Fires2 6/15/09 8:14 PM Page 1

2

United Nations Environment Programme (UNEP)Division of Early Warning and AssessmentUNEP Complex, Gigiri, NairobiP.O. Box 30552, 00100, KenyaTel: +254-20-7623785Fax: + 254 –20- 7624309Web: http://www.unep

© UNEP, SARDC, 2009

ISBN 92-807-2634-X

Information in this publication can be reproduced, used and shared, with full acknowledgementof the co-publishers, UNEP and SARDC.

Citation: SAFNet, Africa Environment Outlook Case Studies: Impacts of Fires on the Environment,UNEP and SARDC, 2009

Cover and Text Design SARDC: Paul Wade and Tonely NgwenyaPrint Coordination: DS Print Media, Johannesburg

Fires2 6/15/09 8:14 PM Page 2

Background to the Case Studies 5

BOTSWANA 7Abstract 7Introduction 7Socio-Economic Profile of Botswana 7Climate 8Land Use Systems and Patterns of Fire 9Causes and Effects of Uncontrolled Fires 11

Causes of Fires 11Effects of Fires 11

Evolution of Fire Management 12Current Status of Fire Monitoring and Management 14

Fire Prevention 14Fire Monitoring 14

Towards Reducing Vulnerability to Fire Outbreaks 15Conclusion 16References 17

NAMIBIA 19Abstract 19Socio-Economic Characteristics 19Climate, Land Cover and Land Use 19

Fire 19Causes and Effects of Fires 20

Landscape Characteristics and Fire Regimes 20Past Fire Uses in Namibia 22Fire Management 22

Evolution of Fire Management 22Current Policy Status 23Human Vulnerability to Fire 23Conclusion 23References 24

SENEGAL 25Abstract 25Introduction 25Landscape Characteristics and Fire Regimes 25Description of the Main Fire Regimes 28

1. The River Valley 282. The Ferlo 283. The Peanut Basin 284. The Transition Zone 285. Eastern Senegal 286. Central South 287-8. Mangrove 28

Temporal Distribution of Bushfires in West Africa 29Fire Management 29Conclusion 30References 32

UNITED REPUBLIC OF TANZANIA 33Abstract 33Introduction 33Physiographic Regions and Climate 33Land Use and Vegetation Types 33Causes and Effects of Fires 34Protected Areas and Forest Plantations 34Public Land 35Fire Management and Monitoring - Tarangire National Park 35

3

CONTENTS

Fires2 6/15/09 8:14 PM Page 3

4

Fire Issues in Tarangire National Park 36Spatial Distribution of Fires in Tarangire National Park 37

Evolution of Fire Management in Tanzania 38Evolution of Fire Management in Urban Areas 38Evolution of Fire Management in Forest Areas 38Current Statutes 38Limitations of the Fire Management Strategy 38Human Vulnerability to Fire 38Conclusion 38References 39

ZIMBABWE 41Abstract 41Introduction 41Background Information about Zimbabwe 41Land Use, Climate and Vegetation Cover 41Fire 43

Use of Fire in Zimbabwe 44Causes of Fire in Zimbabwe 44Impact of Fire 44

Forest Fire Management and Monitoring 45Use of Remote Sensing to Demonstrate Fire Distribution 46Forest Fire Policy and Control in Zimbabwe 46Early Warning Systems on Fire 47Conclusion 47References 48

List of BoxesBox 1 Community experience in effective fire suppression in Mgori Forest Reserve 35

List of FiguresFigure 1 Active Forest Fires in Botswana 10Figure 2 Temporal Distribution of Fires in West Africa 29Figure 3 An Example of Landsat Burnt Area Mapping in North Western Matabeleland, Zimbabwe 46

List of GraphsGraph 1 Fire Outbreaks in Botswana (1991-2001) 11Graph 2 Fire Damage in Botswana (1991-2001) 12Graph 3 Land Use/Cover for Tanzania 34Graph 4 Causes of Fire Outbreaks in Forest Plantations 35Graph 5 Common Disaster Occurrences in Tanzania 38Graph 6 Causes of Fire in Demarcated Forests in Zimbabwe (1985-1988) 44Graph 7 Areas Burnt in Demarcated Forest Areas in Zimbabwe (1989-1992) 45

List of MapsMap 1 Land Use in Botswana 8Map 2 Botswana Ecological Zones 9Map 3 Fire Frequency over Botswana from 1996 to 1998 10Map 4 Fire Regions in Namibia 20Map 5 Monthly Time Series of Active Fires 21Map 6 Fire Frequency in West Africa (dry season 2003-2004) 26Map 7 Fire Frequency in Senegal (dry season 2003-2004) 26Map 8 Inter-Annual Rainfall, 1961 -1990 26Map 9 Senegal: Isohyetes of the Rainy Season, 2003 26Map 10 Vegetation and Land Use in Senegal 27Map 11 Major Land Use and Cover Types in Tanzania 34Map 12 Major Vegetation Types of Tarangire National Park 36Map 13 Distinction of Burnt Areas in Tarangire National Park 36Map 14 Zimbabwe Vegetation Cover 42Map 15 Zimbabwe Land Tenure 43Map 16 MODIS Active Fires Detected over Zimbabwean Between 7 and 8 August, 2004 46

List of TablesTable 1 Potential Sources of Fire Outbreaks in Botswana 11Table 2 Use of Bush Fires in West Africa 26Table 3 Fire Incidences in Reserved Forests of Morogoro Region 35Table 4 Details of the Fire and Rescue Services Act No.3 (1985) and the Forest Act No.14 of 2002 38Table 5 Estimates of the Extent of Land Cover Types of Zimbabwe 42

Fires2 6/15/09 8:14 PM Page 4

5

Background to the Case Studies

The 8th Session of the African MinisterialConference on Environment (AMCEN)endorsed the preparation of an Africa

Environment Outlook (AEO) report with theassistance of UNEP, to provide a comprehen-sive scientific assessment of the environment,policies and environmental management pro-grammes in Africa. The report was launched atthe 9th Session of the AMCEN, which adoptedthe AEO as a tool for monitoring sustainableenvironmental management in Africa and alsoto act as the environmental reporting frame-work at national levels.

The first AEO report (AEO-1) was publishedin 2002. The second AEO report (AEO-2) waslaunched in 2006. AEO-2 aims to:Profile Africa’s environmental resources as anasset for the continent’s development. It high-lights the potential of the region’s naturalresource base to support the development agendaof NEPAD and the opportunity cost of its misuse.AEO also serves as a monitoring and evaluationtool for the implementation of the programmesand activities contained in the Action Plan on theEnvironment Initiative of NEPAD.

Fire was not considered a major environ-mental issue in Africa in both AEO-1 andAEO-2 reports. However, SAFNet was invitedto participate in the AEO-2 and GEO-4Regional Consultative Meeting in April 2004,Nairobi, Kenya and this provided the oppor-tunity to introduce fire in the AEO reportingprocess.

The frequency of fires and total surface areaburnt annually has increased over Africa. Thetiming of fires has also changed. These new fireregimes have a potential to cause significantnegative impacts on ecosystems goods andservices and ultimately affect livelihoods,reducing the potential to take advantage ofenvironmental opportunities.

Fires contribute green house gases and havea role in climate change processes. AfricanGovernments have limited capacity to addressfire issues.

Widespread poverty and other stresses con-tinue to make fire an integral part of land usemanagement for the rural based African com-munities. Under natural conditions, fires have a

positive role in ecosystem processes. As a resultfire needs to be included in all environmentalassessments to facilitate the development ofguided policies on fire as an environmentalasset and reduce its potential to constraindevelopment.

It was behind this background that UNEPDivision of Early Warning and Assessment(UNEP/DEWA) engaged the Southern AfricaFire Network (SAFNet) to facilitate the prepa-ration of focused information on fire and a net-work of fire experts in the continent to guidethe process of incorporating fire issues in theAEO process.

SAFNet facilitated the production of fivecase studies over Africa on:

Veld fires and their impacts on the environmentand human vulnerability, the different policyresponses and institutional frameworks. The casestudies use satellite imagery to highlight environ-mental change due to fire and its impact onhuman vulnerability.

SAFNet organize a planning workshop thatreviewed the case studies, determined a strate-gy for contributing in the AEO process, fireearly warning and policy issues.

These case studies covering Botswana,Namibia, Zimbabwe, Tanzania and Senegalwere presented at the Fifth SAFNet workshop:“Towards Meeting Fire ManagementChallenges in southern Africa” held inMangochi, Malawi, 9-13 August 2004. Theworkshop was sponsored by UNEP/DEWA,the governments of Belgium, Ireland, Norwayand the Netherlands, and UNEP EnvironmentFund. Technical support and satellite imageswere provided by NASA Land Cover Land UseChange and NASA Application ProgrammesSTART.

The case studies in this report provide a pro-file of the fire issues across Africa. Taken togeth-er they provide good comparison betweenmajor sub-regions of Africa in terms of use,causes, effects and human vulnerability to fire,technologies for monitoring fires, current andhistorical policy and institutional frameworksused to address the issue of fire.

The case studies were edited and designedby the Southern African Research andDocumentation Centre (SARDC). SARDC alsocoordinated the printing of the book.

Fires2 6/15/09 8:14 PM Page 5

6

Fire Case Study Countries

Fires2 6/17/09 4:29 PM Page 6

Opha P. Dube1 and J.G. Mafoko2

AbstractFires are common in Botswana and are consid-ered part of the evolution of the savanna vege-tation type found in the country. Different landuse systems in Botswana have also evolvedwith fire.

The extent of burning in the country is notonly closely linked to rainfall, which deter-mines the amount of fuel load, but also landuse, which results in most fire outbreaks.

The rise in population coupled with changesin land use management, poverty and otherstresses such as HIV and AIDS have led to greaterreliance on fire as a land management tool. Thisincreases the risk of fire outbreaks, which con-tributes to human vulnerability. Frequent firesresult in changes in the vegetation cover density,species and structural composition that ultimate-ly influence hydrological processes, wildlife habi-tats and the supply of veld products.

A large section of the population relies onveld products for food and as a source ofincome. The tourism and livestock industriesare directly dependent on the supply of naturalresources, which are susceptible to frequentburning. It is for this reason that the fire man-agement strategy in Botswana is focused on firesuppression. However, the leading agency infire management, the Agricultural ResourcesBoard (ARB) in the Ministry of Environment,Wildlife and Tourism has limited resources toeffectively monitor and manage fire. TheHerbage Preservation Act of 1977 guides firemanagement but there is no fire policy.

Most fires are not recorded and informationon the area burnt and the frequency of fire isbased on rough estimates. It is difficult to assesstrends in fire history and the implication of thison natural resources and human livelihoods.Incorporating geo-spatial technologies in firemanagement in Botswana could greatlyimprove this situation. Satellite data is valuablein assessing the risk of fire danger, monitoringactive fires, determining extent of burn scars,and in assessing post fire recovery.

Botswana’s role in the Southern Africa FireNetwork (SAFNet) is to facilitate access to infor-mation on the use of satellite based products in

fire management and to help raise the profile offire in the country and in the region given thatfire is a potential trans-boundary problem.

IntroductionFire, as part of natural processes, has a positiverole in the vegetation structure and composi-tion, and helps recycle nutrients contained inold and dead trees. There is, however, concernthat the frequency, extent and pattern of burn-ing are increasing due to human activity. Forexample, fires burnt about 20-30 per cent ofBotswana between 1996 and 1997 (CentralStatistics Office, 2000). It is believed that thedamage from these fires has grown to outweighthe benefits of fire on the ecosystem.

Botswana is semi-arid, and there are severalinstances when rains fail following burning.This leaves the ground exposed for long peri-ods which exacerbates land degradation anddeprives livestock, wildlife and people of liveli-hood resources especially during periods whendemand for these resources exceeds supply.Fires destroy soil microbia, timber and non-tim-ber products, including forest-based servicessuch as recreation, and cultural heritage sites. Inaddition, infrastructure and property are lostwhile uncontrolled fires also threaten lives ofpeople, livestock and wildlife. Uncontrolledfires have negative impacts on the economyalthough so far there are no studies that haveattempted to quantify the cost and benefits offires in the country.

It is, however, worth noting the importanceof fire as a management tool in the regenerationof natural and pasture vegetation. Not only is itimportant in the regeneration of vegetation butit also reduces levels of unwanted pests. Fire asa cheap management tool should not be usedwithout enough knowledge of the season out-look that gives the likely rainfall amounts, windspeed and its direction.

Socio-economic Profile of BotswanaBotswana has a population of 1.7 million peo-ple, 70 per cent of whom live in rural areas(Kalabamu, 1994; Republic of Botswana, 2003).With a per capita GDP of US$3 135 in 2001, thecountry fell under the middle-income develop-ing country category (Ministry of Works,

7

BOTSWANA

1. Corresponding author: Department of Environmental Science, University of Botswana Private Bag 0022, Gaborone. Email;dubeop@mopipi.ub.bw or dubemop@yahoo.com

2. Department of Agricultural Resources Board, Ministry of Environment, Wildlife and Tourism, Private Bag 00424 Gaborone, Botswana,Email: gmafoko@gov.bw

Fires2 6/15/09 8:14 PM Page 7

8

Transport and Communication, 2001; BotswanaNational Agenda 21 Coordinating Committee,2002). Despite growth in the economy, about 48per cent of households in 1985/6, and 38 percent in 1993/4 were below the poverty datumline (Buchanan-Smith, 1998; Botswana NationalAgenda 21 Coordinating Committee, 2002).

Most of the poor households are female-headed and found in rural areas where they aremainly engaged in subsistence mixed farming(livestock rearing, arable farming and gatheringveld products). Fire is the cheapest and easiesttool to use for these households to meet differ-ent needs such as clearing fields.

The communal land tenure system wheremost of the subsistence farming is practiced ismakes up about 70 per cent of the country (Map1). Protected areas account for 17 per cent of thecountry’s total surface area, while 13 per cent is

made up of freehold and Government farms(Moyo et al., 1993; Ministry of Works, Transportand Communication, 2001). A significant num-ber of fires are lit in communal areas but even-tually spread into other land tenure systems.

ClimateBotswana falls within the centre of the southernAfrica landmass extending over an area of about582 000 square kilometres (Ministry of Works,Transport and Communication, 2001; BotswanaNational Agenda 21 Coordinating Committee,2002). The country has a semi-arid climate char-acterised by extreme events particularly, droughtwhich reduces the risk of fire. Periods of aboveaverage rainfall also occur and these are usuallyfollowed by widespread fire events due to abun-dant fuel load. Rainfall declines towards the westof the country and the average annual rainfall

Land Use in Botswana Map 1

Source: Department of Surveys and Mapping, 1999

Fires2 6/15/09 8:14 PM Page 8

9

ranges from 200mm in the southwest to 650mmin the northeast (Ministry of Works, Transportand Communication, 2001).

The mean monthly maximum temperatureranges from 20 - 29oC in winter to 30 - 35oC insummer. Due to low humidity night time meantemperatures regularly drop to near zero inwinter (Ministry of Works, Transport andCommunication, 2001). The long dry seasonspanning from around May to October pro-vides favourable fire conditions particularlyafter sufficient rainfall has been experienced.Conditions are generally windy with lowhumidity and hot temperatures after July. Thepeak of the fire season is September, but mostnatural fires occur between October andNovember when thunderstorms are experi-enced. This was probably the peak of the natu-ral fire season before human induced firesbecame widespread.

Land Use Systems and Patterns of FireIn general, there are three major land use sys-tems, the hardveld, the dry sandveld and thewet sandveld (Map 2). Most of the populationand development infrastructure is found in theeastern part of the country, the hardveld, whichhas loamy soils that are more suitable for culti-vation. The hardveld supports more palatable

pastures and as a result has a longer history ofhuman settlements and livestock production.For this reason it generally has lower fuel loadsto sustain large fires (Map 3 and Figure 1). Mostfire scars in the hardveld tend to be smallbecause fires can be spotted and put out quick-ly or infrastructure such as roads acts as barri-ers preventing the fire from spreading. This isin contrast to the Kalahari sands, known as thesandveld, covering about 80 percent of thecountry (Thomas and Show, 1991; BotswanaNational Agenda 21 Coordinating Committee.2002.) (Map 2, Map 3 and Figure 1).

The dry sandveld found in the central andwestern parts of the country is dominated byshrubland with a dense, generally unpalatablegrass layer and this combined with the dry hotweather results in high risk of fire (Maps 2 and3). The far south-western corner where rainfallis below 200 mm has a more desert environ-ment (Thomas and Shaw, 1991), and as a resultthere are limited fire incidents. Although thecattle industry expanded into the sandveldfrom the 1930s, areas of intensive grazing arelocalised due to limited availability of freshwa-ter. Further, a large section of the sandveldcomprises protected areas such as the CentralKalahari Game Reserve (CKGR) and theseaccumulate large fuel load which when ignited

Botswana Ecological Zones Map 2

Source: Ministry of Agriculture, 2003.

Fires2 6/15/09 8:14 PM Page 9

10

results in large fires. Map 2, shows that most ofthe areas that burnt twice in the central parts ofthe country between 1996 and 1998 fall underthe CKGR. Most of the sandveld is still not eas-ily accessible which makes it difficult to detectfires in their early stages, resulting in largeareas being burnt in each fire incident in con-trast to the hardveld.

