd damage of an area.
7.2 R«ommelldations
The analyses of this study have bcen done with the interpolated water level bascd on
distance but this procedure act\lally docs not represent the original water level of the
locality. Thi, is why the two flmxl inundation maps ha;ed on Satellite image and
OEM show some discrepancy.
This study is conducted considering only the monsoon flood, but sometimes
significant damages occur due w f1a,h flood that occurs early in the year. Due to flash
t100d vegetables and few rice crops are damaged which al.,,, alIeet economy of the
78
locality. So it is recommended that further study be conducted considering the whole
time period of flooding.
Only the agriculture has been considered for vulnerability mapping in the study area.
Further studies can include other damages including damages 10 household and
infrastructure.
It is clear that proper damage assessment i~the key to prepare vulnerability function.
Damage uS'essmen! in Bangladesh is still not systemalic and consistent. for example
not much data were available on damage to agriculture, infra,trllc!urc and other
sectors for the 2004 flood,
The flood vulnerability map"' developed In this study can also be used for developing
a flood zoning map. This zoning map can also be u,ed for selectlng the types of crop"
and area for cultivation during the monsoon period on the basis of magnitudes of
vulnerability of difterent wnes.
Operational applications of remotely sensed information arc growing. Opportunities
arc cVCnmore with the introdudion of higher resolution satellite images. However,
high costs of image data and technology restrict their use in a developing country like
Banglade,>h.Improvements in this context will certainly enhance an even wider scale
application of the spatial information sciences for sustainahle integrated development.
•
REFERENCES
Alexander, D. (1993). Natural Disaster. wndon, UCL Press Limited.
Aziz, F., Tripathi, N. K., Mark, 0., and Ku,angi, M. (2002). A Customized Flood
Evacuation System to Alleviate Flood by Integrating Hydrodynamic ModelMike I I and G]S, 1st A,ia-Pacifie DBI Software Conference, Bangkok,Thailand,
Blasco, F., Bellan, M.f., and Chaudhury, M.U. (l992). EstimatIng the extent off]oodsin Bangladesh usIng SPOT data. Remote Sensing ofEnvironrnent 39: 167-178
Brouder, J.M. de. (1994). Flood Study In Mcghna - Dhonagoda Polder, BangladeshInlernational Institute for Acrospaee Survey and liarth Scienccs, Netherlands
Bania, S., and Nagasawa, Y. (2003). 'Risk Assessment and Empowering CommunilyHealth Care Facilities in Coastal Cities of Bangladesh'. Proceedings of the 2'dInternational Symposillm On i\'e" Technologies for Urban Safety of MegaCities in Asia, University of'[ okyo, Japan, pp, 449-456.
Bes (1994). Banglade,;h Population Census, Districl Gaihandha, Governmentof Bangladesh, Dhaka.
BBS (1996). Bangladesh Agricu[tural Census, District Gaihandha, GovernmentofBangladcsh, Dhaka.
BBS (2001). Bangladesh Population Census, Dislricl Gaibandha, Governmcntof Bangladesh, Dhaka.
BWDB (1998), Flood Report, Flood Forecasting & Warning Centre Processing &Flood FOrecasling Circle, Bangladesh Watcr Development Board,Governmet\t oflhe Pcople's Republic ofBat\gladesh.
Clirter, S. L (1996), Vulnerability 10 environmental hazards. Progrc,;,; in HumanGeography 20 (4), pp. 529-539.
Durta, D. (1999). Dislribuled Modeling of Flood Inundalion and Damage Estimation.Doctoral Di>sertat;on, University of Tokyo,
Dutta, D., and Tingsanchali, T. (2003). 'Development of Loss functions for UrbanFlood Risk Analysis in Bangkok', Proceedings of the 2nd InternationalSympo,ium on New Technologics for Urban Safety of Mega Cities in Asia,ICUS, The University of Tokyo, pp. 229-238.
-
Dulta,
80
Oulta, D., Herath, S., and Musiuke, K. (2003), 'A Mathematical Model for Flood LossEstimation'. Journal of Hydrology, Volume 277, pp. 24-49.
D.. and Herath, S. (2004). 'Trend of floods in A~ia and Flood RiskManagement with integrated River BU5in Approach'. Proceedings of the 2'0International Conference of Asia-Pacific Hydrology and Waler ResourcesAssociation, Singapore, Volume I, pp. 55-63,
EGIS (1997). Mapping and Monitoring Floods with I{ADARSAT Image,Environmental and GIS Support Project for Water Sector Planning,Dhaka.
EGIS (1998). Microwave Remote Sensing Applications for Flood Monitoring.Environmental and GIS Support Project for Water Sector Planning, WaterResource, Planning OrganiLation, Ministry ofWatcr Resources,Dhaka, Flanglade,h.
