26123 Water Resources and Environment Technical Note C.3 Environmental Flows: Flood Flows Series Editors Richard Davis Rafik Hirji WATER RESOURCES AND ENVIRONMENT TECHNICAL NOTE C.3 Environmental Flows: Flood Flows SERIES EDITORS RICHARD DAVIS, RAFIK HIRJI The World Bank Washington, D.C. Water Resources and Environment Technical Notes A. Environmental Issues and Lessons Note A.1 Environmental Aspects of Water Resources Management Note A.2 Water Resources Management Policy Implementation: Early Lessons B. Institutional and Regulatory Issues Note B.1 Strategic Environmental Assessment: A Watershed Approach Note B.2 Water Resources Management: Regulatory Dimensions Note B.3 Regulations for Private Sector Utilities C. Environmental Flow Assessment Note C.1 Environmental Flows: Concepts and Methods Note C.2 Environmental Flows: Case Studies Note C.3 Environmental Flows: Flood Flows Note C.4 Environmental Flows: Social Issues D. Water Quality Management Note D.1 Water Quality: Assessment and Protection Note D.2 Water Quality: Wastewater Treatment Note D.3 Water Quality: Nonpoint-Source Pollution E. Irrigation and Drainage Note E.1 Irrigation and Drainage: Development Note E.2 Irrigation and Drainage: Rehabilitation F. Water Conservation and Demand Management Note F.1 Water Conservation: Urban Utilities Note F.2 Water Conservation: Irrigation Note F.3 Wastewater Reuse G. Waterbody Management Note G.1 Groundwater Management Note G.2 Lake Management Note G.3 Wetlands Management Note G.4 Management of Aquatic Plants H. Selected topics Note H.1 Interbasin Transfers Note H.2 Desalination Note H.3 Climate Variability and Climate Change Copyright © 2003 The International Bank for Reconstruction and Development/THE WORLD BANK 1818 H Street, N.W., Washington, D.C. 20433, U.S.A. All rights reserved. Manufactured in the United States of America First printing March 2003 2 CONTENTS Foreword 5 Acknowledgments 7 Introduction 8 Establishing the Appropriateness of Managed Floods 9 Clear objectives for managed flood releases must be defined through a study of the ecosystem, economy, and social structure of the floodplain. Designing Flood Releases 11 Designing flood releases requires incorporating the ex- pectations of stakeholders; understanding linkages be- tween floods and the affected ecosystem; defining Author flood-release options; and understanding the impacts Michael C. Acreman of those options. Technical Adviser Stephen Lintner Implementation 16 Dam outlet structures need to be designed and con- Editor structed or adapted to allow necessary and feasible Robert Livernash flood releases. Pilot flood releases should be made to test the various water-release options. Production Staff Cover Design: Cathe Fadel A Practical Example ­ Design and Production: The Waza-Logone Floodplain 18 The Word Express, Inc. The Waza-Logone Conservation Project in Cameroon, coordinated by IUCN, undertook a wide range of ac- tivities in support of the design and implementation of Notes managed flood releases. Unless otherwise stated, all dollars = U.S. dollars. Conclusion All tons are metric tons. 20 The appropriateness of managed flood releases will depend on the relative merits of using water for flood Cover photo by releases compared to using it for other uses such as David Eastburn (MDBC) irrigation, public/industrial supply, or hydropower. This Macquarie Marshes, Murray decision can only be made on a case-by-base basis. Darling Basin, Australia Further Information 20 3 WATER RESOURCESANDENVIRONMENT · TECHNICAL NOTE C.3 Boxes Box 1. The Senegal valley: a suitable situation for managed flood releases 10 Box 2. Mahaweli: a river where flood releases are not applicable 11 Box 3. Participatory Rural Appraisal in the Waza-Logone project, Cameroon 12 Box 4. Community-based institutions in the Phongolo River 12 Box 5. Flood release options for the Tana River, Kenya 14 Box 6. Valuation in the Hadejia-Jama'are River Basin, Northern Nigeria 16 Box 7. Releasing sediment from Tarbela Dam, Pakistan 17 Figures Figure 1. Flow chart showing the trade-off between using water for managed flood releases and for reservoir-based activities 9 Figure 2. Relationship between flooded area and fish catch 13 4 ENVIRONMENTAL FLOWS: FLOOD FLOWS FOREWORD The environmentally sustainable development and priority in Bank lending. Many lessons have been management of water resources is a critical and learned, and these have contributed to changing