The wet sandveld in the northern part of thecountry is dominated by the lacustrine systems(Lake Ngami, Mababe and Makgadikgadipans) and the alluvial systems such as theChobe-Linyanti and the Okavango Delta in

Ngamiland District, a classified wetland whichis protected under the Ramsar WetlandsConvention (Thomas and Shaw, 19991). Thewet sandveld is the most attractive to touristsdue to availability of water and abundantwildlife. Because these areas generally havehigher moisture content, large protected areasin the form of Game, Forest Reserves andNational Parks, and are sparsely populated,they support more dense vegetation and as aresult are prone to annual burning (Fig. 1).About 110 fires were reported between Januaryand April in 1981 in Ngamiland District only(ARB, 1981).

There are concerns that increased fire activi-ty over the Delta will have negative impacts onbiodiversity of this wetland. In Map 3 mostareas with fire frequencies of between 4 – 10times in 15 years are found in the Okavangofloodplains grasslands which are flooded fre-quently.

In some parts of the wet sandveld, forinstance the Chobe District, vegetation dam-age by elephants contributes to availability ofdry material that can be ignited easily. TheChobe District is home to at least 60 000 ele-phants out of a large population of 130 000found in Botswana (Department of Wildlifeand National Parks, 2003). However, this partof the country also experiences the highestannual rainfall in the country (650 mm) and asa result miombo forests are supported. Fires inthis area generally burn slowly, are moreintense and exhibit crown torching on windydays. This is in contrast to the rest of the coun-try where limited wood fuel results in fasterand less intense fires.

Active Forest Fires in Botswana Figure 1

Source: Ntabeni, 1999

Fire Frequency over Botswana Map 3from 1996 to 1998

Active Fires (shown as red dots) captured on MODIS 1 km spatial resolution fire product. On the left are active fires inthe dry sandveld - central to west and on the hardveld south-eastern parts of Botswana: 29th July 2002. On the right, theimage shows active fires on the wet sandveld north-western part of the country (on the border with Namibia) with a lagerfire west of the Okavango Delta: July 2001 (http://rapidfire.sci.gsfc.nasa.gov.).

Fires2 6/15/09 8:14 PM Page 10

Causes and Effects of UncontrolledFires

Causes of Fires The Agricultural Resource Board (ARB) fieldreports (unpublished material) attribute 90 percent of all fires to human causes. Most causes ofwildfires in Botswana are related to land use.For the rural population fire is a cheap and eas-ily available alternative to expensive humanlabour or modern equipment (Table 1).Although the use of fire has a long history,widespread poverty and other socio-economicchallenges such as migration to urban areas anddiseases such as HIV and AIDS, result in short-age of labour in rural areas, causing an increasein reliance on fire. Female-headed householdsconstitute the majority of the poor and are morelikely to suffer shortage of labour making theuse of fire a necessity for this group (BuchananSmith, 1998). Unfortunately it is during thesepositive uses of fire that uncontrolled fire out-

breaks occur and because the users are alreadyresource poor they are less likely to deploy ade-quate resources to suppress an active fire(Graph 1). Ironically it is also the same groupthat is most affected by fire damage.

The number of fire outbreaks is related torainfall and as a result few outbreaks werereported over the drought years; 1992, 1993,1995 and 1998 due to limited fuel load (MinistryOf Environment Wildlife And Tourism – ARB,2003).

Indications are that farmers, hunters and gath-erers of veld products cause the majority of fires(Table 1 and Graph1). There are few cases wherewildfires are caused by lightning. It is estimatedthat lightning accounts for about 10 per cent ofthe fires in the country (Ministry of Works,Transport and Communication, 2001; BotswanaNational Agenda 21 Coordinating Committee.2002). This is due to the fact that thunderstormsseldom occur in the period between May andOctober, when dry fuel load is abundant butoccur mostly from November to April.

Effects of FiresUsed properly, fire is a very essential manage-ment tool in Botswana as in the rest of Africa(Table 1) (Frost, 1998). For most rural communi-ties controlled use of fire is a necessity and ispart of sustainable livelihoods.

If the frequency is within the required thresh-old wildfires have a positive role in the vegeta-tion structure and composition, and help torecycle nutrients. The ecological benefits of fireon the savanna ecosystem in general have beendocumented (Bond et al., 2004) but there are no

11

Fire Outbreaks in Botswana Graph 1(1991 to 2001)

Potential Sources of Fire Outbreaks in Botswana Table 1

Type of user Description of fire use

Livestock farmers Fire lit towards end of the dry season to remove litter and stimulate growth of new grass.Other uses include controlling pests such as ticks and tsetse-fly.

Arable farmers To remove unwanted biomass while clearing lands

Hunters Hunters use fires to open up dense forest to ease their movement and detection(Also poachers) of wild animals.

Fishers In the Okavango Delta floodplain fires are set just before flooding to open the area for fishing and increase production of phytoplankton.

Safari companies Burning is done to attract wildlife to the new re-growth and hence increase chances for tourists to view and photograph game.

Road maintenance Fire is used to assist in clearing vegetation that obstructs visibility along roads and attract animals to roads resulting in accidents.

Smoking Accidental fires: These result from burning cigarette stumps and are visible along railways and major roads.

Domestic uses Fires resulting from use of fire for warming, cooking and for security by hunters, herders, and gatherers of veld products such as thatching grass. Some of these may start by the roadside where campsites are usually located.

Fires2 6/15/09 8:14 PM Page 11

12

studies focused specifically on Botswana.Understanding the ecological benefits of firewill enhance the benefit of fire in land use man-agement. For example, fire can kill emergingbush species, which helps to control bushencroachment and maintain open grasslandsthat are more appropriate for livestock produc-tion. Strategic application of fire could provevery useful for livestock managers in Botswanaunder future climate change when woody veg-etation will be more supported due to a combi-nation of warmer and drier climate conditionsand elevated carbon dioxide (Bond andArchibald, 2003). Others believe that the cur-rent widespread bush encroachment in thecountry is partly due to fire suppression strate-gies that have been applied since the colonialperiod (Sporton and Thomas, 2002).

The problem associated with using fire as amanagement tool is that most of the uses occurover the dry season when dry flammable fuelload is abundant and weather conditionsfavour quick spreading of fire. As a result it iscommon for such fires to go out of control andextend beyond the original area planned for.Frequent occurrence of fires result in severalnegative ecological and economic conse-quences. People may die or get injured from thefire, lose property or they may be displacedfrom their homes (Dipholo, 1985). For the poorwithout insurance, loss of property due to firescontributes directly to poverty. Further, duringan outbreak of fire many government officialsand equipment are diverted from their normalduties to assist with suppression.

Graph 2, shows large losses in small stocksuch as goats and this is a concern becausethese normally contribute significantly to thewell-being of the family, providing milk andmeat, and are used to generate income for otherhousehold needs. Women usually have greateraccess to goats hence the important role of goatsin the well being of the family. Because smallstock is important for the poor households, fireshave a potential to impoverish further thosewho are already poor.

The estimated losses shown in Graph 2 donot represent the full extent of the loss causedby fire, leaving out such resources as timber inForestry Reserves, pasture and veld products incommunal areas. For wildlife, fire results inchanges in vegetation structure, which givesrise to changes in wildlife habitat structure.Predators such as lions, leopards, and cheetahsare attracted to burnt areas because of the avail-ability and vulnerability of prey attracted toemerging green fodder in burnt areas but withno hiding places. A burning fire coincidingwith animal migration periods will change theannual migration pattern of wild animals withimplications on their survival as their migrationroutes will be impassable during the fires.However, Wright and Bailey (1982) have notedthat, death of large animals directly due to fireis rare. Death is usually due to suffocation andprimarily affects species with small homeranges (Chew et al., 1958).

There are a number of Community BasedNatural Resources Management (CBNRM)projects in the country that rely on wildlife togenerate income (Botswana National Agenda21 Coordinating Committee, 2002).

Fires that occur in the early part of the dryseason (June/July) result in cattle losing feedfor the whole dry season and if rains are late ordo not come, this will have a devastating long-term effect on farmers. After losing grazingpasture to fire, farmers are forced to divertresources to either buying livestock feed ormoving animals to other places where they maycause congestion resulting in long-term loss ofproductivity due to land degradation process-es. In a semi-arid climate post burn recovery isgenerally slow, and if rains fail, burnt areas areleft exposed to agents of soil erosion such asheat and wind for long periods. Some of the offsite consequences of veld fires include sedimen-tation of streams and rivers and reduction ofthe capacity of water reservoirs. However,research is required to assess how fire con-tributes to land degradation in Botswana rela-tive to other activities such as overgrazing(Dube and Kwerepe, 2000).

Evolution of Fire Management There are no studies on the history of fire man-agement in Botswana. However, from generalhistory traditional institutions and culture playeda significant role (Tlou, 1985). Most of the currentuses of fire were practiced in the past but the roleof fire for security purposes, communication,hunting and disease control was probably moreimportant than is the case now.

The Basarwa used fire, as part of hunting butit is not known exactly when fire was lit andwhat precautions were taken before burning.Basarwa lived in small nomadic groups, and asa result it is less likely that they neither engaged

Fire Damage Botswana 1991-2001 Graph 2

Source: Ministry Of Environment Wildlife and Tourism - ARB (2003)

vehi

cles

Cattle

Goa

ts

Peop

le

Fiel

ds

House

s

Casualties/Victims

Horse

sDon

keys

Shee

p

wildlife

Ranch

es

Fires2 6/15/09 8:14 PM Page 12

in organised fire suppression expeditions norwas this necessary for their lifestyle (Dube,1992).

For the pastoral Batswana ethnic groups liv-ing in large communities the Chief (Kgosi), wasthe head of the group and he/she managedresources on behalf of the community with theaid of his/her uncles or juniors (Dikgosana), andwith labour organised in the form of regiments(Mephato) (Schapera, 1951; Dube, 1995). All fireshad to be reported to the chief so as not to mis-take them for communication fires. Where a firewas a threat to property and people, Mephatomust have been engaged to suppress the fire aswas the case with other dangers such as anattack by wildlife or another ethnic group.Judging from other activities such as settingdates for ploughing, the Chief with the aid ofDikgosana had a role on the use of fire in termsof permission to burn, where and when to burnfor common land uses such as hunting andregeneration of pasture. There were measurestaken against those failing to comply (Schapera,1951). The chief also had the power to punishthose who refused to help in fire fighting. As aresult it can be inferred from the historical peri-od that:� Since the population was sparse, large fires

were supported as fuel loads were generallyhigh. It must have been common for firesthat were not a danger to people and proper-ty to be left to burn until natural barrierssuch as rivers or rain stopped them as largeareas were unoccupied and there were limit-ed resources to handle large fierce fires. It islikely that people and wildlife died fromsome of these fires.

� The frequency of fire was low since the pop-ulation was sparse and there was a coordi-nated system on when to burn for varioususes resulting in a more predictable fire pat-tern. Fires due to lightning dominated someareas.

� Cooperation to suppress fires was achieved,as orders from the chief were obeyed.The colonial period leading to the introduc-

tion of western model systems of governancemarked the beginning of a process that eventu-ally weakened chieftainship and changedlifestyles of groups such as the Basarwa result-ing in different fire regimes. Colonial recordsshow that a fire suppression approach wasmore favoured primarily due to the unpre-dictable rainfall (Zimmerman, 1985). However,Chiefs continued to have a role in mobilisinglabour to suppress burning fires.

The introduction of technologies to drillboreholes and improvements in road transportfacilitated the expansion of the cattle industryinto the Kalahari sandveld and this increasedthe frequency of fires (Dube, 1992). Due toproblems of accessibility such fires burnt for

days. Expansion into the sandveld and othersocio-economic changes introduced dispersedsettlements making it difficult for the Chiefs tomaintain coordinated management and use ofland resources including management of fire(Dube, 1995). These changes gave birth to dif-ferent fire regimes as different users set fire fortheir own needs and at their own time withoutconsulting other parties. Such practices resultedin conflicting periods of burning and change inthe timing of fires, a problem that is widespreadtoday.

The removal of the administration of landfrom the Chiefs after independence and theintroduction of different land tenure systemssuch as communal land, National Parks, Gameand Forestry Reserves, freehold and state landfurther weakened the role of traditional insti-tutions in fire management and also changedfire patterns in the country. In general, in com-munal areas there are no clear individualresponsibilities to land resources and as aresult uncoordinated burning has continuedand there is growing reluctance to assist in firesuppression. Fire is prohibited in protectedareas while in freehold areas the owners are atliberty to burn as planned. However, fireshave no respect for boundaries, and as a resultfires from communal areas often spread intoprotected areas and into freehold land, wherelarge fuel exists resulting in greater damage. Itis partly against this background that theAgricultural Resources Conservation Act of1974, which provides for the conservation ofall natural resources in Botswana from anydamage, be it natural (fire, floods) or artificialwas put in place. The same Act also providedfor the establishment of the AgriculturalResources Board to implement provisions ofthe Act.

Later the Herbage Preservation Act alsoknown as Fire Prevention Act of 1977 wasenacted and this provides for regulations gov-erning the use and control of fires as follows:� The issuing of burning permits to carry out

prescribed fires.� The power to request assistance in putting

out fires.� The protection of life, person or property by

counter-firing.� Construction of fire breaks to reduce wild-

fires and their impacts.The division of the ARB, which recently,

moved to the Ministry of Environment, Wildlifeand Tourism is the custodian of this act. WhileARB has the overall responsibility to managefire, the Department of Wildlife and NationalParks has a similar mandate but restricted toprotected areas such as National Parks andGame Reserves. This can be a problem as mostfires in protected areas spread from the commu-nal lands adjacent to these areas.

13

Fires2 6/15/09 8:14 PM Page 13

14

Current Status of Fire Monitoring andManagement

Fire PreventionARB has applied different approaches over theyears to implement the Fire Prevention Act of1977, including: Firebreaks: ARB has over the past yearsembarked on the construction of firebreaksthroughout the country. A network of firebreaksup to a total of 10 000 km has been constructedcountrywide and is repaired every year toreduce the risk of spreading fires (ARBSecretary’s Report, 2000).

The Agricultural Resources ConservationAct of 1974 requires construction and mainte-nance of firebreaks. As a result the ARB can besued for failure to maintain firebreaks. In somecases private companies have been engaged toconstruct firebreaks (ARB Secretary’s Report,2000).

The construction and maintenance of fire-breaks has proved to be very expensive. By theend of 2003, only 45 per cent of the total 10 000km firebreaks were maintained due to financialconstraints and lack of equipment.Furthermore, there are no indications that fire-breaks have helped to reduce fire outbreaksjudging by the incidents of fire noted above.However, firebreaks have reduced cases wherefires from one land tenure system spreads toanother, in particular from communal areas toother land tenure such as Game Reserves.

Public education: ARB also provides educationthrough public gatherings, radio, national tele-vision, and other educational materials in orderto enlighten people about wildfires. Signboardsbearing messages such as “Se Tshube Naga”and “Somarela Matlotlo A Tlholego”Interpreted As “Do Not Burn the Forest” and“Conserve Natural Resources” are commonthroughout the country. In additionConservation Committees are formed in differ-ent villages to assist in engaging communitiesin natural resource conservation and control offire. It is not clear if these education campaignshave changed attitudes on fire.

It has been shown in Ghanzi in the westernpart of the country that communities did nottake ARB education efforts seriously and theyfelt that the campaigns were inadequate(Mosweu, 2004). There is need to evaluate theeducation strategies used including the mes-sage signboards to formulate more effectiveapproaches.

Under the Fire Act it is an offense punishableby law to cause a fire and anybody using fireneeds to seek permission and also informneighbours of their intention before doing so.Most people find it cumbersome to apply for afire permit and there is no information to indi-

cate how people may apply for fire permits,and how long it takes to get one. As such theARB finds it difficult to take measures againstthose breaking laws.

First there are no resources for policing andsecondly the public is reluctant to come for-ward with information on fire outbreaks. Sincefire is widely used, individuals realise thatthere is always a chance that they too couldcause an uncontrolled fire by accident and getreported.

As a result most fires are never reported andtherefore, suspects are never arrested and it isdifficult to analyse trends in causes of fire. Thisis one of the negative consequences of a strin-gent fire suppression approach. A more partici-patory approach to fire management needs tobe considered.

Fire MonitoringFire detection relies on patrols by the ARB, butthis Department is generally understaffed andpoorly resourced. Reports by villagers, civil ser-vants, lodge owners as well as small aircraftSafari pilots complement the ARB effort.Members of the public are likely to be the firstto detect a fire.

As noted above there is general unwilling-ness by members of the public to report fire out-breaks as people fear that they may end uphelping with investigation on causes of the fire.However, in other cases failure to report is dueto ignorance on the wider impacts of fire on theenvironment and on the productivity of naturalresources, which sustain their livelihoods.There is generally a tendency not to be con-cerned if the fire is not a direct threat to one’sproperty.

Another form of detection is by use of firetowers, which is common in Chobe District.Deploying field assistants to work on fire look-out towers is costly as this is a 24-hour task dur-ing the fire season. Delayed deployment resultsin greater expenses because if a fire escapes itsinitial stages it becomes difficult to suppress. Atpresent the posts are under utilised.

The current fire detection methods used inBotswana are therefore not systematic and aregenerally unreliable (Monageng, 1985). Further,where a fire is detected it is not always possibleto deploy fire control forces early enough tosuppress it before it spreads over large areas.

In order to improve fire monitoring, attemptshave been made to incorporate informationfrom satellite data.

Initial attempts in the late 1980s used hard-copies of Landsat MSS data to map fire scarsbut this was constrained by the cost of the data.