Elsner, A., Mai, S" Meyer, V., and Zimmennann, C. (2003). In!~gration of the floodrisk in coastal hinterland management. Proc~~dings of !h~ InternationalConference. Coast Gis. Genua, Italy, 2003,
FAP.19 (1995), Satellite Based Radar for flood Monitoring in Bangladesh,ISPANIUSAID, Flood Plan Coordination Organization. Ministry of WaterReSO\lrces, Dhaka, Banglade,h.
Goswami, 0, C. (1985). Brahmaputra River, A88am, India: Physiography, BasinDenudation, and Channel Aggradation. Water Reso\lr<;e; Resear<;h, V 21, pp959-978,
Goswami, D. C. (1998). Flood Foreca8ting in lhe Brahmaputra River, India: A CaseStudy. http://www.southasiafloods.orglDDcumentlffh/index.html
(;lade, T. (2003). Vuln~rability Assessmcnl in Landslide Risk Analy:;is, Die Erde-Beitrag zurerdsystemforschung 134 (2): 123-146.
G\lpta A,D., !:label, M.S., and Ngoc, P, (2004). 'Flood Damage A:;sessment in theMekong d~lta, Vidnam'. Proceedings of the 2nd 1ntemationai Conference ofA.lia.Pacifie A"ociation of Hydrology and Water Rewur<;es, Singapor~,Volume I, pp. 109-117, Hanoi Statistical Year Book (2003), Hanoi Stati,liealOffice.
Hutehjwn, 1., and Watt, M. (1978). A s}stemalie approach to the flood riskmapping.
Higgins. R.I., and Robinson, OJ, (1981). An Economic CDmpari:;Dn of Differentflood Mitigation Stralegi~s in Au,lralia: A case study, Canberra: AustralianGovernment Puhli.,hing S~rvic~.
.'.
81
Heyman, B,N., Davis C. and Krumpc, p.r. (1991). An 3>scssmcnt of world widecii>aster vulnerability. Di,a,ler Management 43-36.
Huque, I. (2000). National Report of Bangladesh, Environmental and GIS SupportProject for Water Sedor Planning, IAPPRS, vol. XXXIII, Amsterdam.
lCID (1997). 'Evaluation of a scientific System of Flood Forecasting and Warning inthe Ganges, Brahmaputra and Meghna River Basins'. InternationalCommission on Irrigation and Drainage, Proceeding~ of the Seminar OnBangladesh National Committee of !CID (I3ANCID).
JICA (2003). 'Feasibility Study for Improvement of Flood rorecasting and WarningServices in Ihe Peoples Republic of Bangladesh' , Draft Final Report, vol. II,Main Report, Bangladesh Wale, Development Board, Ministry of WaterResources. The People's Repubiic ofBangiadesh.
ILWIS 3.0 AcademIc (200I). Integrated Land and Water Information S}stem, UnitGeo Sofuvare Development, IT Department, lnternational Institute forAerospace Survey and Earth Sciences.
"lam, M. M., and Sado. K. (2000a). 'Development of flood hazard maps ofBangladesb using NOAA-AVHRR images witb GIS'. Hydro!. Sci. JI, 45(3),337-355.
Islam, M. M., and SadCl,K. (2000b). Ranglade,b 'Flood HaLard Asses,ment inBangladesh using NOAA AVI-mR Data with Geographical InformationSystem'. Journal of HydrologIcal Processes. 14(3), pp-605-620.
Islam, M. M., and Sado, K. (2000e). 'Satellite remote sensing data anal}sis for flooddamaged zoning with GIS for lJood management". Journal of HydraulicEngineering, 44, 301-306.
Khalequ=man, M. (2000).1'IClodControl in Bangladesh through Resl Man"gernentPractices.
Khatun, F. (2005). Vulnerabilily Assessment for Urban Flood Disaster RiskManagemenl in Hanoi and Bangkok, lhe capitals of Vietnam and Thailand.MSe Thesis, Asian Institute ofTcchnology.
Mitchell, J.K. (1989). Hazards researeh. In: Gaile, G.L., Willmot, C.J. (Eds),Geography in America. Columbus, OH: Merrill. pp. 410-424.
Maathuis, 8.. and SchierI', M. (2004). flood RIsk Assessment of the 'neler andCulemborgcrwaard, Intrcnalional Institute for Geo-Informaties Science andEarth Ob,ervali"n (lTC).
-
82
Messner, F., and Mc}cr, V. (2005). Flood damage, vulnerability and risk pcrccption-challenges for flood damage research
Ngoc, P. (2003). Flood Damage Assessment in the Mekong Delta, Vietnam, AITThesis no. WM 02-17, Asian Instilule of Technology
Oehi, S., Rahman, M,N .. and Kakiuchi, IL (1991). A study On flood risk evaluationin Bangladesh using remote sensing and GIS. Journal "rthe Japan Society ofPholOgmmmctry and Remote Sensing, Vo1.30, NO.6, 1991, P34,
Parker, D.1., Green. C.H., and Thompson, P.M. (1987). Urban flood prolecti()nbenefits: a project appraisal guide (The Red Manual). Aldershot, UK: GowerTeGhni~alPre".