In 1996 the Department of MeteorologicalServices (DMS) acquired a PC-based NOAAreceiver to provide data from the NOAAAdvanced Very High Resolution Radiometer

Fires2 6/15/09 8:14 PM Page 14

(AVHRR) sensor. AVHRR data was used tomap fire scars and vegetation cover. For earlywarning purposes DMS issued two reportseach day – the Pilot Active Fire Action Report(PAFAR) and the Active Fire Impact Report(AFIR). The objective of the PAFAR was toalert fire management agencies (ARB) of possi-ble fire, with a target time of 1100 hrs each day.The AFIR was issued with a target time of 1600hrs each day but in contrast to PAFAR alsoincluded information on communities, landuse and infrastructure that could have beenaffected by the fires. These products were eval-uated under the Range Inventory andMonitoring Project (BRIMP) carried out by theRange Ecology Division of the Ministry ofAgriculture in collaboration with ARB (Graph2) (BRIMP, 1998).

However, the use of NOAA AVHRR data infire monitoring was short lived due high inci-dents of false fire and fire scar detection.Limited skills in image processing also com-pounded this problem (Flasse, 1998).Nevertheless the development of theMODerate Resolution, ImagingSpectoradiometre (MODIS) that has special firechannels designed to saturate at about 500Kand 400K, respectively has significantlyreduced the potential of false fire detectionexperienced with AVHRR data.

Botswana participated in the validation ofthe MODIS fire products under the Universityof Maryland and NASA project facilitatedthrough SAFNet (Roy et al, 2005). MODISactive fire products are available in the publicdomain on a website hosted by the GeographyDepartment of the University of Maryland,USA (Website: http://maps.geog.umd.edu).However, these products remain inaccessiblefor use in operational fire monitoring in thecountry due to a number of constraints some ofwhich include, insufficient Internet services toaccess the data and limited human resources inARB.

Towards Reducing Vulnerability to Fire Outbreaks In order to effectively control fire there is needto take into account different land use needs forfire so as to incorporate these into fire controlstrategies. This will improve coordination ofburning by different users and facilitate focusededucation on how to prevent unplanned fireoutbreaks.

Education on preventing fire outbreaksneeds to be carried out through existing com-munity based organisations and NGOs that arealready dealing with natural resources. As aresult there is need to review the current man-agement approaches to clearly reflect the fireneeds of different land uses because fire is inte-gral to maintaining sustainable rural liveli-

hoods while at the same time land use is themajor source of unwanted fires.

A national fire danger rating system shouldalso be developed and communicated to landuse managers. Information on the major landsystems and land use types of Botswana isavailable as baseline data from the Departmentof Meteorological Services (DMS) to assess firerisks relative to rainfall variability and otherweather parameters.

Annual change in fuel load can be providedfrom satellite data products such as NOAAAVHRR, MODIS and more recently theMeteosat Second Generation data which is alsobeing received by DMS. However, resources toaccess, process and use these different datatypes for fire early warning needs to be put inplace in different institutions concerned withfire management. Without this structure inoperational form the country remains vulnera-ble to fires particularly after very wet periods.

Fire management should incorporate theecological role of fire. This requires informationon the interactions between fire and differentvegetation types found in the country. This willhelp to determine if current fire frequencies area danger to the ecosystem.

Such knowledge is important for planningprescribed burning. So far there are limitedstudies focusing on the ecological role of fire inthe country and there is limited prescribedburning practiced. Total exclusion of fire partic-ularly in protected or sparsely populated areaspromotes conditions for increased vulnerabilityto fire danger resulting from accumulation offuel load.

Archiving and easy retrieval of informationto assess fire history and trends is an importantpart of effective fire management. All satelliteproducts need to be validated and the availabil-ity of both current and past ground based firerecords is important in this process. Currentlythere is no fire database in Botswana. Lack ofhistorical records constrains attempts to con-struct fire frequency records, to make predic-tions and to assess future vulnerability to fire.

On a regional scale fires have a role in green-house gas emissions. Botswana is a signatory ofthe United Nations Framework Convention onClimate Change and is therefore expected toprovide a national inventory of sources andsinks of greenhouse gases. So far it is difficultto accomplish this requirement for fire becauseof the limitations noted above on fire monitor-ing. An ineffective fire management systemincreases the vulnerability of the country to cli-mate change in future as studies have shownthat fierce fires are likely to occur in future fol-lowing very wet periods (Rutherford et al.,1999; Dube, 2003).

Botswana has taken a lead in building theSouthern Africa Fire Network (SAFNet) prima-

15

Fires2 6/15/09 8:14 PM Page 15

16

rily aimed at providing motivation to addressthe fire problem. SAFNet facilitates more expo-sure to the use of geo-spatial data in fire man-agement, issues of cross border fires, debates onengaging communities in fire control andappropriate institutional frameworks and firepolicy requirements.

ConclusionAlthough the frequency of fire in Botswana islow compared to areas of higher rainfall such asAngola and the Democratic Republic of Congo,the country remains vulnerable to fire out-breaks. This is because rainfall is unpredictableand a large section of the population dependson resources that rely on rainfall. As a result,loss of these resources due to fire is a concern,as supply may not be realised after a long timedue to drought.

Fire occurs mostly over the dry season whenthe demand for most natural resources is veryhigh. However, vulnerability to fire is partlylinked to the fact that fire is part of land use

management and as a result fire outbreaks areinevitable, particularly after wet periods.Added to this, is that although the country hasa Fire Prevention Act, the implementation ofthis Act has not been effective. For instancethere is no fire early warning system and themanagement of fire is not guided by scientificunderstanding of the role of fire in ecologicalprocesses and in land use management.

The use of geo-spatial technology could helpimprove fire management but the institutionalcapacity to incorporate this data operationallyis limited. This is a major concern as initial cli-mate change assessment shows that the countrywill experience more fire incidents in future.

The most vulnerable groups are the poorwho rely on veld products for food and incomegeneration. However, fire is also a threat to thecountry’s livestock and tourism industries.There is need for studies to provide informationon the cost and benefit of fires to assist inincreasing awareness at all levels on the signifi-cance of fire in sustainable development.

Fires2 6/15/09 8:14 PM Page 16

References

ARB, (1981). Annual Report. Agricultural resource Board, Ministry of Agriculture, Gaborone, Botswana.

Agee James K., (1993). Fire Ecology of Pacific Northwest Forests. Island Press, Washington, D. C., and Covelo,California. U.S.A. pp. 113- 171.

Bond W. J. and Archibald, S. (2003). Confronting complexity: fire policy choices in South African savannaparks. International Journal of Wildland Fire. 12:381- 389.

Bond, W. J., Geldenhuys, C. J., Everson, T. M., Everson, C. S. and Calvin M. C., (2004). Fire ecology:Characteristics of some important biomes of sub-Sahara Africa. In J. G. Goldammer and Cornelis de Ronde,(eds) Wildland fire management handbook for sub-Sahara Africa. Global Fire Monitoring Centre (GFMC). pp. 11-21.

Botswana National Agenda 21 Coordinating Committee. 2002. Botswana National Report on the implementa-tion of Agenda 21 and other Rio Earth Summit decisions. Gaborone. p. 277.

Buchanan-Smith M., (1998). Food and nutrition insecurity: The impact of drought. In M. Mugabe, K.Gobotswang and G. Holmboe-Ottesen (eds). From food security to nutrition security in Botswana. NIR/IASAMcollaborative Research Programme on Health. Population and Development. UB. Gaborone. pp. 233-245.

Central Statistics Office., (2000). Environmental Statistics. Department of Printing and Publishing Services.Gaborone. pp. 100-178

Chew, R.M., B.B. Butterworth, and R. Grechman. (1958). The Effects of Fire on Fish and Wildlife. Oregon StateUniversity Press, Corvallis, Oregon, U.S.A. p. 233

Department of Wildlife and National Parks, (2003). Dry Season Aerial Census. Gaborone

Ditlhogo, M. K. (1996). The Natural History of Imbrasia Belina westwood (Lepidoptera;saturniidae) and somefactors affecting its abundance in North Eastern Botswana. PhD thesis. Unpublished. University of Manitoba.P. 300.

Dube, O. P., (2003). Impacts of climate change, vulnerability and adaptation options: Exploring the case forBotswana through Southern Africa: A review. Botswana notes and records 35:147-168.

Dube, O. P. and Kwerepe, R., (2000). Human induced change in the Kgalagadi Sands: beyond the year 2000.In Ringrose S. and Chanda R. (Eds). Towards sustainable management in the Kalahari region - Some essential back-ground and critical issues. Directorate of Research and Development, University of Botswana. Gaborone. pp.244-258.

Dube O. P. (1992a). Women and Environmental Management in Botswana. Population and Environment in Africa.Conference proceedings. The Union for African Population Studies. September 1992. Gaborone. pp. 462-475.

Dube O. P., (1992b). Notes on Ecology and Society in Southern Africa. In A. Erele, (ed), African Education andIdentity. HansZell, London. pp. 358-368

Dube O.P. (1995). Monitoring vegetation and soil degradation in Botswana. Progress Report. AustralianCenter for International Agricultural research. ACIAR Project No. S010001. p52.

Dipholo, K. P D. (1985). Steps being taken to control the loss of resources caused by bush fires. In Proceedingsof the 2nd Conservation Seminar for District Authorities. Department of Wildlife and National Parks, Maun.

Flasse, S. (1998). Quick Assessment of semi-automatic AVHRR burn scar detection for Botswana, prototypeversion 1.0. Technical Report for DFID-BRIMP. Internal draft. NRI, The University of Greenwich.

Frost, P. G. H. (1998). Community-based management of fire: lessons from the western province in Zambia.A paper prepared for a meeting on Public Policies Affecting Forest fires, FAO, Rome, 1998. p. 13.

Heinl M. (2004). Reconstructing the fire history of the Okavango Delta, Botswana. Final Report.Conservation International Botswana, Okavango Program. p12.

Hulme, M. (1996). Climate change and Southern Africa: an exploration of some potential impacts and impli-cations for the SADC region. Report commissioned by WWF. Climate Research Unit, University of EastAnglia, Norwich, U.K. p. 104.

Kalabamu, F. T. (1994). Urban challenge and development policies: Lessons from Botswana. InternationalConference on Growth Management for cities in crisis. Cape Town, South Africa. 8-9 September. pp. 21

Kaufman Y. and Justice C. O. (1998). Algorithm technical background document. MODIS Fire Products.Version 2.2 EOS ID 2741. MODIS ATBD:Fires. http://modis.gsfc.nasa.gov/MODIS/ATBD/atbd_mod14.pdf.

Letsie, L. (1996). A gendered Socio-economic Study of Phane. Proceedings of the First Multidisciplinary Symposium onPhane, 18 June. Department of Biological Sciences, University of Botswana, Gaborone.

Ministry Of Environment Wildlife And Tourism - ARB (2003). Report for the Early Warning TechnicalCommittee Meeting. Ministry Of Finance Development Planning, 11th December 2003. Gaborone. p. 5.

McMahon, T.E., and D.S. de Calesta, (1990). Effects of Fire on the small mammal populations of chaparral. OregonState University Press, Corvallis, Oregon, U.S.A. pp. 233.

17

Fires2 6/15/09 8:14 PM Page 17

18

Ministry of Works, Transport and Communications, (2001). Botswana Initial National Communication to theUnited Nations Framework Convention on Climate Change. Bay Publishing Pty Ltd., Gaborone. p. 96

Monageng, G. G. (1985). Fires in forestry areas of northern Botswana. In Proceedings of the 2nd ConservationSeminar for District Authorities. Department of Wildlife and National Parks, Maun.

Mosweu, T. (2004). Causes and socio-economic impacts of veld fires in Ghanzi. Undergraduate dissertation-unpublished. University of Botswana. p. 38.

Ntabeni, T. N., (1999). Fire potential modeling using NOAA AVHRR data and other spatial data of Botswana.Msc dissertation, unpublished. University of Greenwich.

Pyne Stephen J., (1984). Introduction to Wildland Fire-Fire Management in United States. A Wiley-IntersciencePublication, New York, U.S.A. pp 225- 310

Republic of Botswana, (2003). National development plan 9 2003/04-2008/9. Ministry of Finance and devel-opment Planning, Gaborone. p. 427.

Roy, D., Frost, P., Justice, C., Landmann, T., Le Roux, J., Gumbo, K., Makungwa, S., Dunham, K., Du Toit, R.,Mhwandagara, K., Zacarias, A, Tacheba, B., Dube, O. P, Pereira, J., Mushove, P., Morisette, J., SanthanaVannan, S., Davies, D., (2005). The Southern Africa Fire Network (SAFNet) regional burned area product vali-dation protocol. International Journal of Remote Sensing Vol. 26 (19): 4265–4292.

Rutherford, M.C., Midgley, G.F., Bond, W.J., Powrie, L.W., Roberts, R., Allsopp, J., (1999). South AfricanCountry Study on Climate Change. Plant Biodiversity: Vulnerability and Adaptation Assessment. RefereedVersion. Department of Environmental Affairs and Tourism. Pretoria, p. 34.

Schapera, I. (1951). The ethnic composition of Tswana tribes. Monographs on Social Anthropology No. 11.The London School of Economics and Political Science. London. p133.

Sekhwela, M.B.M., (1994). Grapple Plant (Harpagophytum procumbens, DC) Resource potential and manage-ment studies. Final Report. National institute of development Research and documentation. University ofBotswana. p. 89.

Sporton, D. and Thomas, D.S.G. (eds). (2002). Sustainable livelihoods in the Kalahari environments. A contributionto global debates. Oxford University Press. , Oxford. , p. 231

Thomas, D. S. G. and Shaw, P. A., (1991). The Kalahari environment. Cambridge University Press, London

Tlotlego, S. (2004). Remote Sensing and GIS and veld fire management: Exploring possibilities for an inter-face for the rural communities. Undergraduate dissertation-unpublished. University of Botswana. p. 62.

Tlou, T. (1985). A history of Ngamiland – 1750-1906 – the formation of an African state. MacMillan,Botswana. p. 174.

Zimmermann, I. (1985). Are bush fires good or bad? An ecological appraisal. In Proceedings of the 2ndConservation Seminar for District Authorities. Department of Wildlife and National Parks, Maun. pp. 8 -10.

Fires2 6/15/09 8:14 PM Page 18

Johan le Roux1 and Rabanus Shoopala2

Abstract Namibia is a large semi arid country coveringa surface area of 824,392 sq km with a smallhuman population of 1.8 million people. Theseasonality of the rainfall provides dry plantmaterial that burns between April andNovember, with a peak during August andSeptember. Lower rainfall in the southern andwestern parts of the country means less fueland consequently fewer fires. However, suffi-cient biomass may accumulate over severalseasons to sustain large but infrequent fires.

The predominance of fires during the cloud-less dry season is indicative of anthropogenicfires that burn approximately five per cent or40,000 sq km of the country every year. Themost fire prone area is in the Kavango andCaprivi region where the rainfall is higher.Between 25 and 50 per cent of this area burnsevery year, possibly affecting as many as 280,000 people. Fires are lit for reasons which areclosely related to land use, including land clear-ing for cultivation, application of natural fer-tilisers, hunting and pasture rejuvenation.Frequent fires are seen as detrimental to theindigenous timber industry in particular, andmay also have direct and indirect adverseaffects on humans, livestock and wildlife. Localcommunities are actively involved in fire man-agement, and the process of developing nation-al fire policies is under way.

Socio-Economic CharacteristicsNamibia is a large, arid, sparsely populatedsouthern African country, bordering Angola,Zambia, Botswana, South Africa and Zimbabwe,and the Atlantic Ocean in the west. Human pop-ulation density is about 2 people per square kilo-metre, but settlement is not uniformly distributedacross the country (Mendelsohn et al. 2002). Lessthan 40 per cent of the country’s population of 1.8million people live in urban areas (Mendelsohnet al. 2002; Republic of Namibia, 2004; Universityof Namibia, 2005), while the majority of the ruralcommunities live near perennial rivers in thenorthern parts of the country. More than 50 percent of the population is under the age of 20 yearsand have yet to enter the labour market. Only

about 50 per cent of the country’s workforcereceives a cash income (African DevelopmentBank, 2002). Those who do not earn wages arenot necessarily jobless since many are involved insubsistence farming at or near their homesteads.

Climate, Land Cover and Land UseNamibia’s rainfall is largely derived from themoist Inter-Tropical Convergence Zone (ITCZ),which moves from the equator southwards.However, in some seasons the ITCZ is keptnorth by a dry cold air mass of the SubtropicalHigh Pressure Zone. Namibia has a pro-nounced rainy season from about November toMarch. The country has a steep rainfall gradientfrom the north-east to the south-west, withmedian annual figures of 550 mm and less than50 mm respectively (Mendelsohn et al. 2002).As expected, these rainfall patterns give rise toa land cover gradient from tall moist wood-lands in the north-east, to short dry scrub in thesouth-west of the country.

Land use in Namibia is dominated by agri-culture. About 80 per cent of the country isreserved for farming on either freehold or com-munally owned land. Eco-tourism and miningactivities are interspersed with farming in theseareas. About 14 per cent of the country consistsof state owned protected areas. Livestock farm-ing is the main agricultural pursuit, but largeareas are cleared for dryland (rain fed) crop cul-tivation, particularly in the northern, commu-nally owned areas (Mendelsohn et al. 2002).

FireNamibia is not fire-prone due to low biomassload. Much of the country receives too little rainto generate sufficient amounts of combustibleplant material. Nevertheless, in those areas thatreceive enough rain every year, fire is a regularand widespread phenomenon. The most fre-quent, intense and extensive fires occur in thenorth-east, while in the south and west fires arerare. The fire season is generally restricted tothe dry season (April to October), pointing topeople rather than lightning as the main sourceof ignition. Most of these fires spread throughthe grass and shrub layer as surface fires, whilecrown and subterranean (peat) fires are eitherrare or confined to small areas.