Pramanik. M.A.H. (1994). Remote Sensing Applications in I)i~asters MOllitming inBangladesh. GIS development. net.
Parker, O.J., and Cannon, T. (2002). Vulnerability Analysis and Disastcrs, Floods,Volumc I, wndon and New York: Routledge, pp. 45-55.
Rasid, H., and Pramanik, M. (1990). 'Visual interprelation of satcllitc imagcry formonitoring floods in Bangladesh'. Journal of Environmen! Managemcnt.14(6),815-821.
Rahman, A.K.M. (1992). Use of GIS, Remole Sensing and Models for flood studiesin Bangladesh: An Analytical Study in a nond prone poidcr in Bangladesh.MSc Thesis, Internationai Institute tn' Aerospace Survey and Earth Scicnccs,Enschede.
Siovic, P. (1987), Per~eption of risk. Scicncc 236:280-285
Sado, K., and I,lam, M. M. (1997), 'Salellite remote sensing data analysis for tloodedarea and weather study: case study of Dhaka city, Bangladesh'. Journal ofIlydraulic Engineering, 4 I, 945.950.
Shrestha, M,S., Heggen, R" Thapa, K.B., Ghimire, M., and Shakya, N, (2004). 'FloodRisk and Vulncrability Mapping using GIS: A Nepal case study'. Proceedingsof the 2nd Intcrnational Conference of Asia-Pacific Association of I Iydrologyand Wa!cr Rcsourecs, Singapore, Volume 1, pp. 180-190, July.
Stcdingcr, l.R., Vogel, R.M., and Georgiou, E.f. (1993). Frequen~y Analysis ofExtreme Evenls, Chap,18, Handbook of Hydrology.ed. D.R.Maidment,McGraw-Hili Inc., New York.
Sundarganj Upa-7i lia Agricultural Department (2005). Rcport on general andagricultural information.
83
United Nation, (199]). Methodolog;. of ri.,k a~ses.lrnen!for mitigating naturaldisasters.
Valkengoed, E.V., Looijcn, W., Huque, I., Hassan, A., and Vente, J.D. (2001),'Market acceSs in Banglade,h for sustainable data services for rice and floodapplications using SAR data' _Fillal report, October.
Wagner, T.W. (1989). Preparing for noodplain mapping and flood monitoring withremote sensing and GIS, Rep(lrt of the w(}rbhop On remOle sensing forfloodplain mapping and flood monitoring, Dhaka, Bangladesh.
World Bank (1989). 'Bangladesh action plan for Hood control,' 91, Asian Region,Country Dept. 1, Washington D.c.
Wu, B., and Xla, F, (1990). Flood damage evaluation sy,tern de,ign for a pilot area OnBangladesh tloodpJain using remote sensing and GIS. European Conferenceand GIS.
WARPO (2002). State of Water Resources System 2001. Water Resources PlanningOrganization, Ministry ofWatcr Resources, Dhaka, Bangladesh, 2002.
i\1'1'lc~lll){-i\
QUESTIONNAIRE FOR LOCAL RESPONDENTSTitle of the Thesis Work: Application ofGeo-illformatics for Flood Stndyat
Tarapur Union ofGaibandha
A. General information of the respondentNo...l.ldentification of the respondent:
Name of the respondent.
rather's name
Age: .
2. Location:Village:
3. Education:.:. Don't know how to read and write-:. Passed ------ class.:. Pa"ed S.S.c.!H.S.C.!Bacheior degree!Ma,ler degree!More.:. Madrasha
Date:
Serial
4. Occupation:D BusincssJTmder,D Tramporl operalor,
o Other '<0"
D Agriculture, D General Labour,o Unemployed, 0 OldlHandieappecl,
o Service,D Artesian,
5. Period ofliving in the village/locality:o U.s, lhanlO ycars,D 10 to 15 year;, 015 10 20 years. 0 20 to 25 years, 0 Morelhan 25 years
B. Agricultural information of Tara pur union6. Type, of crop' cullivate in Tarapur union:
o Au, 0 Aman 0 lRRl 0 B\1f\1 0 Olher; (vegelabl~s, wheat ele.)
7. Crop cultivalion period:
o January-February 0 March-April 0 May-June 0 July-Aug\lst 0 September-
October 0 November-December
-
85
8, Crop harve,Ung period:
o January-February0 March-April 0 May-June 0 July-August 0 Septernb<:r-
October 0 November-December
9. Normal agricultural production (Pcr Acre) for Rab! crop------------------
10. Normal agricultural production (?er Acre) for Kharif crop ._-----------
C. General flood information of Tara pur union11, Types of floods (lccur in Tarapur union:
o Only carly tlash flood 0 Only monsoon flood 0 Both types of flooding
12. Nonnally early flash flood {lccur:; in the monlh-------------------------------------------
13. Normally mon>oon tlood occur in the month-- ••- •••---------------------------------
14. Flood become more deva<;tating if occur in the month-----------------------------------
15. Frequency of early flash tlood in Tarapur union:
o Flood occurs every ye~r 0 Flood occurs every alternative year 0 r lood occurs
every three to five ye~rs Ormore
16. frequency of monsoon flood in I"arapur union:
o Flood oecurs every ye~r 0 Flood occurs every alternative year 0 Flood occur~
every three to five years or mOre
17. Causes of early t1ash t1ood:
o Due to early rainfall in the locality 0 RUl1-offdue to early exce;si,e
rainfall in upstream boundary 0 Due to over flow fr<;>mTee,ta River 0 Due
to over flow from I:lrahmaputra
18. Causes of monsoon tlood:
o Due to mOnSOOnrainfall in the locality 0 Run-off duc to excessive mintidl
in upstream boundary 0 Due to over floy,. from I"eesta River 0 Due to over
flow from Brahm~plltm
19. Intensity ofcarly flash tlood in the locality:
o Occurn once in a year 0 Occurs twice or more in a year 0 Occurs every
alternative year 0 Occurs every three to five years or more
20.lntensity afmon,oon nood in the locality:
o Occurs once in a year 0 Occurs twice or more in a year 0 Occurs every
alternalive year 0 Occurs every three to five years or more
21. Maximum depth of water due to early nash nood in Tarapur unian:
o One feet 0 Two feet 0 Three feet 0 Four feet 0 Four and half feet 0More
than five feet
22. Maximum depth of water due ta man.,oon nood in Tarapur union:
o Onc feet 0 Two feet 0 Three feet 0 Four feet 0 Four and halffect 0More
than fi~e feet
D. Agricultural damage due to early flash flood23. Relation between duration af standing nash nood watcr and damage of crops:
o No damage if duratian become not more thal1------------ day~
o One-fourth of total crops damage if duration---------- day' 0 One-lhird of
total crops damage if duration-------- days 0 Half of total craps damage if
duration----------- days 0 .J wo-third of lolal erops damage if duration---------
days 0 TOlalcrops damagc if duration----------- days
24. Damage duc to onc feet dcpth of water due to early flash flood:
o No damage of crops 0 One-fourth oHotal crops damage 0 One-third "flolal
crops damage 0 Half of lOla1crops damage 0 Two-third of total crops damage
o Total crops damage
25. Damage due 10lwo feel deplh ofwaler duc to early tlash tlood:
o No damage of crops 0 One-faurth oftolal crops damage 0 One-third oftola!
crops damage 0 Half oftolal craps damage 0 Two-third oftolal crops damage
o Tatal CrOP"damage
86
•
87
26. Damage due Ie>three feet depth of water due to early flash flood,
o No damage of crops 0 One-fourth of lola I crops damage 0 One-third oftDlal
crops damage 0 Half artolal crops damage 0 T"o-lhird artolal crop' damage 0
Total ~rops damage
27. Damage due to four feel depth ,,[waler due IIIearl) flash flood:
o No damage of crop, 0 One-fourth of lola I crops damage 0 One-third oftolal
crops damage 0 I lalf of lOlaI crops damage 0 I'wo-third oftolal crops damage 0
Total crops damage
28. Damage due 10 four and half feet depth of water due to earl} flash f1(lod,
o No damage of crop.' 0 One-fClorth of 101a! crops damage 0 One-third of lola I
crops damage 0 Half oftoral crops damage 0 Two-third orlolal crops damage 0
Total crops damage