19

NAMIBIA

1 Etosha Ecological Institute, P.O. Box 6, Okaukuejo, Via Outjo, Namibia. Corresponding author, e-mail:leroux@mweb.com.na

2 National Remote Sensing Centre, Directorate of Forestry, Windhoek, Namibia.

Fires2 6/15/09 8:14 PM Page 19

20

The issue of fire in Namibia can be discussedin relation to the four broad geographic regionsshown in Map 4 below.

The coloured areas represent burned areas fromfive fire seasons. Adapted from Trigg and leRoux, 2001.

Region 1 is situated in the densely populat-ed north-central part of the country. The vege-tation consists mainly of mixed grassland andsavanna woodland. The land use is predomi-nantly small-scale farming on communallyowned land where fires are ignited for landclearing and for grazing rejuvenation. Region 2also in north-central but located immediatelysouth of Region 1, comprises the EtoshaNational Park, an area of more than 22,000 sqkm where large areas are burned annually aspart of the Park’s fire management strategy.Region 3 extends immediately south of Region2 but covers a large area up to the borders ofBotswana and South Africa. Region 3 consistsof sparsely populated freehold farmland wherethe vegetation is predominantly mixed treeand shrub savanna. In this region every effortis made to prevent and combat fires. Region 4is situated in the high rainfall north-easternpart of the country and covers all of the Capriviand Kavango regions, as well as the northernpart of the Otjozondjupa region where the veg-etation is a mixture of grassland, savannawoodland and dry forest. This is the most fireprone area in the country, with fires being litmainly for land clearing, grazing rejuvenationand hunting.

Causes and Effects of FiresIn regions 1 and 4, fires are lit to stimulate thegrowth of fresh grass for cattle, to attract game,to flush out game that can then be hunted, toclear vegetation around waterholes and honey-gathering areas, to clear land for cultivation, orto promote the regeneration of grasses used forthatching. Large areas are sometimes burned

accidentally through the uncontrolled spread offires deliberately lit to clear new fields. Inregion 3 fires are a result of lightning strikes,and accidental due to poorly supervised char-coal making and careless discarding of ciga-rettes.

Landscape Characteristics andFire RegimesThe progression of fires during the 2003 burn-ing season is shown in Figure 2. The scarcity ofdetected fires in the images for April andNovember is due to a combination of cloudcover that block the satellite sensors’ view ofthe earth’s surface, and actual absence of fires atthat time. Fire occurrences peak in August andSeptember. Given that several fires occur oversuch large areas between the months of Mayand October, it is not surprising that there issuch a dramatic drop in active fires inNovember, as much of the combustible fuelwould have been consumed either by fire orlivestock.

Burnt areas that were mapped at the nation-al level for five years (Map 4), show that aboutfour million hectares burn every year. Thisamounts to a mere 4.8 per cent of Namibia’stotal area. The picture changes dramaticallyhowever, when one focuses on the north-east-ern part of Namibia where between 25 and 50per cent of the area burns every year. Forregions 1, 2 and 3 where rainfall is more vari-able and lower, fires usually occur after a wetyear. The unpredictability of rainfall has poten-tial to lead to land degradation processesbecause vegetation regeneration may not besupported following the burning due to lack ofmoisture.

The fact that the burning season peakstowards the second half of the dry season indi-cates that management fires that could reducethe fuel load are not widely used. This is per-haps not surprising in a semi-arid country likeNamibia, where livestock owners are oftenreluctant to burn fodder that will not bereplaced if the rains fail and consider the poten-tial effect of this on land degradation processes.The extent to which grazing acts as a substitutefor management fires is not known, but judgingby the size of the area that burns every year, thisform of fuel reduction is ineffective at regionalscales.

The frequent burning of large parts of north-western Namibia is responsible for damage tolarge areas of woodland, with a significantreduction in commercial and non-commercialtimber resources, loss of habitat, reduction inwoodland biodiversity, and loss of woodresources for domestic use. The high fire fre-quencies also promote the establishment of fire-resistant, but often less useful woody speciessuch as Dichrostachys cinerea. All these processes

Fire Regions in Namibia Map 4

Fires2 6/15/09 8:14 PM Page 20

21

Monthly Time Series of Map 5Active Fires

April

May

August

September

November

June

July Source http://maps.geog.umd.edu/

October

Fires2 6/15/09 8:14 PM Page 21

22

eventually contribute to land degradationwhich is a major environmental concern inNamibia.

Fires also kill large mammals, with peatground fires in the Malangalenga area ofCaprivi killing 170 buffalo in 1996 (Mendelsohnand Roberts, 1997). In exceptional cases, peopleare trapped and suffer fatal burns. During theheight of the dry season, large smoke plumesand high levels of background smoke are pres-ent. The effect that reduced air quality has onpublic health still remains to be quantified. Inmany areas where migratory pastoralism is nolonger practiced, the removal of pasture by fireis highly undesirable as it threatens the survivalof livestock. This situation is compounded dur-ing years when rains fail after burning.

A study by the National Remote SensingCentre (2002), found that even though most ofthe Kavango and Caprivi area burns at very reg-ular intervals as shown in figure 3, the effect onthe woodlands is possibly not as devastating aswas once believed. A comparison of field basedforest inventory results with satellite derivedburned area maps shows that the apparentdecrease in tree cover could not be attributed toobserved high fire frequencies. The study pointsout that although large trees might be on thedecline, smaller trees of 3 – 4 metres in height areon the increase. The overall result of the currentfire regime therefore favours the development ofyounger, denser woodland, at the expense ofmore mature wooded grassland.

Past Fire Uses in NamibiaAlthough the relationship between fire andland use has come to focus more on pasturesmanagement and clearing for agriculture, thereis evidence to show that there were other usesof fire among Namibian ethnic groups. Forexample, Herero and the Himba people havebeen practicing holy fire for centuries. TheHoly Fire serves as the medium for communi-cating with the ancestral spirits (Malan, 1995).All fires in the village are lit by the holy firewhich was kept by the village-head. Among theOvambo fire symbolised hope and luck for thenation. It gave the group existence, and gavethe Chief his power. If this fire got extin-guished, it meant the decay of the nation and italso foretold the Chief’s loss of power or hisdeath by witchcraft (Hiltunen, 1986). While inthe Kavango region, the Ovambukushu used fireto symbolise the sun and moon, as helpers ofNyambi (God) to guide and protect the peopleon earth. During the day this fire (sun) burnedfiercely, giving off immense heat, while it wasjust kept alive to give light at night (moon), andthis allowed people to rest (Hiltunen, 1986).Although not documented it is likely that underfavourable conditions such fires escaped andburnt large areas.

Another use of fire was in ritual ceremonies(Malan, 1995). The ritual fire was generated bymeans of ritual fire sticks called ‘ozondume’, ofwhich each lineage kept a special set. This fireconsisted of a single smouldering stump lyingin the ashes and it was only stirred into flamewhen important rituals are performed. The firewas kept burning between the entrances to themain hut and the cattle enclosure. Burning firesymbolised sustained contact between the liv-ing and their ancestors and provided spiritualsecurity. The practice of Holy Fires was discour-aged by missionaries and is no longer wide-spread (Malan, 1995). Because of the controllednature of these fires they are less likely to havespread beyond where they were lit.

The San people used fire for hunting purpos-es (Malan, 1995). The use of fire, cut the longdistances travelled for hunting as sproutinggrass brought game closer. The San people alsoused fire to encourage the sprouting of tubeplants which they consume as food. Such firesare likely to have burnt large areas.

The Ovambo, Kavango and Caprivian peo-ple continue to use burning to clear land foragricultural purposes. Fire is also used by localpeople in the region to control diseases causedby ticks and other parasites. Although fires areeasily started at cattle posts, these are often con-fined to small areas and they come after longintervals (fire returns) due to overgrazing andrainfall variability. According to Sheuyange(2002) the regions with adequate grass cover arethose subjected to heavy grazing pressure, apractice that leaves limited fuel load to supportburning fires.

During colonial times, the Soil ConservationAct No. 6 passed in 1949 prohibited the use of firein commercial farming areas of the country. Theintroduction of cattle farming in these areas andthe suppression of fires are regarded as some ofthe contributing factors towards bush encroach-ment (de Klerk, 2004). Bush encroachment is aphenomenon where bush thickets have replacedsavanna woodland characterised by tall trees ,scattered bushes and a carpet of pasture. Bushencroachment has a major impact on livestockfarming since it reduces the carrying capacity ofthe land (Mendelson and O’beid, 2005).

Fire Management

Evolution of Fire ManagementFire management is still evolving, primarily driv-en by Forestry officials with support from theNamibia/Finland Forestry Programme. The pro-gramme involves local communities in fire“management” through the establishment andmaintenance of firebreaks, and is aimed atstretching the fire return period in the forestreserves and other wooded areas of East Caprivi.

Fires2 6/15/09 8:14 PM Page 22

Current Policy StatusThe formulation of policy and regulations con-cerning fire is also an ongoing process. Some ofthe recent efforts include the National FirePolicy and Guidelines on Fire Management inNamibia developed by the Directorate ofForestry (Goldammer, 1998). The policy andguidelines build on the Namibia ForestryStrategic Plan, which provided the basis for firepolicy and management planning. TheStrategic plan recognises many of the complex-ities of fire, including:� The need to reduce the occurrence and sever-

ity of uncontrolled and accidental forest firewhile allowing controlled fire under specificcircumstances;

� The need for community participation in theprotection of forest resources; and

� The need for different policies for differentparts of the country.Other fire policy development efforts include

the Round Table Meeting on Fire held in 1999with the aim of defining the country’s fire poli-cy and coordinating responsibility for fire man-agement (Goldammer, 1999). Draft proposalsfor regulations on bush fire management havealso been prepared (Piepmeyer, 2000), and thecountry ratified the UN FrameworkConvention on Climate Change, which requiresit to provide quantitative assessments of free-burning vegetation fires (Goldammer, 1998).

Human Vulnerability to FireAccording to Kamminga (2001), the risk of wildfires is strongly related to the prevailing land usesystem and tenure. Traditional extensive systems

have built-in strategies such as migratory pas-toralism and crop cultivation, which reduce thecommunity’s vulnerability to fire and theabsence of rainfall. Intensive grazing is anothercommon if inadvertent means of reducing wild-fire risk by reducing fuel loads. The combinedeffect of trampling and grazing around settle-ments is very effective as a means of protectinghighly combustible wood and thatch houses.

About 280,000 people live in the fire pronenorth-western part of Namibia (Mendelsohn etal., 2002). While all of them are affected by theannual fires in some way, most live in a narrowstrip along the banks of the Okavango River,where a combination of dense settlement andheavy grazing has virtually eliminated fire. Inthe extreme eastern part, between the Zambeziand Chobe/Linyanti rivers, about 80,000 peo-ple live in much closer proximity to these fireevents. This amounts to approximately four percent of the country’s population.

The great majority of these fires are human-caused. It is therefore reasonable to expect thatif enough people were adversely and directlyaffected by fires, they would stop igniting them.Judging by the extent and frequency of theburning in this part of the country, the benefitsof the use of fire that has been going on for cen-turies, still outweighs the disadvantages.

ConclusionIt can be concluded that fire, although wide-spread, frequent and harmful to humans andthe environment, affects rather small propor-tions of the country and its human population,and not always in a detrimental way.

23

Fires2 6/15/09 8:14 PM Page 23

24

References

African Development Bank, (2002). Gender, Poverty and Environmental Indicators on African Countries 2002-2003,Vol. III. Statistics Division, Development Research Department. African Development Bank Group. pp 304.

De Klerk J. N. (2004). Bush Encroachment in Namibia. Ministry of Environment and Tourism, GovernmentRepublic of Namibia.

Goldammer, J. G., (1998). Development of National fire policy and guidelines on fire management inNamibia. Report, Directorate of Forestry, Ministry of Environment and Tourism, Windhoek.

Goldammer, J. G., (1999). Namibia Round Table Meeting on Fire, Windhoek 10-11 November 1998. Report,Directorate of Forestry, Ministry of Environment and Tourism, Windhoek.

Hiltunen, M. (1986). Witchcraft and Sorcery in Owambo. The Finnish Anthropological Society. Helsinki.

Kamminga, E. M., (2001). Impact of the Integrated Forest Fire Management Program on Rural Livelihoods inEast Caprivi Region, Namibia. Report, Directorate of Forestry, Ministry of Environment and Tourism,Windhoek.

Malan, J. S. (1995). People of Namibia. Rhino Publishers. University of the North.

Mendelsohn, J. and, Roberts, C., (1997). An Environmental Profile and Atlas of Caprivi. Directorate ofEnvironmental Affairs, Windhoek.

Mendelsohn, J. C., Jarvis, A., Roberts, C. and Robertson, T., (2002). Atlas of Namibia: A Portrait of the Land andits People. Published for the Ministry of Environment & Tourism, Namibia, David Philip Publishers, CapeTown.

Mendelson, J. and O’beid, S. (2005). Forests and Woodlands of Namibia. Directorate of Forestry, Ministry ofAgriculture Water and Forestry, Government Republic of Namibia.

Republic of Namibia, (2004). Namibia’s Third National Report On The Implementation Of The UnConvention To Combat Desertification. Desert Research Foundation of Namibia, Windhoek. p. 14

National Remote Sensing Centre. (2002). A local operational tool for fire monitoring and management for theKavango and Caprivi Regions. Final report, engineering and consulting contract No. Nam/328-02-132-A.Directorate of Forestry, Ministry of Environment and Tourism, Windhoek.

NRSC. (2002). A local operational tool for fire monitoring and management for the Kavango and Caprivi Regions.National Remote Sensing Centre (NRSC), Ministry of Environment and Tourism, Directorate of Forestry.Windhoek

Piepmeyer, W., (2000). Draft Proposals for regulations on bush fire management. Report, Directorate ofForestry, Ministry of Environment and Tourism, Windhoek.

Sheuyange, A. (2002). Landscape level vegetation change in relation to fire history in eastern Ohangwena region,Namibia. Noragric MSc. Thesis. NLH Agricultural University of Norway. Norway.

Trigg, S.N. and le Roux, J. L. (2001). Fire Situation in Namibia: Fire Environment, Fire Regime and EcologicalRole. In International Forest Fire News No. 25 - July 2001.

University of Namibia, (2005). Beyond Inequalities. Women in Namibia. Southern African Research and docu-mentation centre. SARDC Information Resource Centre. Virtual Libraryhttp://databases.sardc.net/books/namibia2005/index.php- consulted August 2006

Fires2 6/15/09 8:14 PM Page 24

Cheikh Mbow1 , Thomas Theis Nielsen2,

AbstractWest African savannas are characterised by per-manent burning associated with several landmanagement and natural resource extractionactivities. The use of fire is seen as a necessarytool in agro-pastoral activities, national parksand game reserve management, and in forestproduct collection. Several fires occur annuallyin the savanna grasslands where a continuousgrass layer prompts extensive burning everydry season. Uncontrolled burning is common,mainly when the herbaceous layer is very dry.

There are both biophysical and socio-eco-nomic negative impacts of fire in West Africa,including loss of biodiversity, vegetation frag-mentation, loss of soil fertility, degradation offodder, and reduction of rural incomes. Thereare also some positive impacts of fire such asthe promotion of herbaceous re-growth for live-stock, pest control, release of organic matter forsoil fertility, and the opening up of land for cropfarming. As such several attempts that havebeen made in the past to ban or mitigate fireshave not been successful. In the colonial era thecommonest management strategy was to banfire mainly in the forest reserves.

The failure of the total ban along with posi-tive scientific outputs on the use of fires drew anacceptance of early fires to achieve some man-agement objectives. During the cyclic drought ofthe seventies and eighties, the West African gov-ernments adopted an extensive strategy for firefighting. The lack of sustainability of thisapproach led to a shift from coercive to persua-sive strategies. The new fire managementregime gives responsibility to local communitiesto manage their natural resources. In order tosupport the new fire management approaches,which also include early warning systems, satel-lite data are required to document the temporaland spatial distribution of burning.

This case study is a summary of the state ofbushfires in West Africa, focusing mainly on thehistorical background of fire, the main causes offire, the fire regime in relation to landscapediversity, the timing of fire, the impact of fire,and fire management options.

IntroductionBush fires have been a common occurrence inWest Africa since pre-historic times(Goldammer and de Ronde, 2004). Althoughthe overall consequences as well as the extent ofthe damage by fire are difficult to assess, it isgenerally agreed that the propagation andoccurrence of bush fires is closely associatedwith several triggers linking the bio-physicaland social environments.

Biomass production through the rainy sea-son limits the extent of fires since fuel-loadshave to be high enough for fires to spread(Nielsen and Rasmussen, 2001). In addition, themoisture content of the biomass strongly influ-ences the propagation of fire (Saarnak et al,2003; Mbow, 2000).

Fires are used for religious purposes by somecommunities. The same communities are alsodependent on the use of fire to clear land for agri-culture or to promote fodder production for theircattle. The majority of fires occurring in WestAfrican savannas are therefore human-causedand occur as part of an integral land and naturalresource management regime (Goldammer, 1993;Schmitz, 1996; Hough, 1993).

Whereas fires have been endemic in thesavanna regions, acting as a major factor shap-ing ecosystem productivity and functionality, itis clear that without regular, targeted fires, thesavanna environment will have denser vegeta-tion preventing the development of major agro-pastoral activities (Orgle, 1998).