29, Damage due to more than I1vc feet depth of water due to early tlash tlood :
o No damage of croP" 0 One-fourth "ftotal cmp' damage 0 One-third "ftolal
crops damage 0 Half of total crops damage 0 Two-third of total crops damage 0
Total crops damage
E. Agricultural damage due to monsoon flood30. Relation between duration of standing flash tlood water and damage of erops:
o No damage jf duration beeome nOImOrClhan------------- days
o One-fourth of total crops damage if duralion---------- days 0 One-third of IOtal
crops damage if duration-------- days 0 Half of tOlal erops damage if duration-----
------ days 0 Tw()-lhird of lolal croP" damage if duration----------- da)'~ 0 T"lal
crop~ damage if duration----------- days
31. Damage due to one teet depth of water due to monsoon tlood:
o No damage of crops 0 One-fourth of IOta1crops damage 0 One-third oftotal
erops damage 0 Halfoftotal crops damage 0 Two-third of total crops damage 0
Total crops damage
32. Damage due to two feel depth ofwaler due 10mOnSOOnnood:
o No damage of crops 0 One-fourth oftolal cmp" damage 0 One-third oflotal
crops damage 0 Half of IotaI crops damage 0 Two-third oftola! crop~ damage 0
Total crops damage
33. Damage due to three feet depth of water due to monsoon flood:
o No damage OraTOpS0 One-fourth artolal crops damage 0 One-third of total
crops damage 0 Hair orlolal crops damage 0 Two-third of Iota I crops damage 0
Total crops damage
34. Damage due to four feel depth of water due to monsoon flood
o No damage of crops 0 One-fourth "flolal crops damage 0 One-third of total
crops damage 0 Half of total crops damage 0 Two-third orIOla] crops damage 0
Total crops damage
35. Damage due 10 fOll[ and haIr feet depth of water due to monsoon flood:
o No damage of crops 0 One-fourth of lola I crops damage 0 One-third of total
crops damage 0 Half of lOlaI crops damage 0 Two-third of lOla1crops damage 0
Total crops damage
36. Damage due to more than five feet deplh of water due to monsoon flood:
o No damage of crops 0 One-fourlh oflotal crops damage 0 One-third of total
crops damage 0 Half of lOlaI crops damage 0 Two-third of Iota I crops damage 0
Total crops damage
Signature (,flhe Enumemlor
88
Maximum water levels data of different stations (July 10Seplemher):
89
]'I,'oonkhaua Tarapur Chilmari BabadurabadYear Dalia WL WL WL WL WL1988 52,89 27.5 24.7744847 23.58 19.391989 52.69 26,2 24.8718512 23.69 20,08
1990 51.83 25.91 24.1653306 22.99 19.21991 52.8 26.9 25.3847347 24.22 19.91992 52.3 j 26.39 24.961882 23.8 1,<1.751993 51.91 27.03 24.6769915 23.52 20.361994 52,6 25.42 25,5300634 24,38 19.991995 52,OS 2733 25.0484333 23,9 19.941996 52,3 26.88 25.2684333 24.12 20.371997 52,03 26.82 25.0655806 23.92 19.821998 52.6 2735 25.6547655 24.51 20.171999 52.28 26.52 25.5074299 24.37 19.322000 52.02 26.6 25.2762073 24.14 20.092001 52.42 25.9 J 25.752947 24,62 19.891002 52.74 27.19 26.1688716 25,04 20.182003 52,16 26.65 25.8286833 24.71 20.622004 52.2 27.07 25.8878682 24.77 19.58
Source: BWDB
API'l:NDlX B-I1
Different Methods of Flood Frequency Analysis
Log Normal Distribntion
If a random variable x has a log normal distribulion (LN2) dislribution lhen the log
transformed variable has a normal distribution. The trdnsformed variable is denoled
by
y = In[x]
Wherex)O,
The maximum likelihood method is generally best for fJtling the LN2 dislribution.
The maximum likelihood method gives:
"y=I/"Lln(x,),~I
".,~=1/"L[ln(x,)-yj',=,
Eslimate offlood Uowcorresponding to T-year return period can be oblained from
Where for Uood now z, = ((I _ 11T)0 '.15 _ (II T)o 1l5) / 0.1975
And for low now ZI = ((11 T)Ol15 - (1-1/ T)""") I 0.1975
Where Zr is the standard normal variate, y is mean and S, .tandard de~iation of the
log transformed data.
Detail calculation for esumming of flood flow using I,og Normal Distri'
corresponding 10T"ycar return period is given in Table-A
"•>~••••~~0
="8""=~•"•'"'••• 1>0..l •" "•••i<",£=0'00
~-a-•8"0z~0..l
." . • • • • • • • • • • • rn rn • • •0" • , ,,,,,,,,,,, ,, ,,"N ,,• • N N N N N • N N ,N •~~x. • • • • • • • • • • • • • • •N N N N N N N N N N N • N • N • N