Due to the high fire frequency in West Africa,several attempts to ban or fight fires have beenintroduced in the past with little success. Thefire mitigation strategy should be orientedtowards changing rural practices rather thaninvesting in substantial logistics and heavyequipment to fight fires, as has been the caseduring the last 30 years.

Landscape Characteristics and Fire RegimesMaps 6 and 7, derived from MODIS and NOAAAVHRR respectively, reveal a fairly consistentpicture of the location of fires across WestAfrica. To the north, a low fuel load limits theoccurrence of fires, while in the south fires arelimited by humid conditions.

25

SENEGAL

1 Institut des Sciences de l’Environnement Faculté des Sciences et Techniques, Université Cheikh Anta Diop de Dakar Sénégal.2 Institute of Geography, University of Copenhagen, Denmark.

Fires2 6/15/09 8:14 PM Page 25

26

Use of Bush Fires in West Africa Table 2

Source of burning Reasons for burning

Agricultural Shifting cultivation, field preparation, fertility improvement, residues removal, pest control

Forest management Early fires to minimise late burning, forest accessibility, improvement of animal visibility for tourists in national parks, control of bush encroachment, stimulate the flowering of some plants

Rangeland management Early burning for grass re-growth, counter season fire to remove ‘bad grass’

Forest product extraction Honey collection, hunting game, gum extraction, charcoal making, firewood collection, fruit collection

Crime Elimination of grass to keep away nomads, avoid clashes between communities, track livestock raiders

Accidents Non extinguished firewood, cigarettes, transportation of embers Other sources Clearing around villages to prevent house-burning, clearing along rail and roads to

prevent animal accidents, traditional festivities in some ethnic groups, etc.

Source: Mbow et al. (2000)

Fire Frequency in West Africa Map 6(dry season 2003-2004)

Inter Annual Rainfall, Map 81961 to 1990

Fire Frequency in Senegal Map 7(dry season 2003-2004)

Senegal: Isohyets of of the Map 9Rainy Season 2003

Source: MODIS data

Source: MODIS data

Source: Météo-Nationale, Senegal

Source: SDRN-FAO, Roma

Fires2 6/15/09 8:14 PM Page 26

The temporal distribution of fires can beassessed by the same methods. In Map 6 andFigure 3 the total number of detected fires in thesame data window Map 6 and 7 is illustrated asa function of the time of year. A pronouncedfire-season corresponding to the onset of thedry season can be identified. Studies haveshown that the fire season in West Africa sets into the north as soon as the biomass dries up andmoves southwards along a precipitation gradi-ent (Nielsen and Rasmussen, 1997).

More detailed information can be obtainedfrom the same data sources by analysing thedetected fires on a more regional scale. In the pres-ent case, Senegal will be used as an example.Figure 2 illustrates the spatial density of detectedfires. Here it is clearly seen that while the northernmost regions of Senegal are seemingly unaffectedby fires in the fire season of 2003-2004, a high den-sity of fires is found in the south of the country.While these results were produced on the basis ofMODIS data they still illustrate the results pro-duced for Senegal for previous fire seasons basedon data from the NOAA AVHRR sensor.

The key aspects of current fire trends are gen-erally the same in the Sahelo-Sudanian zone ofWest Africa. The variation in the use of fire formanagement purposes depends on the status ofthe vegetation in a given area and the relation-ships between the local livelihood and naturalresources. In Senegal one can identify a north-south variation with a gradual shift, in landcover types, rainfall, and agricultural/fire prac-tices.

The fire regimes, causes, effects and fire man-agement issues observed in Senegal arebelieved to be representative for most of theSahel-Sudan region. In Maps 8 and 9 the pat-tern of fire occurrence in one particular dry sea-son shows that a few fires occur where rainfallis below 800 mm. The same pattern is observedin other countries with slight difference in rain-fall limit of fire occurrence.

A combination of the fire distribution andtiming derived from several years of NOAAAVHRR and MODIS with the land cover mapenable a stratification of Senegal into regions ofuniform fire regime.

27

Vegetation and Land Use in Senegal Map 10

Red lines delineate the fire regime zones http://www.bcconline.org/wolof/Maps/VegetationMap.htm

Fires2 6/15/09 8:14 PM Page 27

28

Description of the Main Fire Regimes

1. The River ValleyThe main land cover types in the Senegaleseriver valley are marshes (with dense invasiveplants), agricultural land (rice, sugar cane,vegetables), and steeps (with grass andshrubs). Given the humidity and vegetationcharacteristics a few fires associated with theremoval of crop residue in rice fields are notedevery year. In addition, sugar cane extractionis done with extensive use of fire to facilitatecropping. Recently it has been noted that thefight against the encroachment into waterbodies by Phragmites australis, necessitatedextensive use of fire to remove the invasiveplants hindering fishing and agriculturalactivities.

2. The FerloThe dominant land cover types are rangelandswith a continuous grass layer. Although this isa fire prone zone, few accidental fires are notedearly in the dry season (September, October andNovember). These fires can cover large areasbecause of extensive plateaux and continuousgrass layer. In this area, fires are prevented bylocal people because of the scarcity of theherbaceous biomass (low fuel load) which is themain fodder for livestock. There is a deep com-mitment of local stakeholders to ban fires inorder to preserve the main source of income,pastoralism.

3. The Peanut BasinNo fires have been noted in this area becauseof the absence of a continuous natural grasscover that could burn extensively. The land ismainly used for groundnut cultivation inalong with millet and secondary crops likebeans and maize. However, some small firesassociated with field preparation can occurprior to the rainy season. These fires are dis-persed and small such that only a few areobserved by use of satellite images. As a resultthe satellite based fire regime characterisationdescribes the Peanut Basin as an area of lowfire occurrence.

4. The Transition ZoneTrees and shrub savannas dominate the landcover of this area. This zone is characterised bya continuous grass layer and the presence ofsparse trees. The use of fire is subject to greatcontrol because of its effects on the ecosystemand on the income of local people.

5. Eastern SenegalThe savanna woodland ecosystems in thiszone are particularly favourable for fire prop-agation (Map 10). Rural activities are based ona mixture of crop and livestock production

(Table 2). These two factors require the use offire to prepare the land for new cultivationand to remove some patches of the herba-ceous carpet to promote grass re-growth anddestroy non-valuable grass. The NiokoloKoba National Park is located in this zone andfire is used as a management tool to promotetourist activities.

6. Central SouthThe high fire frequency in this area is associat-ed with a relatively high human density andseveral activities causing fires are noted: agri-culture, charcoal making, pastoralism, andhunting. Due to the high fire incidence in thisarea most of the past and current projects onbushfires have been implemented here. Thepopulation density is relatively high with a pos-itive migration responsible for rapid spatialexpansion of slash and burn agriculture forcash crops production (peanuts and cotton).The improved accessibility of this area withnew roads made by the cotton company (SOD-EFITEX) could be one of the reasons why farm-ing settlements are spreading through theregion.

7-8. MangroveThe main mangrove zone of Senegal is locatedin the western side of the country in the Deltavalley of Casamance and Saloum rivers. Theseareas are characterised by the presence of moistvegetation with little grass. These ecosystemscannot catch fire except the low plateaus inbetween the salty depressions. Some few casesof fires have been reported in these plateauswhere the terrestrial vegetation is dominatedby savanna types with a consistent grass layer.In the Delta du Saloum (8) there is a NationalPark where fires are used at the onset of the dryseason. The Casamance area is characterised byeffective control of fires. There is a fear of set-ting fire in this area because of traditional/cul-tural censure associated with deliberate burn-ing. Pasture land is mainly found on the smallplateaus where livestock raising is an importantactivity.

The savanna ecosystem is quite heteroge-neous in terms of fire regime. This difference islinked to the subtle variation of parameterssuch as the herbaceous layer continuity and thefuel load. Where the tree cover increases, theherbaceous carpet becomes patchy. It is notedfrom field observation that the highest biomassload is found in Trees and Woodland savannas,which are known to experience intensive fires(Mbow, 2000).

The analysis of the fire regimes clearly showswhich areas have high risk and frequency offire. This could form the basis on which to pri-oritise areas for the allocation of resources forbush fire management (Orgle, 1998).

Fires2 6/15/09 8:14 PM Page 28

Temporal Distribution of Bushfires in West AfricaThe temporal distribution of bushfires in WestAfrica can be divided into three main periods,each of which corresponds to the end of the dryseason following a north-south trend. Itappears from MODIS data analysis (November2003 to March 2004) that fires appear first in thenorth (Sahelian zone) as the grass layer driesout earlier in this zone (September, October andNovember). The fire frequency is not very highduring this period because of fuel scarcity andlocal fire prevention strategies to protect fodderfor livestock. In the same period the fuel mois-ture content in the southern part is still relative-ly high to support ignition.

The second period corresponds to the peakof fire frequency in West Africa spanning themonths of December, January, and February.Most of the management fires occur during thisperiod. It corresponds to the end of the harvestand the start of non-agricultural activities –hunting, honey collection, charcoal making,management fires, among others – that requirethe extensive use of burning. This trend is clear-ly illustrated in Figure 2, which shows a rapidincrease of fires and a slow decrease of burningas we go further into the dry season.

This trend of fire occurrence is very relevantto fire protection planning. Figure 2 shows thatthe most critical period for the occurrence offire in West Africa is from December toFebruary.

Fire Management The fire management history in West Africa isdifficult to address and differs according tocolonial legacies of the former French coloniesand English ones. In general terms there aretwo main periods in the fire management evo-lution in West Africa. The first period corre-sponds with the colonial time, while the secondperiod is the post-colonial era.

During the pre-colonial era, fire was consid-ered as a destructive phenomenon and waslargely banned. Perceptions of, and attitudes tobushfires – and to the African environment ingeneral – changed during the colonial era whennew institutional responses emerged. Attemptsto introduce repressive bushfire regimes as anintegral part of forest reservation policies werechallenged throughout the colonial period, par-ticularly in terms of local struggles over landand access to woodland resources. Prominencewas given to the perceived threats posed bybushfires even though some agricultural offi-cers also questioned the wisdom of burningbans, and recognised the many economic andcultural uses of fire.

The use of fire as a land-clearing tool duringthe colonial era was meant to fill the labour gapcreated by slavery which left many agriculturalareas (such as Haute Volta, Soudan Francais,Niger and the northern frontiers of the GoldCoast colony) with few, if any able-bodied men.

The total ban option did not give the expect-ed results because of the role and importance offire in achieving several land managementobjectives and in the promotion of social andeconomic activities. Thereafter, some scientifictests, such as those carried out by Aubrevilleand Chevalier (Brookman, Emissah, 1980;Louppe et. al., 1995; Schmitz et. al., 1996), on therole of fire in ecosystems showed that earlyburning, when performed correctly can be ben-eficial to the environment. The results of thesetests carried out around 1950 showed, firstly,that early fires do not totally destroy the herbsbut save part of the natural woody regenerationand, secondly, that these early fires are efficientin preventing late fire. Based on this evidence,legislation was modified to accept the use ofearly fires. This move led to a more flexible atti-tude with regard to use of fire, and this is clear-ly stated by Moor (1935) who stresses that“Complete protection is both impracticable andundesirable”.

The application of the legislation promotingearly fires faced some difficulties in its imple-mentation. Forest Departments did not havethe means to control the use of fires everywhereand the so-called early controlled burningbecame was not adequately controlled.Therefore, use of fire increased in all the WestAfrican countries. After the peak of fire occur-rence during the drought of the seventies and in

29

Temporal Distribution of Figure 2Fires in West Africa

Source: Modis Data 2003-2004

Fires2 6/15/09 8:14 PM Page 29

30

the early eighties, some Forest Departmentswere provided with equipment and consider-able budgetary allocations. In some countriessuch as Ghana, a National Fire Service was cre-ated to focus on fire fighting strategies and aBushfire Prevention Law was promulgated in1983 to address the problem of bushfires byenforcing measures aimed at curtailing the inci-dence and magnitude of burning (Ampadu-Agyei and Atsiatorme, 1998). At the same timethe legislation for fire prevention was strength-ened in other countries. For example, inBurkina Faso fire was considered a crime pun-ishable by heavy penalties. In Benin and Mali,fire laws were enacted in 1982 and 1986, respec-tively, to formalise the fire policy at nationallevel (Schmitz et al. 1996).

The concern about bushfires is one of themain interests of forest commissions as clearlystated in National Forest Action Plans. Throughsensitisation, construction of firebreaks, earlyfires and formation of local fire committees andproviding equipment, the Forest Departmentsaim to prevent drastic burning to mitigate for-est loss due to fire. Sensitisation is becoming themajor strategy to increase awareness of theenvironmental impacts of fire. Through RuralRadio, the local population has the opportunityto broadcast important information on bush firecontrol. This technology is widely used in mostWest African countries, including Senegal,Mali, Guinea, Niger, Burkina Faso and Benin. Inparallel, active fire fighting is applied to stopactive fires. This has been done since thedrought of the 1970s. The active fight againstfires is based on mobilisation of considerablepersonnel and logistics to stop fires.

Current efforts are shifting to management offires, which depict both the acceptance of theimportance of fires in rural activities and therequirement to control fire for better ecosystemrenewal. This new approach is being hamperedby a general perception of the forest resource asbeing owned by the State and that responsibili-ty for its protection falls under official services.This situation results in the continued exclusionof local people from obtaining economic bene-fits from the use of natural forest resources, andconsequently there is no local incentive to stopor minimise the use of fire since the forestresources will benefit only the “foreigners”(Mbow et al., 2001).

Current fire management strategies seek topromote the full commitment of local peoplein the fire management process. To achievethis goal, the decentralisation process takingplace in most West African countries puts agreat emphasis on allocation of forestresources to local stakeholders. The logic is toplace the fire mitigation policy within theframework of the local agro-forestry activities,controlled by local people. The actual policy

for fire management shows a conceptual shiftfrom the coercive approach to a more persua-sive one. This requires greater involvement oflocal people through key stakeholders with aclear and careful explanation of reasons ofwhy the control of grass burning (limited useof fires) is relevant for the local context. In thatrespect, forest agents have been trained in par-ticipatory approaches to better facilitate thejoint effort between the Forest Department andthe local government in the context of decen-tralisation. Moreover, it is noted that while thelegislation is clear about the responsibility oflocal government towards natural resourcesmanagement there is not a clear back-up of the“community-based” fire management policies.There is a transfer of the decision makingprocess and not a transfer of the means bywhich to apply the local decisions. In addition,local stakeholders have not yet achieved aneffective transfer of managerial responsibili-ties, which requires budget transfer and plan-ning skills.

The use of early fires for land management isclearly expressed in the forest legislation inmost West African countries. Its applicationrequires some caution, which is not alwaystaken into account. It is noted at least inNational Parks where this management burn-ing takes place every year that the period ofburning is far from optimal. Burning tends totake place when the level of fuel dryness is con-ducive to severe and uncontrolled burning.Outside the protected forests the local govern-ment, in consultation with the ForestDepartment, decides the use of early fires. Thebeginning of the early burning season does not,in most cases, take into account the variationand state of herbaceous layer.

For these reasons, better early burning policyshould be based on monitoring of the herba-ceous cover to set up a spatial and temporalplan for the use of fire for management purpos-es. This can be achieved by setting up a firedanger rating programme using up to datetechniques for biological monitoring includingsatellite surveying of the fire prone areas. TheSenegalese Laboratoire de Géomatique (LERG)and the Centre de Suivi Ecologique (CSE) areknown to derive various environmental prod-ucts from satellite data including fire monitor-ing and could therefore, play an important rolein producing relevant information for the deci-sion makers.

ConclusionWest African savannas are known to be fireprone because of regular burning that occursevery dry season. The complexity of the land-scape (savanna mosaic) and the variability oflocal livelihoods towards natural resource man-agement lead to different fire regimes.

Fires2 6/15/09 8:14 PM Page 30

Whatever the socio-economic motivations forthe use of fire, the multiple consequences andside effects resulting from burning should beconsidered and should motivate a limited useof fire, given that for some purposes there is noalternative for land management.

One of the key challenges highlighted in thiscase study is the understanding of how andwhy fire is used in West African landscapeswith a focus on Senegal. This will ultimatelyhelp to unravel the complex historical, ecologi-cal and socio-cultural realities and, ultimately,provide opportunities for State institutions tobetter harness the knowledge and skills of localresource managers.

The fire management constraints emanatefrom budgetary difficulties and sectoralismassociated with the decentralisation process.

This results in conflicting policies and limita-tions in community based natural resourcemanagement at the local level (Goldammer andDe Ronde, 2004)

The use of fire should be included in theoverall forestry programme and not consideredin isolation. Controlled use of fire is a goodpathway to promote: the renewal of degradedland, the reduction of greenhouse gasesreleased into the atmosphere, the protection ofsoil against erosion and the improvement of thecarbon storage possibilities in line with theClean Development Mechanism initiative.

Fire vulnerability in West African countriescould only be mitigated if decision-making isbased on an early warning system that givesinformation on fire risk and active fire distribu-tion.

31

Fires2 6/15/09 8:14 PM Page 31

32

References

Ampadu-Agyei O., Atsiatorme, L.D., (1998). Management and support for anti-bushfire programmes. Proceedings,Strategies for bushfire intervention in Ghana. The National Workshop, Accra February 1998, pp. 18-27.

Brookman-Ammissah, J., Hall, J.B., Swaine, M.D., Attakorah, J.Y., (1980). A re-assessment of fire protectionexperiment in north-eastern Ghana savanna. Journal of Applied Ecology, 17, pp. 85-99.

Goldammer, J.G., de Ronde, C (Eds)., (2004). Wildland Fire Management Handbook for Sub-Sahara Africa. Preface.Global Fire Monitoring Center (GFMC), p. 432.

Hough, J.H., (1993). Why burn the bush ? Social approaches to bush fire management in West African national parks.Biological Conservation 65, pp. 23-28.

Korem, A., (1985). Bush fire and agricultural development in Ghana. Ghana Publishing Corporation, Accra. p.220.

Louppe, D. Ouattara, N., Coulibaly, A., (1995). Effet des feux de brousse sur la végétation. Bois et Forêts desTropiques, N° 245, pp. 59-73.

Mbow, C., (2000). Caractéristiques spatio-temporelles des feux de brousse et leurs relations avec la vegetation dans leParc National du Niokolo Koba (Sud-est Sénégal). Thèse de Troisième Cycle, Institut des sciences del’Environnement, Université de Dakar. p. 125.

Mbow, C. Nielsen, T.T., Rasmussen, K., (2000). Savanna fires in east-central Senegal: Distribution patterns,resource management and perceptions. Human Ecology, Vol. 28, N°4 , pp. 561-583.

MBow, C., Goïta, K., Bénié, G.B., (2004). Spectral indices and fire behaviour simulation for fire risk assessmentin savanna ecosystems. Remote Sensing of Environment, 91: 1-13.

Mbow, C., Goudiaby, A., Sambou, B., Ba, A.T., (2002). Towards a community forest strategy to minimise theoccurrence of bushfires in Senegal. Serein Occasional Paper N° 13; pp. 137-146.

Nielsen, T.T., Rasmussen, K., (1997). The distribution in time and space of savanna fires in Burkina Faso asdetermined form NOAA AVHRR data. Danish Journal of Geography, 97, 1: 86-98.

Nielsen, T.T., Rasmussen, K., (2001). Utilization of NOAA AVHRR for assessing the determinants of savannafire distribution of Burkina Faso. International Journal of Wildland Fire, 10: 129-135.

Nielsen, T.T., Rasmussen, K., Mbow, C., Touré, A., (2003). The fire regime of Senegal and its determinants.Danish Journal of Geography. 103 (1), pp. 43-54.

Orgle, T.K., (1998). Forest Fire Management in Ghana. In Strategies for bushfire intervention in Ghana. Proceedings,National Workshop, Accra 3-4 February,1998. pp. 42-47.

Saarnak, C., Nielsen, T.T., and Mbow, C., (2003). A local scale study of the trace gas emissions from vegetationburning around the village of Dalun, Ghana, with respect to seasonal vegetation change and burning prac-tices. Climatic Change, 56: 321-338.

Schmitz, A., Fall., A.O., Rouchiche, S., (1996). Contrôle et utilisation du feu en zones arides et subhumidesafricaines. Cahiers FAO Conservation, N° 29, p.211.

Wardell, D.A., Nielsen, T.T., Rasmussen, K. and Mbow, C., (2004). Fire history, Fire regimes and fire manage-ment in West Africa - an overview. In: Goldammer, J.G. and de Ronde, C. (eds.) Wild land Fire ManagementHandbook for Sub-Sahara Africa. Global Fire Management Centre/Oneworldbooks: 350-381.

Fires2 6/15/09 8:14 PM Page 32

Simon Mwansasu1

AbstractTanzania has a population of 34 million people.Forests and woodlands cover 38.8 million ha,constituting about 40 per cent of the country’stotal land area. The woodlands consist of second-ary forest dominated by extensive savannahs thathave developed as a result of frequent fires andovergrazing. Rainfall averages 400mm. Fire out-breaks are mainly attributed to preparation offarms, accounting to 35 per cent of all possiblefire causes. Prescribed burning is practiced espe-cially in protected areas like National Parks andGame Reserves. The use of remote sensed data isincreasingly becoming important as demonstrat-ed in the case of Tarangire National Park. Firemanagement falls under two different ministries– the Fire Services Force under the Ministry ofHome Affairs and the Forest and BeekeepingDivision under the Ministry of Natural Resourcesand Tourism. Each ministry is guided by differ-ent laws and regulations such as Fire and RescueServices Force Act No. 3 of 1985 and the ForestAct No. 14 of 2002.

IntroductionTanzania lies between Latitudes 1 and 12 degreessouth and Longitudes 29 and 41 degrees east. It isbound by the Democratic Republic of Congo,Burundi and Rwanda to the west, Zambia to thesouthwest, Uganda to the northwest, Kenya tothe northeast, Malawi and Mozambique to thesouth and the Indian Ocean to the east. Tanzaniahas a population of 34 million people (Census,2002). About 51 per cent are women and 46 percent of the total population is under 15 years ofage. Around 70 per cent of the population lives inrural areas (United Republic of Tanzania, 2004).Agriculture is an important sector in rural areascontributing 45.1 per cent of the GDP in 2002while other sectors had the following contribu-tion: Transport (4.9 per cent), Commerce (12.4 percent), Construction (5.2 per cent), Manufacturing(7.5 per cent) and others (24.9 percent). Tourism isincreasingly becoming an important sector.

Physiographic Regions and ClimateThe landscape of Tanzania consists of mainlythree physiographic regions, namely, theislands and the coastal plains to the east; theinland saucer shaped plateau – forming the

major part of the county; and the highlands.The Great Rift Valley cuts through the middleof the country, running north to south and split-ting at Lake Nyasa/Niassa/Malawi.

The climate of Tanzania varies with altitudefrom the highest point on the continent, Mt.Kilimanjaro – which peaks at 5 895m – to sealevel. The narrow lowland coastal region is con-sistently hot and humid, while the central regionsare sufficiently elevated with cooler tempera-tures. In these two regions temperatures neverfall below 20°C and the hottest period rangesfrom 25-31°C. In the highlands, temperaturesrange between 10 and 20°C. There are two rain-fall regimes. One is unimodal (December – April)and the other is bimodal (October – Decemberand March – May). The former is experienced inthe southern, south-western, central and westernparts of the country and the later is found to thenorth and northern coast. Average annual rainfallvaries from 200 mm to 1,000 mm over most partsof the country. Higher rainfall, often over 2,000mm, is commonly recorded in the highlands.Central Tanzania, comprising mostly the post firesecondary forests dominated by the savannareceives an annual average of 400 mm(http://statehouse.go.tz/abouttanzania/).

Land Use and Vegetation TypesForests and woodlands cover 38.8 million ha,more than 40 per cent of the total land area inTanzania. Woodland is the dominant vegetationtype covering large areas in the western andsouthern parts of the country (Map 11). Othermajor land use types are presented in Graph 3.Most forests and woodland (in public land andto some extent in protected areas) are secondaryforests having regenerated naturally after sig-nificant human and natural disturbances of theoriginal forest vegetation. With fire playing amajor role in these disturbances the forests dis-play major differences in structure, canopy andspecies composition with respect to nearby pri-mary forest on similar sites. There are at leastfive types of secondary forests:� Post extraction secondary forest – these

include miombo woodlands found almostthroughout the country.

� Fallow secondary forest – these are generallywoodland with scattered cropland. They aremainly found in the drier central plateau ofthe country.

33

UNITED REPUBLIC OF TANZANIA

1 Institute of Resource Assessment, University of Dar es Salaamsmwansasu@ira.udsm.ac.tz

Fires2 6/15/09 8:14 PM Page 33

34

� Rehabilitated secondary forest – these arelargely degraded lands due to overgrazingand shifting cultivation.

� Post fire secondary forest – these forests coverlarge areas in Tanzania. They have developedas a result of frequent fires and overgrazing.The extensive savanna woodlands in the semi-arid central, western and southern parts of thecountry constitute a post fire secondary zone.Grazing is an important activity in this zoneoften regarded as rangelands.

� Post abandonment secondary forests – theseare generally dry areas with low potential ofregeneration.

Causes and Effects of FiresThere are frequent fire outbreaks or occurrencesin both protected areas and public land thathave resulted in the secondary forest typesdescribed above. The causes of fire in Tanzaniahave a lot to do with land tenure systems, espe-cially where the land is protected (NationalPark, Game Reserve, Forest Reserve) or public.Public land is often viewed as belonging to noone and there is little incentive to protect it fromfire.

Protected Areas and Forest PlantationsFire is often used as an ecological manage-ment tool in many National Parks and GameReserves. Fire is set in selected areas duringthe dry season in what is described as earlyburning. The aim is to induce the growth ofgrass at the beginning of the rainy season.Early burning is done under supervision toavoid fire spreading to other areas notintended for burning. Though done undersupervision, controlled burning cannotexclude the possibility of fire being spread toadjacent areas, be it within the protected areaor the adjacent public land, by various factorslike wind. Between 1995 and 2000, there were35 incidences of fire in Reserved Forestsaffecting 11,166 ha in Morogoro Region alone(Table 3).

Major Land Use and Cover Types in Tanzania Map 11

Land Use/Cover Distribution Graph 3for Tanzania

Source: Based on Landsat TM 1994/95

Source: Landsat TM of 1994/95

Fires2 6/15/09 8:14 PM Page 34

Public LandThere are a number of reasons for fire outbreaksin the public lands. Some of the main causes offires are:� Cattle herders set fire when the vegetation is

dry during the dry season to induce regener-ation of fresh grass. The herders usually donot have resources to control the extent of thearea to be burnt.

� Subsistence farmers use fire during the dryseason to prepare their plots. This process isoften difficult to control and the possibility of afire spreading to other areas is high. In thoseareas with inadequate rainfall, fire is used toclear the land before sowing the seeds.

� When new agricultural areas are opened aspart of shifting cultivation, burning is almostinevitable.

� Fire is also used during hunting to drive outanimals from their hideouts.

� In tsetse fly infected areas, fire is deliberatelyset during the dry season to reduce the men-ace of tsetse flies.

� Charcoal burning is always cited as anothersource of fires. There is some controversy tothis as bush fires especially those with devas-tating effects are not in the interest of char-coal burners.

� Natural fires ignited by lightning.According to Dallu (2003), records with the

Forest and Beekeeping Division between1990/91-1999/2000 show that the major causesof fire in Forest Plantations are field preparation(36 per cent), hunting (27 per cent) and arson(14 Per cent). Other causes for the same dura-tion are indicated in Graph 4.

Fires started intentionally have in manycases been very destructive, reducing for exam-ple, the grazing potential of land and exposingit to soil erosion. The species composition of thepost fire secondary forest is determined by fire.The forest is dominated by Afromosia andCombretum species. It has been argued that ifthese forests remain protected from fire for a

number of years plant fuel accumulates on theground and when a fire occurs the accumulatedplant material fuels fierce burning which expos-es the land to severe soil erosion making regen-eration difficult (Box 1).

Though most fires occur in miombo wood-lands and rangelands, they are not limited tothose areas. For example, between 1997 and1999 bush fires damaged 5,000 ha of forest andpeat land on Mt. Kilimanjaro and Mt. Kenya inneighbouring Kenya (Global Forest FireAssessment, 1999/2000).

Fire Management and Monitoring - TarangireNational ParkTarangire National Park provides a good exam-ple of fire vulnerability assessment in Tanzania.Established in 1970, the park measures approx-imately 2,600 km

2, and is the fifth largest of the

twelve national parks in Tanzania (Mchallo,1994; Hjert, 2006). The park is named after the

35

Causes of Fire Outbreak in Graph 4Forest Plantations Fire Incidences in Reserved Table 3

Forests of Morogoro Region

Year Fire incidences Area affected(ha)

1995/96 7 3 4951996/97 18 5 2701997/98 4 1 2021998/99 4 4721999/00 2 727Total 35 11 166

Source: Forest and Beekeeping Division, Ministry of NaturalResources and Tourism, 2001.

Community experience in effect Box 1 of fire suppression in Mgori ForestReserve

In Singida Region, Mgori Forest Reserve has beenunder community based management with each sur-rounding village managing a portion of it. Forest offi-cials have been emphasizing the importance of elimi-nating fire to keep the forest intact, and the extent towhich the villages could achieve this was regarded bythose government officers to be a key measure oftheir success. Many village leaders accordinglypressed for control and instituted rules which resultedin a dramatic decline in grass fires during the earlypart of the dry season in both 1995 and 1996, only tolead however to much more widespread fires in theforest later in the dry season, given the greateramount of dry biomass on the ground. The heat fromthese late fires managed to damage standing trees, aneffect that was not normally seen during earlier fires,when the grass was less dry. Villagers came to theconclusion that the traditional burning of Mgori byhunters, which without controls, tend to occur in ear-lier months, were in fact protecting the forest fromlater fires spreading and doing more damage.

Source: Wily, Liz (1996)

Source: Forest and Beekeeping Division

Fires2 6/15/09 8:14 PM Page 35

Tarangire River which provides the only per-manent water source for wildlife in the area.Tarangire River runs through the centre of thepark passing through diverse habitats and var-ied topography, gentle rolling hills and season-al swamps dominated by black cotton volcanicsoils. The park is located in a wooded steppewith an arid acacia savanna belt that is domi-nated by Acacia and Commiphora species.

The various types of woodland includeAcacia – Commiphora woodland, Acaciadrepanolobium woodland, Combretum –Dalbergia woodland, Riparian woodland, andAcacia tortilis Parkland. Other vegetation typesinclude wetlands and seasonal flood plainsvegetation, riverine grassland, deep gully veg-etation and grasslands with plenty of baobabtrees (Map 12). Tarangire National Park has thehighest concentration of wildlife during thedry season due to Tarangire River (Forey, 2002).The dry season occurs from July - Septemberand temperatures range between 18 and 32 oC.The wet season occurs from mid-October tomid-May and temperatures are between 21 and27oC. Rainfall is usually in the region of 1,600mm per annum and is heaviest fromNovember through to April (Kahurananga andSilkiluwasha, 1997). During the rainy season,animals disperse to the greater Tarangireecosystem.

Fire Issues in Tarangire National ParkThe park resembles an island surrounded bythe pastoralist Maasai except on the eastern partbordering Babati District where farmers rely oncultivation although they may keep some live-stock like cattle and goats. Fire is widely used inthe preparation of farms. In recent years, theMaasai pastoralists have engaged in agricultur-al activities for subsistence. Fire has been partof the ecology for many years and is responsi-ble for the Park’s high biological diversity.

Observations mainly based on years of experi-ence, show that the growth rate of new vegetationat the onset of the rains is very different on burntareas compared to unburnt areas and this differ-ence is evident throughout the park (Paul andMarc, 2003). After burning, soil is exposed to thedrying effects of the sun and wind. The exposedsoil may be colonised by other species that havebetter dispersal ability than the previous species.The exposure may also result in soil erosion anddesertification if the rains are late or come asfloods, which are a common phenomenon.

Recent research in the Ngorongoro Crater,where fire has been artificially suppressed,showed that tick numbers increase two-foldwith every 0.5m increase in grass height andthat it was the control of fire which resulted inthe increase in parasites noted on Buffalos in2001/2002 (Mills, et al, 2003) This ‘parasite

36

Major Vegetation Types of Map 12Tarangire National Park

Mapped using Landsat TM 1994/95 data

Distribution of Burnt Areas in Map 13Tarangire National Park

Based on Landsat TM 1994/95 and GPS survey of 2003

Fires2 6/15/09 8:14 PM Page 36

load’ along with late rains in 2002 resulted inthe deaths of hundreds of Buffalo in the crater.

Prescribed burning is done in TarangireNational Park according to the ecological man-agement of the park. However, most fires affect-ing the park are ignited outside and spread intothe park burning considerable areas.

Spatial Distribution of Fires in Tarangire National ParkData derived from the Landsat TM 1995 imageshows that eight per cent of the park was burntin that year (Map 13). The burnt areas com-prised mainly of the grass plains. In 2003, thearea burnt increased to 15 per cent and morethan 80 per cent of these were grassland plains.The rest was bush land and very little woodedvegetation. Analysis of Maps 12 and 13, showthat the area burnt in both 1995 and 2003 repre-sented two per cent of the total park area.

Evolution of Fire Management in TanzaniaThe evolution and management of fire issues inTanzania falls under two categories. There isthe evolution of fire management in urbanareas, specifically in town councils, airports andsea ports and the evolution of fire managementin forest areas (both protected and public land).There are, therefore, two Acts dealing with firemanagement. One under the Ministry of HomeAffairs and another under the Ministry ofNatural Resources and Tourism.

Evolution of Fire Management in Urban AreasThe first Police Fire Brigade was established in1945. Between 1953 and 1981, the Police FireBrigade was overwhelmed and a number ofFire Brigades were established. In 1972, thePolice Fire Brigade was transferred from theMinistry of Home Affairs to the Ministry ofWorks and then back to Home Affairs in 1978.In response to the increasing number of fireincidents and the need to have organised serv-ices pertaining to fire hazard rescue operations,

the “Fire and Rescue Services Force Act No.3 of1985” was passed to provide for the betterorganisation, administration, discipline andoperation of fire and rescue brigade services.

Evolution of Fire Management in Forest AreasIn principle all forest areas, whether protected ornot, were protected against fire under the ForestOrdinance Act of 1957 Cap 389 and the GrassFires Control Ordinance Cap 135 of 1943. The aimof Cap 389 was to protect and manage forestareas under the central government and towncouncils in Tanzania. The Grass Fires ControlOrdinance prohibited setting fire on natural veg-etation although this could be done with permis-sion of the District Commissioner. Once the per-mission was granted, the law also required theperson to inform in writing, the DistrictAdministrative Officer and neighbours withinhalf a kilometre of the area to be burnt, two daysbefore actual setting of fire. The New Forest ActNo. 14 of 2002 has merged the two ordinances tofacilitate among other things, the management offire in forest areas. Part IX sets conditions for pre-venting fires, procedures to follow for setting fire,measures to be taken once one is caught settingfire and outlines the responsibilities of each citi-zen to extinguish fire once there is a an outbreak.In principle there are no major differences withthe Grass Fire Control Ordinance except that thelatter emphasises enforcement.