• rn 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0, 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0,• rn rn rn rn rn rn rn rn rn rn rn rn rn M rn MN N N •~~~•~~~• • • • N Nrn • rn rn 0 rn • rn • rn •rn0
• N00
0rn••< < • • 0 • • • • < N < • • 0 0 •< • • • • • • • • rn rn • < • • • <
< • • • rn • • rn • • 0 N • N • , 0 NN 0 M 0 ;: • • rn < 0 0 N M • 0 • < •< < < 0 0 0 0 0 0 0 0 0 0 0 < ",
-, • 0 rn - • rn, • , 0 ,• , • • " N N • < ,• ,• •- 0 • 0 • • 0 , - , • • • • rn
< M • 0 • , 0 rn • - • rn N rn •" rn - N N
~• • • • • N 0 • •, < • • ~" ~
< • • • M N - 00 0 - - - - - - -
rn 0 • 0 • < ,•• • 0 • 0 0 < , Nrn • • 0 • - 0 N , • • • 0 • < • •,0 0 • 0 • • ~, rn • • 0 • - 0 N
0 0~~
0 ;: :; • 0 , ,• • •0 d d d d d 0 0 0 0 0 0, .'. -,, 0 0 , • 0 0 • • " " "• , • • ,• 0 • • • rn 0,• rn < • • 0 • • 0 0 • 0 , • 0 • •• • • • • rn rn rn N N N ~0 0 0,N N N N N N •~~~ ~~~• -rn 0 rn rn rn rn rn rn rn 0 M
0
"•0• , ,• • , 0 rn 0 < M 0 0" • • • 0 0
~]• 0 • 0 , 0 • ,N • • • • • • 0 rn• , • N • 0 ,•• • • • • < • •0 • N • • rn 0 , 0 • • , ,, 0< • • :;~• • rn • N 0 0 • • ,• <
•• • • • • • • • • • • • •,0 • N N N N N N N N N N N N N • N0C0•,, 0 < N rn • • 0 ," < N rn • •.,• 0 < < < < - - - -0
92
3-1'aramet~r Lug Normal Distribution
The three-parameter log-normal (LN3) distribution has one more parameter Ihan the
LN2 distribution has. When a lower bound parameter; is subtracted from dala and
then logarithms arc taken. the transformed variable has a normal di<'lriblltion. The
transfQrmed variable is denoted by
y=ln(x-;l
A simple and defined e,timator of ,; is the quantile-lower-bound estimator
When x, +xn -2Xrn",,,,.>O, where XI and Xnare the large~t and smallesl observed
value~ re'pecti~ely; i, the sample medium equal to X'" for odd sample SilCS
n=2k+land Y,(x,+Xf+') ror n=2k.Whcn x,+xn+2xm",,""< 0 the formula
provides an upper bound so that y = In(.; - x) would be nonnaily di,tribuled.
Median 18the middle mo,l value ora given sample.
Again';, y and S, (standard deviation) can be obtained from the transformed data.
Estimate of flood flow corresponding to T-year return period based on y = In(.; - x)
can he obtained from
And the ealeulated tlood flow using 3-Pararneter I.og Normal Di,lribulion is given in
Table-B
"•>01••••""00.,0
'""0",~-"••0•" ~•~ "" ~••,g0•.,-.,Q.,,"~•0•"0",,~"
j, • • N • ,• • ~0 N " 0 ,• " 0 •"0- , • • • • 0 ~• • • • " " 0 ,I~!;:' N • , • • 0 ~
~~0 • • • • ,• •• " • • " •• • • • • • • • •'i N N N N N N N N N N N N N N N N•
-' N• • • • " • • 0 • • • • • • • N NNilh • • ,• • • • ~0 • • N • ,• • •• N 0 • • N • •• • 0 0 N ~0 • •,• • • • • • • • • • • " • " • ,.; •"-.• 0 " • N " N • 0 N 0 • 0 ,• •.oN ,• 0 •~• • N ,~• ,• ,~•"~• N 0 ,• • N • • ~" N 0 • •, N •• • • • • • • " " " • • 0 0~~~~- ~~~~~~~~~~~~
~• •• 0 • • •,~N ~• • 0 0 •~ • 0 • • 0 • 0 0 • ~• • • ~N " • • 0 • • 0 • • N • N • , 0 NN • 0 0 :; :; " •~0 0 N " • • • ~"~~~ 0 0 0 0 9 9 9 9 9 9 c c.- • " - " N 0 • ,,- ~~• - •• • • "," • • " , • •• N ~ • • • 0 • • • • •~0 0 0 • 0 , • • 0 • • • • • • N
" N - • • ,•~0 •• 0
, • , • N •0 0N ~ " • - •• • N ~ 0 •• • 0 " ,~- - ~- -~~0
00 • • • • • • • • • • • •,0 ~N N
,• N , , • N N, • N ,,• • • 0 ~ 0 • • • - •• 0 ~N
~ ~• • 0 • ,, • :; 0:; 0 0 0 0 0
0 0 0 0 0 ~
•,--.,~0
•,N•0•-~~
N 0 0 - • • ,,• • N • •~•• •• • • • • • ~• 0 • • N 0 • • , N•~0 N • • • - - • • • • 0 0
- " N • • • • • •• • • • • • •• ,"• ~- - - - - ~~~~~~~- - - -••0
0 N0•N~
0 N
~•,• ~,N
•0 0
'" • • • 0 • 0 , • • • • • N N • , •• • N • • • 0 • ,•• • • , N N •~ ~ E - • • ,~• • 0~NO 0 • • , •~
• • • • .,.;.,; • •• • • • • • •o • N N N N N N N N N N N N N N N N N••• 0 - N N • • • ,~,~N 0 • • • ,•• - - - - ~~~~0
94
Pearson Type III Distribution
Th" Pearson Type III Distribution (P3) !la, three parameters, A P3 distribution can be
fitted to a ,ample of data by using the product moment oflile data.
Estimate {If flood flow corresponding to T-year return period can be obtained from
Xc =x+KjS.