Current StatutesCurrently, there are two statutes dealing withfire management in Tanzania. These are the Fireand Rescue Services Force Act No.3 of 1985 andthe Forest Act No.14 of 2002 (Table 4).

Part IX of the Forest Act No.14 of 2002 dealswith fires. This part has largely been inheritedfrom the Grass Fires Control Ordinance Cap 135of 1943. Under part XI (Penalties and Offences) ofthe Forest Act, any person found guilty shall beliable to a fine or to imprisonment for a term notexceeding one year or to both.

37

Details of the Fire and Rescue Services Force Act No. 3 of 1985 and Table 4 the Forest Act No.14 of 2002

Fire and Rescue Services Act, No.3 of 1985: Part IX of the Forest Act No.14 of 2002:- Duties of the Fire force under section Provisions.

5 (1) of the Act.

Extinguish fires; 70. Restriction on burning of vegetation.Protecting life and property in case of fire or 71. Power to require persons to assist in other calamities; extinguishing fire.Discharging such other duties as may be imposed 72. Orders in relation to firebreaks.on it by law or by any direction of the Minister. 73. Fires kindled on land of another to be controlled

and extinguished.74. Saving of counter firing.75. Saving of right to recover damages.76. Definition of the words owner or occupier.

Source: Ntamuturano, 2003.

Fires2 6/15/09 8:14 PM Page 37

38

Limitations of the Fire Management StrategyThere are two organisations dealing with firemanagement in Tanzania. These are the Fireand Rescue Services and the Forestry Department. The Fire and Rescue Services servesmostly the urban areas, covering industries,towns, air and seaports, while the ForestryDepartment covers forests in public and pro-tected areas. There are no indications of coordi-nated efforts between the two organisations. Inboth cases, there is lack of proper equipmentand sufficient trained human resources for firemanagement activities. Enforcement of theexisting laws also appears to be a problem.

There are no early warning systems in placefor both urban and rural areas. There are how-ever, some preventive measures that includethe construction of fire breaks and planting fireresistant species particularly in forest planta-tions. The Sao Hill forest plantation in Iringahas some observation towers for detection offire outbreaks.

Human Vulnerability to FireTanzania is prone to disasters. Most of the dis-asters are linked to its topography, climate, nat-ural resources and human activities. Accordingto the Disaster Vulnerability Analysis carriedout by the Disaster Management Department in2001 the most common disasters in over 57Districts in mainland Tanzania were related toepidemics, pest infestation, drought/famine,floods, fire, major accidents, cyclone/strongwinds, refugees, conflicts/Internally DisplacedPersons (IDPs), landslides, explosions, earth-quakes and technological disasters. Fire consti-tutes 5.4 per cent of the occurrences of disasters(Graph 5). Most fires that claim lives and prop-erty occur in urban areas. Most of the fire inci-dents are closely related to the nature of thelocation they occur. The common causes areelectrical faults and industrial inflammable

substances such as chemical and petroleumproducts.

ConclusionFire is, and will continue to be an importantecological tool for the foreseeable future. TheNew Forest Act of 2002 has now incorporatedaspects of community participation in the man-agement of forest resources. Fire issues can nolonger be ignored in those joint managementendeavours. The Mgori experience has demon-strated this reality. What is required is sustain-able use of fire. The use of remote sensed datacan greatly assist in coming up with sustainablemanagement plans and early warning systems.However, constraints such as a poor communi-cation network, lack of software, hardware andskilled manpower continue to limit the use ofgeo-spatial data even where this data is avail-able for free (http:/rapidfire.sci.gsfc.nasa.gov/).There is a need to tap into the local knowledgeof fire management that is often disregarded by“experts”. There is also need to harmonise firemanagement by having one fire act possiblyunder one body.

Common Disaster Occurrences Graph 5in Tanzania

Source: Prime Minister’s Office – Disaster ManagementDepartment (2001)

Fires2 6/15/09 8:15 PM Page 38

References

Bloesch, U., (2001). The Use of fire in the environmental rehabilitation on the site of a former refugee camp at Benaco,Tanzania.

http://www.forest.ch/zeitschrift/archiv/2001/092001.htm#Einsatz%20von%20Feuer

Dallu, A.I.M., (2002). Tropical Secondary Forest Management in Africa – Reality and Perspectives. TanzaniaCountry Paper presented at the Workshop on Tropical Secondary Forest Management in Africa. Nairobi,Kenya.

Dallu, A.I.M., (2003). Fire fighting in forest areas. Workshop proceedings: Improving the structure and service provi-sion of fire fighting and rescue services. Bagamoyo (5 – 7 May 2003)

Forey, L.S. (2002). The Influence of Environmental Factors and Human Activity on Elephant Distribution in TarangireNational Park, Tanzania. Thesis – International Institute for Geo-Information Science and Earth Observation,Enschede, The Netherlands.

Global Forest Fire Assessment (1999/2000). Forestry Department (FAO). Forest Resources AssessmentProgramme – Working paper 55 (Rome 2001) p.95

Hjert, C.J. (2006). People vs Wildlife: Buffer zones to integrate wildlife conservation and development?.Supervisor- Per Wramner. Södertörns Högskola, Miljö- och vecklingsprogrammet, C-uppsats 10 poäng

http://www.diva-portal.org/diva/getDocument?urn_nbn_se_sh_diva-728-1__fulltext.pdf

http://statehouse.go.tz/abouttanzania/ (Tanzania Climate)

Kahurananga, J. and Silkiluwasha, F (1997). The migration of zebra and wildebeest between TarangireNational Park and Simanjiro Plains, northern Tanzania, in 1972 and recent trends. African Journal of Ecology.Vol. 35 Issue 3 SEPTEMBER 1997. p. 179

Kahwa, R. (2002). The main disaster in Tanzania. Disaster Management Department - Prime Minister’s Office.Paper presented to the United Nations Regional Workshop on the Use of Space Technology for DisasterManagement for Africa (The United Nations Office for Outer Space Affairs/The United Nations EconomicCommission for Africa). Addis Ababa, Ethiopia, 1- 5 July 2002.(http://www.unoosa.org/pdf/sap/2002/ethiopia/presentations/3speaker01.pdf)

Lissu, T. (2000). Policy and Legal Issues on the Wildlife Management in Tanzania’s Pastoral Lands: The Case Study ofthe Ngorongoro Conservation Area. 2000(1) Law, Social Justice and Global Development (LGD).ttp://www2.warwick.ac.uk/fac/soc/law/elj/lgd/2000_1/lissu/

Mchallo, I.A.J (1994). The Impact of Structural Adjustment Programmes on Natural Resources Base: The caseof tourism development. UAFITI (New series), Vol. 1 No. 2, 1994.

Mills, A; Moekel, P; Runyoro, V; Amiyo, A; Muruthi, P; Binamungu, T; Borner, M and Thoirgood, S. (2003).Management of Black Rhino in the Ngorongoro Crater – A Report on the workshop held at Serena Lodge,Ngorongoro 3-4th September 2003.

National Bureau of Statistics, Planning and Privatization,(2001). Tanzania in figures 2000.

http://www.tanzania.go.tz/census/ 25th August 2004

Nssoko, E., (2001). Fire in miombo woodlands: a case study of Bukombe District Shinyanga, Tanzania. Communitiesin flames – Proceedings of an International Conference on Community Involvement in Fire Management (FAORegional Office for Asia and the Pacific). Bangkok, Thailand.

Ntamuturano, S., (2003). Fire and Rescue situation in Tanzania before and after independence. Workshop proceedings:Improving the structure and service provision of fire fighting and rescue services. Bagamoyo (5 – 7th May 2003).

Oliver, P. and Baker, M, (2003). The effect of fire in Tarangire National Park. June 2003 Bush Newsletter.http://www.oliverscamp.com/oca-tarangire/news2003.html

United Republic of Tanzania, Ministry of Natural Resources and Tourism, (1997). Forest ResourcesManagement Project – National Reconnaissance Level Land Use and Natural Resources Mapping Project –Final Report Volume II Data Dictionary. Hunting Technical Services.

Wily, Liz, (1996). Collaborative Forest Management Villagers and Government – The Case of Mgori Forest.Miombo community woodland management case studies. Forests, Trees and People Programme andNetwork (FTP / FTPP), FAO / SLU. FAO 1996

39

Fires2 6/15/09 8:15 PM Page 39

40

Fires2 6/15/09 8:15 PM Page 40

Kolethi C. Gumbo

AbstractFires play a natural and useful role in the lifecycle of forests. However, fire can also have adevastating long-term effect on ecosystems thatare not adapted to frequent burning. Frequentfires in tropical forests such as those found inZimbabwe, alter forest species and structuralcomposition, regeneration and recovery poten-tial and also flammability. Frequent burning hasimplications on carbon stocks and emissions,wildlife habitat, human health and livelihoods.In a mature forest, a fire might burn only the for-est floor and leave larger trees undamaged.

In Zimbabwe where there is a long dry sea-son, dry plant debris in the miombo woodlandsprovide a lot of fuel on the ground and as aresult fire flames are hot enough to reach theforest canopy and spread rapidly over largeareas. Human activity is the major cause of fire.This includes burning to clear forests for theestablishment of agricultural land, improve-ment of grazing, hunting and small fires used inlivelihood activities such as honey gathering,cooking or to keep warm. Fires can burn out ofcontrol in areas where there are no incentivesfor fire protection and spread into protectedareas, grazing land, forests, conservancies andnational parks causing massive damage.

Although there are laws and regulationsgoverning the use of fires, they are difficult toenforce. Fire management is done in protectedareas such as national parks and forests, conser-vancies and commercial farms but not in thecommunal lands where more than 60 per centof the country’s population lives. This casestudy summarises the role of fire in land usemanagement and environmental change; itsimpact on human vulnerability; and forest pol-icy/institutional responses in Zimbabwe.

IntroductionZimbabwe is a land-locked country in southernAfrica, covering 39 million hectares (Ministry ofMines, Environment and Tourism, 1998), andsurrounded by Zambia to the North,Mozambique to the East, South Africa to theSouth and Botswana to the West. The countryhas a population of about 12.6 million peopleaccording to the 2002 National Census. The

soils in Zimbabwe vary from less fertile sandygranite-derived soils, which are found in 70 percent of the country (Zimbabwe Geological map– Surveyor General Department), to the fertileclay igneous intrusion soils. The country has asub-tropical climate with warm wet summersand cool dry winters. The maximum tempera-tures in summer can reach 40oC and the coldtemperatures getting down to about 10oC.

Background Information about ZimbabweHigh fuel loads and a conducive climate renderZimbabwe prone to fires. The country experi-ences two distinct seasons: the wet season(summer/autumn), which facilitates thegrowth of grasses and other herbaceous vegeta-tion, and the dry season (winter/spring) whentemperatures are initially cold and then veryhot, effectively drying the vegetation. The hotand dry season is ideal for fire, as there isenough dry fuel load, low humidity and thetemperatures are high enough to sustain theignition.

According to Goldammer and de Ronde(2004), the majority of wildfires in developingcountries of the tropics and subtropics, and intemperate-boreal countries in transition, arecaused by human activities usually associatedwith land use practices and changes. Manyland use systems in these regions are vulnera-ble to wild fires. The property, health and wel-fare of people in these areas are negativelyaffected by direct and indirect consequences offire such as air pollution. Active involvement ofthe local people is therefore, recognised asbeing necessary for the successful implementa-tion of fire management programmes, especial-ly at the interfaces between wild lands, man-aged systems and residential areas.

Land Use, Climate and Vegetation CoverIn order to understand the fire situation inZimbabwe, it is necessary to look at the coun-try’s vegetation cover and land use, as the veg-etation is the fuel for the fires and the land useprovides the source of most of the fires.Uncontrolled fires are mostly associated withland uses such as farming and pasture manage-ment. Species diversity in Zimbabwe isdescribed within the context of the five wood-land types of the Zambeziaca phyto-region

41

ZIMBABWE

Forestry Commission, P O Box 467, Bulawayo, Zimbabwe.E-mail: chesa@mweb.co.zw

Fires2 6/15/09 8:15 PM Page 41

42

These are Miombo, Mopane, Teak, Acacia andTerminalia (Ministry of Mines, Environment andTourism, 1998). All these vegetation types arevery prone to fire (Table 5).

Much of the country’s forest cover, other thanthe Mopane woodland, is found in the gazettedstate forests, National Parks, commercial farm-ing areas and the Eastern Highlands part of thecountry. About 40 per cent of the country’s nat-

ural forests and woodlands are found in thecommunal areas. These areas are the mostexposed to frequent natural fires, indiscrimi-nate tree cutting, and clearing for agricultureand construction purposes. Indigenous wood-lands cover about 25.8 million ha – 66 per centof the total land area. (VegRIS – ForestryCommission, 1996).

Land use in Zimbabwe is varied, rangingfrom agriculture (which includes crop farmingand livestock ranching), forestry, wildlife, andsettlements. Farming, which includes bothcommunal and commercial agriculture, consti-tutes the major land use found throughout thecountry. The land use systems have a lot ofinfluence on the occurrence of fires. Forestryand wildlife areas are strictly conservationareas, where there are fire prevention and mon-itoring systems. The investment in fire protec-tion mechanisms such as fireguards in somecommercial farming areas has lessened thespread of fire. However, the same cannot besaid for the communal lands. Frost (2003) notedthat as in the rest of Africa, where the landtenure in Zimbabwe is communal the responsi-bility for controlling the use of fires is a collec-tive effort involving the whole community.

Estimates of the Extent of Table 5Land Cover Types of Zimbabwe

Land cover Area % of total(000 ha) area

Afro-montane Rainforest 12 0.03Exotic plantations 156 0.40Indigenous woodlands 25 772 65.93Grasslands 1 894 4.85Cultivated land 10 783 27.56Settlements 139 0.36Other (water, rocky outcrops) 334 0.85Total 39 090 100

Source: Forestry Commission Report, 1996

Zimbabwe Vegetation Cover Map 14

Source: Forestry Commission – adapted from Wild and Barbosa 2002

Fires2 6/15/09 8:15 PM Page 42

Climatic conditions in Zimbabwe are moder-ated by altitude, aspect, proximity to maritimeinfluence and influence of the mid-continentalhigh pressure. There are three seasons inZimbabwe which are the hot dry season, warmto hot wet season and cool to warm dry season.Although the early dry season conditions inMay/June are conducive to fire outbreaks mostfires occur between August and October andsometimes November.

The exact time of the year at which fires startand finish depends mostly on the past rainyseason and the start of the rains in the currentseason (WWF, 2001). Hot fires usually occur latein the dry season when the grass and trees arevery dry. The hot dry season has more decayedhumus content and dried up vegetation therebyincreasing the risk of fire as land is mostly bare.Wind speed, high temperatures and highhumidity contribute to erratic and uncontrol-lable fires in August, September and October.The weather conditions in winter and rainy sea-sons cause less extreme burning during theseperiods.

FireThe Forestry Commission and the Institute ofEnvironmental Studies (IES) of the Universityof Zimbabwe have carried out some research

into wild fires and their effects. Frost (1992)states that miombo fires are fuelled largely bygrass. Fire intensity therefore depends much onthe recent grass production. Fires tend to bemore frequent and intense in areas wherecanopy cover is low, mean annual rainfall highand grazing intensity low. In miombo forests,dry-season fires in the under-stories occur reg-ularly and frequently. At a given site, there isvariation in the seasonal timing of fires and inthe interval between them.

The complete absence of fire is rare and prob-ably limited to dense miombo forests with anevergreen under-storey and limited grass cover(Frost and Robertson, 1987). According to Frost(1992) fire-return intervals at any one locationare likely to be more variable, depending onfuel accumulation and potential sources of igni-tion. Late dry-season fires are more intense anddestructive than early dry season fires,(between June and July) and late wet seasonfires, (between April and May), which are lessintense because the fuel material is still moist.However, if fires are excluded, litter accumula-tion increases the risk of accidental intensivefires.

Many of the woodlands are beginning toshow signs of damage from frequent andintense fires. In Gambiza et al. (2000) it is stated

43

Zimbabwe Land Tenure Map 15

(Adapted from the Department of the Surveyor General Land-use Map)

Key

Fires2 6/15/09 8:15 PM Page 43

44

that timber extraction exacerbates the risk offire by opening up the woodland canopy andby allowing more light to reach the herbaceouslayer as this promotes increased production ofgrass and fire-resistant shrubs that fuel fires.Fires kill the more fire-sensitive trees and sup-press the re-growth of the fire resistantspecies. Calvert (1993), carried out burningtrials for over 30 years and noted that forestfires do not eliminate the woody componentof the vegetation but stimulate the develop-ment of an under ground community. Calvertfurther showed that there has been a shift inthe proportion of individuals away fromexploited timber species and in structure fromsingle stems to coppice as well as a reductionof roughly 15 per cent in species biodiversity.

Use of fire in ZimbabweThe use of fire as a tool for land use manage-ment in Zimbabwe is widespread and histori-cal. Fire is used by farmers to clear land to plantagricultural crops and to improve grazing, andby hunters to drive out or later to attract ani-mals to the re-growth on the burnt areas.Clearing of land for the next cropping season isdone using fire in communal areas, as the use offire is the cheapest method, Frost (2003). Fire isalso used by honey-hunters to smoke out beesand by most people, especially in the ruralareas, for cooking, lighting and warming.Farmers, conservationists and foresters use firein the preparation of fireguards to protect theirareas from wild fires. Fire is also used in thecuring of tobacco.

Causes of Fire in Zimbabwe Sources of fire in Zimbabwe are both naturaland anthropogenic. Gumbo (2001) noted thatalthough lightning causes natural fires, theseare very few in Zimbabwe as in most casesthey are put out by the rains that accompanythe thunderstorms. The common anthro-pogenic causes of uncontrolled fires in publicareas include, hunting practices, clearing landfor farming, poachers, illegal gold panning,unattended fires and throwing away smoul-dering cigarettes. Fire is the cheapest methodused by small-scale farmers to clear the landfor the next cropping season.

Farmers burn the foliage in the evening andif left unattended such fires may spread toother areas. In some cases poachers start firesin order to drive animals in a direction wheretraps have been set. In Manicaland timberforests, honey hunters caused 14.8 per cent ofthe total fires with the 38 reported casesaccounting for 274 hectares, according to theTimber Producers Federation. Border TimbersLimited reported that in the ChimanimaniEstates 3,000 hectares were damaged by arsonfires started by squatters in the estate.

Machinery, including steam locomotives alsocauses fires as shown in Graph 6.

Impact of FireFires as part of the natural process, has a posi-tive role in vegetation structure and composi-tion and helps recycle nutrients contained inold and dead trees. However, if frequency,extent and pattern of burning are not strategi-cally planned they pose devastating effects. Fireimpacts on both human and environmentalprocesses. Strategic application of fires hasproven to be very useful. If the frequency iswithin the required threshold wildfires have apositive role in vegetation structure and com-position. They help to recycle nutrients andbring many benefits in land use management.The overall ecological benefit of fires is to killemerging bush species and help with nutrientrecycling. Further, they maintain open grass-lands which is more appropriate for pastureproduction.

Uncontrolled fires have devastating effectson the vegetation ecosystem. Fires lead tochanges in vegetation structure which gives riseto change in wildlife habitat structure.Predators such as lions, leopards, cheetahs areattracted to burnt areas because of availabilityand vulnerability of prey. Migration patternsalso change due to fires (Ministry of Mines,Environment and Tourism, 1998).

Death of species occurs due to suffocation,however this mostly affect species with smallhome ranges as compared to large animals(Wright and Bailey 1982). Uncontrolled fireshave devastating effects on the vegetationecosystem.

Since 2002 there has been a massive increaseof uncontrolled veld fires which have inflictedsubstantial damage to National Park areas,indigenous forests, commercial timber planta-tions, rangelands and grazing areas have beenlost (Ministry of Mines, Environment andTourism 2006). A study on the effect of fires inthe Gwaai Forest Reserve showed that fires

Causes of Fire in Demarcated Graph 6forests in Zimbabwe (1985 and 1989)

Source: Forestry Commission

Fires2 6/15/09 8:15 PM Page 44

have a depressive effect on teak woodlands asthey shift species composition from Baikiaeaand Guibourtia towards non-commercialspecies and from single to multi-stemmed root-stocks (Shaw and Timberlake 1994). Farquhar(1970) identified two shrubs namely Bauhinapetersiana and Commiphora mossambicencisas indicative of the adverse effects of fires.Other species which tend to dominate afterrepeated fires are Terminalia sericea,Ochna pul-cra and Combretum species. The complete pro-tection of teak woodlands from fires favours theregeneration of fire tolerant species such asPterocarpus angolensis and Burkea Africana(Farquhar 1970).

Although most people rely on fire for theirlivelihood, often such controlled fires breakout and cause wild fires that lead to majordestruction. Damage may constitute heavyfinancial losses and damage to humans, floraand fauna. Fires destroy grazing pastures con-tributing to loss of production in the livestocksector, which leads to poverty, as many ruralcommunities are pastoral farmers.

Fires are known to have destroyed grazingpastures for livestock with animals having tobe moved to other areas for relief grazingbefore the burnt areas recovered.

A case in point is the Save Conservancy,which in 2000 was destroyed by huge wildfires started by new farmers clearing theirlands for agriculture. Moving the cattle was avery costly exercise and those that were notmoved perished. The situation has beensevere in resettlement areas and timber plan-tations. 2004 and 2005 were the worst yearswith an unprecedented 10,925,351 ha and11,508,587 ha of land respectively beingdestroyed by uncontrolled fires.

The damage in 2005 alone is estimated attimber valued at over Z$1.5 trillion (Ministryof Mines, Environment and Tourism 2006).Properties are destroyed during fire outbreaksand in some cases, people have been killed inthese fires. Frequent fires have a role inprocesses of land degradation, which leads topoor harvests causing famine and starvationthat leads to malnutrition and diseases.

Forest Fire Management and MonitoringSome of the indigenous forests of Zimbabweare located in Matabeleland North Provinceand in Western Zimbabwe on loose Kalaharisands found in very fragile eco-systems. Theforest areas, which cover about 800,000 ha, aremanaged by the Zimbabwe ForestryCommission under multiple use principles,which involve timber production, grazing,wildlife, watershed protection and conserva-tion of biodiversity. Judge (1975) noted thatmost trees that are damaged by fires showedserious faults and blemishes that reduced their

value. Forest fires that occur annually duringthe dry season are a major concern for sustain-able management of these forests. As a result,fire protection is one of the major activities inthese forests. Data on fire incidence, spatial dis-tribution and effects are collected in order todevelop strategies aimed at improving fire pro-tection and prevention measures. This enablesthe zoning of the forests into fire protectionzones classified according to the susceptibilityof each area to fire.

During the fire season, which runs betweenJuly and November, fireguards are cleared andfire towers manned during the day and all firesoccurring within the forest area reported.Active fires within the forest are put out andmapped. All the fires mapped are plotted ontoa 1:250,000 map. At the end of the year a reportsummarising all the recorded fires is preparedgiving statistics of the numbers of the fires bycause and the total area burnt (Fig. 4). As aresult, there are several paper copy maps thathave been produced over the years showing thenumber of fires by forest area, causes and areaburnt each year. Mkosana (2000), proposed thatthis information could be used in a GeographicInformation System (GIS) to identify “hotspots” or areas that burn regularly to helpstrategise on fire protection measures.

However, fire mapping in the forest areas is arudimentary process whose accuracy cannot beverified. This usually results in unreliable esti-mates of the actual size of the fire as it is donemanually by simply sketching the fire bound-aries on the map. The documentation of theoccurrence of fire dates back to the 1960s,although there has been no meaningful analysisof the data to come up with trends on fire occur-rences. The collection of fire data is primarilyaimed at improving fire protection strategies inthe demarcated forests thereby reducing fireprotection costs. Zoning of areas into fire sus-ceptibility zones is the major goal of this exer-cise but this has not yet been achieved. As a

45

Areas Burnt in Demarcated Graph 7Forest Areas in Zimbabwe (1989 and 1992)

Source: Forestry Commission

Fires2 6/15/09 8:15 PM Page 45

46

result despite the large investment that is putinto fire protection, fires still occur within theseforests. There is a need to improve and strate-gise on fire protection measures so that they canbe more effective.

Use of Remote Sensing to Demonstrate FireDistributionAs stated in Tacheba et. al. (2001) the use ofremote sensing in monitoring fires providesopportunities of improving data handling,analysing and archiving the data through the useof new technologies such as GeographicInformation Systems (GIS). Fire distribution datacan be analysed in relation to other data sets likeland cover, land use, soils and climate to developeffective fire monitoring and prediction strategiesfor forest fire management. New satellite datasources that can be used for fire studies such asLandsat ETM+, ASTER and the MODerateResolution Imaging Spectoradiometre (MODIS)

are now available. Apart from providing directinformation on fire, such sensors also provideinformation on vegetation from which fuel loadscan be determined and fire risk assessment madewith the consideration of other factors such asland use.

As indicated in Roy (2000), Zimbabwe partic-ipated in studies to find out how MODIS fireproducts could be actively integrated into forestfire mapping and monitoring programs. TheLandsat-7 and comparative MODIS fire prod-ucts were made available for validation follow-ing the SAFARI-2000 data policy. Figure 4below is a Landsat TM False Colour image ofNorth Western Zimbabwe showing burnt areas.Due to land use patterns in the CommunalLands fires are fragmented and very difficult tomap

Forest Fire Policy and Control in Zimbabwe The use of fire in Zimbabwe is controlled by leg-islation under The Environmental ManagementAct of 2000 and The Forest Act, Chapter 19:05 of1996 Sections 67–75 which cover: � The control of fires and burning of vegeta-

tion;� The creation and maintenance of fire guards;� Measures to be taken when using controlled

fires;� Litigation in relation to damages caused by

fire;� Procedures in extinguishing fires and; � Penalties for offences related to fires.

The laws are however, very difficult toenforce, especially in the Communal lands. Ashas been noted by Frost (2003), where centralgovernments have attempted to regulate andcontrol the amount of burning, it has beenpatently impossible to exercise the control.Matose (1997) cites the case of Mafungabusiwhere most fires are caused by honey huntersdespite there being regulations to curb fire.

The solution to the problem could lie withdevolving the authority to control and use firesto the communities themselves. Frost (2003)emphasises that by seeking to institutionalisethe responsibility for burning within communi-ties, more restrained and accountable patternsof fire use should emerge. In the governmentprotected areas there are efforts to haveresource sharing schemes with communitiesliving in or bordering the forests. Communitiesget thatching grass, graze their livestock andare given part of the proceeds of revenueaccrued from the resources by the authoritiesfor community development projects likeschools and clinics. Cases that can be cited arethe Zimbabwean Communal AreasManagement Programme for IndigenousResources (Campfire) and the MafungabusiResource Sharing Scheme (Thomas 1993,Matose, 1997 and Mapedzae 2000).

An Example of Landsat Burnt Area Figure 3Mapping in North Western Matabeleland,Zimbabwe

Landsat TM+ False Colour Composite Image (Bands 4, 5,3)Scene 171 –73, spatial resolution: 30m, dated 27 August, 2000showing mapped fire scars (blue vectors) in MatabelelandNorth in Zimbabwe (Image courtesy of David Roy).

MODIS Active Fires Map 16Detected over Zimbabwe between 7 and 8 August, 2004

Fires2 6/15/09 8:15 PM Page 46

Early Warning Systems on FireDuring the dry season (March/April toOctober/November), daily forecasts of fire riskknown as the Fire Danger Estimate (FDE) areissued by the Meteorological Department andare broadcast over the radio and television bythe Zimbabwe Broadcasting Holdings. FDE arebased on three variables of weather: ambienttemperature, relative humidity and windspeed. These components, plus a factor for thenumber of days since the last rainfall, are com-puted into a model that was developed inChimanimani in 1964 (Elias 1992). This systemhas been refined in recent years.

The fire danger index is expressed in“points” and also indicated in a colour code.Prescribed burning, for instance, may be under-taken under conditions from 0 up to 39 points(Green). Greater care must be taken under con-ditions between 39 and 59 points (Orange), andburning is not permitted at all if the FDE isabove 60 points (Red). Up to 1990 the fire dan-ger estimates were issued only for the Northand the South of Manicaland Province (EasternHighlands). In 1991 this service was extendedto cover the whole of Zimbabwe. This was astep to promote fire awareness all overZimbabwe. The Fire Danger Estimate is appliedequally to forested lands, open savannas andgrasslands, and the transition types betweenforest and savanna.

In addition, the government through conser-vation agencies like the Forestry Commission,Department of Natural Resources, Agricultural

Research and Extension Services and a numberof Non-governmental Organisations, carry outextension programmes to educate the commu-nities on the dangers of uncontrolled fires, con-trolled use of fire and fire management. In pro-tected areas there are billboards discouragingforest fires. Through SAFNet Zimbabwe is alsoparticipating in the AFIS Fire-fax initiative thatis being run from the CSIR SatelliteApplications Centre in South Africa. The fire-fax is generated and sent by email together witha map and the co-ordinates of fires that havebeen detected by MODIS over the area on adaily basis. This is very helpful in identifyingand locating fires in areas where there are nofire lookout facilities.

ConclusionFire remains one of the most integral and wide-ly used tools by the Zimbabwean communitiesfor various livelihood needs. It is thereforeclear that the communities will continue to relyon the use of fire. What is needed is to educatethe communities and encourage their participa-tion in fire management so that fire remainstheir tool and not their master. There is a needfor participatory fire management programmesthat would enable communities to be part of thewhole fire management process. It must benoted that fire also plays an important role inthe maintenance of the ecosystem and thereforeresearch into appropriate fire regimes for thesustained management of the eco-systems isnecessary.

47

Fires2 6/15/09 8:15 PM Page 47

References

Bradley, P. and Dewees, P., (1993). Indigenous woodlands, agricultural production and household economy inthe communal areas. In: Bradley, P.N. & McNamara, K. (Eds.). Living with Trees: Policies for ForestryManagement in Zimbabwe. The World Bank, Washington, DC, pp. 63-137.

Calvert, G.M., (1986). Fire effects in Baikiaea woodland, Gwaai forest. In: Piearce, G.D. (ed). The Zambezi TeakForests. Forest Department, Ndola. pp. 319-325.

Campbell, B.M., Grundy, I. M. and Matose, F., (1993). Tree and woodland resources- the technical practices ofsmall-scale farmers. In: Bradley, P.N. & McNamara, K. (Eds.). Living with Trees: Policies for Forestry Managementin Zimbabwe. The World Bank, Washington, DC, pp. 29-62.

Elias F., (1992). Fire Awareness in Zimbabwe IFFN No. 7 - August 1992, p. 16-17

Farquhar, I. 1970. The effect of fire on baikiaea plurijuga woodlands in the Gwaai Forest Reserve. Field ecolo-gy dissertation, University of Zimbabwe.

Forest Act Chapter 19:05 1996 Revised edition, Government Printers Harare, Zimbabwe.

Frost, P.G.H., (1992). A draft policy for fire management in the rangelands of the Western Province of Zambia. LDP,DVTCS, Department of Agriculture, Mongu, Zambia.

Frost P. G. H., (2003). Southern African Fire Network Statement on the Use and Control of Fire in Sub-Equatorial Africa. Paper presented at the Wildland Fire Summit in Sydney, Australia.

Frost, P.G.H. & Robertson, F., (1987). The ecological effects of fire in savannah. In: B.H. Walker (ed.). Determinantsof tropical savannahs. IRL Press, Oxford. pp. 3-140.

Gambiza, J., Bond, W., Frost, P.G.H. & Higgins, S., (2000). A simulation model of miombo woodland dynam-ics under different management regimes. Ecological Economics 33: 353-368.

Goldammer, J. G. and de Ronde C., (2004). (Eds) Wildland Fire Management Handbook for Sub-Sahara Africa. Apublication of the Global Fire Monitoring Centre (GFMC).

Gumbo K., (2001) Forest Fire Management In The Indigenous Forests of Matabeleland North Province InNorth Western Zimbabwe. Paper presented at the 3rd SAFNet Meeting in Gaborone, Botswana.

Judge J. G., (1975). The Kalahari Sand Forests of Zimbabwe – Management Proposals. Typescript. InternalForestry Commission Report.

Katerere, Y., Moyo, S. & Mujakachi, K., (1993). The national context: land, agriculture and structural adjust-ment, and the Forestry Commission. In: Bradley, P.N. & McNamara, K. (Eds.).

Mapedza E., (2000). An investigation of Land cover change in Mafungabusi Forest, Zimbabwe, using GIS andparticipatory mapping. http://dlc.dlib.indiana.edu/archive/00000mpedzae 1305055 pdf

Matose, F., (1997). Institutional Configurations Around Forest Reserves in Zimbabwe:

http://dlc.indiana.edu.archive/00000863/matosef 160502 pdf

Ministry of Environment and Tourism (2006). National Fire Protection Strategy and Implementation Plan

Ministry of Mines, Environment and Tourism. (1998). Zimbabwe Biodiversity Strategy and Action Plan:Status of biodiversity, unmet needs, strategies and actions.

Mkosana F., (2000). The Possibility of Using MODIS Data in Conjunction With Landsat ETM to SupportForest Management in North Western Matabeleland in Zimbabwe. (Unpublished proposal)

Mufandaedza E., (2002). Topical Secondary Forest Management In Africa: Reality and perspectives.Zimbabwe Country Paper for the FAO/EC LNV/GTZ workshop on tropical secondary forest management inAfrica Reality and perspectives In collaboration with ICRAF and CIFOR Nairobi, Kenya, 9-13 December2002.

Rogers, C.M.L., (1993). A woody vegetation survey of the Hwange National Park. Department of National Parksand Wildlife Management, Harare. p176.

Roy D., (2000). MODIS Southern Africa Fire Products Validation Protocol. Documented after travelling meet-ing in Zimbabwe and Zambia 11 – 19th July, 2000.

Thomas, S.J., (1993). Indigenous woodlands and CAMPFIRE: Complementarities from collaboration. In:Piearce, G.D. & Gumbo, D.J. (Eds.). The Ecology and Management of Indigenous Forests in Southern Africa.Zimbabwe Forestry Commission and SAREC, Harare. pp. 71-82.

Tacheba B, Dube, P. O., Roy, D. et al., (2001). Southern Africa Fire Network (SAFNet) of the GlobalObservation of Forest Cover (GOFC): Developing Fire Monitoring and Management Systems in SouthernAfrica.

Timberlake, J. and Shaw, P., (1994). Chirinda Forest - a visitor’s guide. Forestry Commission, Harare. p. 157

48

Fires2 6/15/09 8:15 PM Page 48