Where K,. is a frequency factor. The frequency factor for T<l 00 years and absolute
value of skewness<2 are well approximate by the Wilson-J-1ilferty tramfo,mation.
Kj = 21 x,{1 + (2T • g, /6) - g: /36)"' - 21 g,
Where ZT is the standard normal variate andS" g, are standard deviation and
coefficienl of ,kewncss of normal data (not transformed) respectively.
The calculated flood flow using Pearsoll Type 111Di,lribulion is given in Table-C.
96
Log Pearson TJpc III Distribution
Log Pearson Type 1lI Distrihuticm (l.P3) dislribution describes a randoln variable
whose logarithm, are P3 distributed. Thus
y = lnx
the method of e,timaling the parameters of the LP3 distribution is similar to lliat for
the P3 distribution except that logarilhms of the data are to he taken before estimating
moments, Thlls for a set of observations {x""" ""'''-"J the transfonned dala arc
given by.", = In x, y, = In x, ,Y, = lnx.
E>tirnale of flood !low corresponding to T-year return period can be obtained fmm
K, =2!g,(I+(Z} *gy!6)_g;1J6)'~2/gy
Where g,. is the coefficient of;kewness of log transformed data.
The ,,,Iculated flood 110wusing Log Pearson Type j II Distribution is given in
Table-D
••
"".:="•'i•e--••.ip"•0•"•e--•]•
~"•~ •e-"••~•"-,~---,•0I:••,•~"••
'"', • 0 • • • ~0 0 • • • • • • • N
• • N ~• • •~• • • N • • • ~•" . • • ::; :;~• • • • N ~~0 • •,••• • • • • • • ••••• • •xl N N 0 0 N N " 0 0 0 N N" 0 0 " 0
" ~ ~0 • • • ~,• ,, • ro • • 0
" • • ~,• 0 ,• • " • 0 • 0 • 0
• • • • • • • • • 0 " 0 • 0 0
~N N 0 0 N ~~N 0 0 N " ~~N 0
• • • • • • • • • • • • •0 • • • • • • 0 •~• • • • • •• • • - • N • • • - 0 ~0 • • •
" • • • • • • • • - • • , • • ~•• 0 0 •~• • N • ~~• • • • - •~• - 0 ci 0 0 0 9 9 9 9 9 c c
0•• , N••0
•"
N0N00
-~ ••
9
• -•~S
•••"- ~> •
~• • • • , • • ~• • • • • • • • •• ro ~• • • • • • •~0 ,• • • •~• • • • • ro •~" 0 N - ~
0 0 •;; ~N " 0 0
~~0 N 0 0
~N ~
• • • • • • • • • • •~- • • 0 • • • • ~N • • 0 0 •~• • • • • • • • • • •~• • • -" • • • • • • • • • • 0 • 0 • • 0 0
• • 0 ~~• • • 0 0 N • • • • • •• - - 0 0 0 0 <; 9 9 9 9 C C
- • • • • - ro • • - • • • ,• , • • • 0 N • - N • " ,• •0 0 0 N • 0 - • • • • • •0 N • • • • , ro , • • • • N • •" • - 0 - 0 • •• ro • 0 N • • •" • • •~" •
0 - • ro • • 0 • 0
• • 0 0 - - ~~~• -• • • 0 • ~, N • • • • 0 • • , 0• • • 0 • - • N , • • • 0 • ~• 0- • • 0 • • 0 • • • • 0 • - • N0 0 - N ~• • • • • d ~"
,• • •0 0 0 0 ci ci 0 0 0 0 0 0 ci
•.':::E • • • • • 0 • • • • • • 0 0 • , •.~"• • N • • • 0 • , • • • • , ,• •,e~:i' ~• • N ~::• • 0 0 0 ::; • • ,• -• • • • • • • • • " • • • ••• 0 N 0 0 N N N 0 N 0 N 0 " 0 0 0••,, 0 - 0 ro • • • ,E ~0 • • •• ,•• ~~- - ~~~•0
98
Gumble Distribution
The Gumblc distribution is (he rno,1 wide!} used distribution in flood frequency
analysis. The Gurnble di<tribution is also known as the Extreme value Type-l (EVI)
diwibut;on. The probability weighted moment (PWM) mdh",l I, employed in this
study to estimate parameters of Gumble dislributi<;ln. Para'neter estimation by the
PWM method is a recent m~thod which is widely used for those distributions that
COP can be explicitly inverted. Estimate of flood 110wcorresponding (0 T-year return
period can be obtained from
XJ =u+O:*Yr
Where u = x - 0.57720:
YI =-In(-In(l-]IT))
The calculated flood flow using GUffiblc Distribution is given in Table-E.
,"
~
Gumble Distribution for Water Level at Tarapur Union ofTeesta River
Table-E
Sorting InOrder, des~~~ding [-tn(1- YT=-I~l-
m ,I T 11 I " iii;, '" ,. ," " ", 26.169 0033 " 1.841 0,034 3.384 05111 25,284 0.399 25.054 26403, 25.888 0,089 11.25 1.348 0,093 2.374 26000, 25.829 0;« 6,923077 1.058 o 156 1.858 25,794, 25.753 0.200 ; 0,839 0223 1.500 25.652
; 25,855 0.256 13913043 0654 0,295 1.220 25.540
" 25530 0.311 3.214286 0490 0.373 0,987 25.447, 25507 0.367 2.727273 0339 0.457 0,784 25.366
• 25,385 0.422 2.368421 0.195 0.549 I 0600 25.293
• 25.276 0,478 2.093023 0.055 0.650 0.431 25228
" 25.268 0533 1,875 .0.083 0762 0.272 25,162
" 25.066 0589 1.698113 -0,223 0,889 0.118 25101
" 25.048 0644 1,551724 -0368 1 034 -0.034 25,040
" 24,962 0700 1428571 .0,522 1,204 -0186 24980
" 24872 0.756 1,323529 -0689 1 409 -0,343 24.917
" 24,774 0.811 1.232877 -0,879 1,667 _0511 24.850
" 24677 0.887 1.153846 _1.109 2.015 -0,701 24.775H 24 165 0.922 1.084337 -1.422 2.554 I -0.938 24.680
Correlation between xi andXT= 0.950228
99
IOU
APPENDIX B-IlI
Goodness-of- fit analysis for selecting the most apprnpriate di,1ributi,m
SLNO. T es of Analysis Value RankI Correlation between Xi& XT (Log Normal 0.4462835 4
Distribution)
2 Correlation between ~~ artd Xl {J-Parameter -0,9888491 5Lo!! Nannal Distribution
3 Correlation between Xi and XT 0.9904610 I
(Pearson Type III Distribution)
4 Correlation between XI and XT (Log Pearson 0,990355 2Type 1II Distribution)
5 Correlation between Xi and XT (Gumble 0950228 3Di,lribution)
APPRi'lDlX H..IVPearson Type III Distribution
Water levels regarding difterent return period hased on Pear"," Type III Di,lriblltion
ReturnPeriod
I /Year\ Z, X S, G, K, XT"'X+KT.ST
2 0 25.284 0.5111 -0.331 0.0550589 25.31193569
5 0.8386 25.284 0.5111 -0.331 0.8503944 25.718407
10 1.2813 25.284 0.5111 -0.331 1.24028342 25.9176589
20 1.6493 25.284 0.5111 -0.331 1.5491136 26.07548876
50 2.0636 25.284 0.5111 -0.331 1.8805788 26.24488964
APPENDIX C_I
101
Items Bangladesh Gaibandha % of Ban ladeshGros, cropped area 28616451 608419 2.13intensity of cfOr:!p'in % 1'4 190 -AmlUl-Local 6036661 84064 1.39
HYV and Pa"am 4510985 153947 3.41Aua-LllCal 3156748 623 j 0.20
I-IYV and Pajam 991867 15189 1.53l3oro- Local 1129701 13965 1.24
IIYY and Pa"am 5007116 175620 3.5 j% of1JYV Rice to gr">s cropped 36,7 56.66 -areaWheat-lA)cai 508042 8780 1.73
HYV 1013947 34222 3.38Jute 1297433 34311 2.64Sugarcane 329016 329016 2.50
Basic Information of Cropped Area (Area in acres) ofFann Holdings:
Source: BSS, Census of Agriculturc-1996
Croppin!!: pattern of Tara pur union
Rabi • KhaTif I _Kharif II
• Boro(rice) - Fallow - Transplanted Aman {Rice}• Wheat -Jute - Transplanted Aman
• Polato - Boro - Transplanted Aman
• Mustard - Boro - Transplanted Aman
• Vegetable -Jute - Transplanted Aman
• Boro - Vegetable - Transplanted Aman
Source: BBS. Census of Agriculturc-J996
APPENDIX C-I1
Comparison of land use practices (If Sumlarganj thana of 1996 with 1983-1984Census,
102
f Area in acre,)Hem, 1983 1984 1996
All Non-farm Tota! All Non-farm TotalHoldings Holdin~~ Boldin ; Hold;n s
Homestead 4042 517 3525 5258 1112 4146AreaPercentage 100 12.79 87,21 100 21.15 18.85]\'1.'1 67882 41 67841 58258 88 58170CultivatedAreaPercenta " 100 0,06 99.94 100 0.15 99.85Gross - - 140820 - 113669CroppedAreaPercentage - - 100 - - 100Cw AreaHYV Aus - - 2876 - 3774and PajamHYVAman - - 6053 - - 30902and Pajam
HYV Boro - - 2430 - - 27863and PaiamWheal - - 10516 - - 9603Pulses - - 3855 - - 265Oil Seeds - 1038 - - 2902Cra,h - - 13759 - - J j 711Crops
Source: BBS. Census of Agriculture-1996