complex issue for both rich and poor countries. It attitudes and practices in World Bank operations. is technically challenging and often entails difficult trade-offs among social, economic, and political con- Water resources management is also a critical de- siderations. Typically, the environment is treated velopment issue because of its many links to pov- as a marginal issue when it is actually key to sus- erty reduction, including health, agricultural tainable water management. productivity, industrial and energy development, and sustainable growth in downstream communi- According to the World Bank's recently approved ties. But strategies to reduce poverty should not lead Water Resources Sector Strategy, "the environment to further degradation of water resources or eco- is a special `water-using sector' in that most envi- logical services. Finding a balance between these ronmental concerns are a central part of overall objectives is an important aspect of the Bank's in- water resources management, and not just a part terest in sustainable development. The 2001 Envi- of a distinct water-using sector" (World Bank 2003: ronment Strategy underscores the linkages among 28). Being integral to overall water resources man- water resources management, environmental agement, the environment is "voiceless" when other sustainability, and poverty, and shows how the 2003 water using sectors have distinct voices. As a con- Water Resources Sector Strategy's call for using sequence, representatives of these other water us- water as a vehicle for increasing growth and re- ing sectors need to be fully aware of the importance ducing poverty can be carried out in a socially and of environmental aspects of water resources man- environmentally responsible manner. agement for the development of their sectoral in- terests. Over the past few decades, many nations have been subjected to the ravages of either droughts or floods. For us in the World Bank, water resources man- Unsustainable land and water use practices have agement--including the development of surface and contributed to the degradation of the water resources groundwater resources for urban, rural, agriculture, base and are undermining the primary investments energy, mining, and industrial uses, as well as the in water supply, energy and irrigation infrastruc- protection of surface and groundwater sources, pol- ture, often also contributing to loss of biodiversity. lution control, watershed management, control of In response, new policy and institutional reforms water weeds, and restoration of degraded ecosys- are being developed to ensure responsible and sus- tems such as lakes and wetlands--is an important tainable practices are put in place, and new predic- element of our lending, supporting one of the es- tive and forecasting techniques are being developed sential building blocks for sustaining livelihoods and that can help to reduce the impacts and manage for social and economic development in general. the consequences of such events. The Environment Prior to 1993, environmental considerations of such and Water Resources Sector Strategies make it clear investments were addressed reactively and prima- that water must be treated as a resource that spans rily through the Bank's safeguard policies. The 1993 multiple uses in a river basin, particularly to main- Water Resources Management Policy Paper broad- tain sufficient flows of sufficient quality at the ap- ened the development focus to include the protec- propriate times to offset upstream abstraction and tion and management of water resources in an pollution and sustain the downstream social, eco- environmentally sustainable, socially acceptable, logical, and hydrological functions of watersheds and economically efficient manner as an emerging and wetlands. 5 WATER RESOURCES ANDENVIRONMENT · TECHNICAL NOTE C.3 With the support of the Government of the Nether- The Notes are in eight categories: environmental lands, the Environment Department has prepared issues and lessons; institutional and regulatory is- an initial series of Water Resources and Environ- sues; environmental flow assessment; water qual- ment Technical Notes to improve the knowledge ity management; irrigation and drainage; water base about applying environmental management conservation (demand management); waterbody principles to water resources management. The management; and selected topics. The series may Technical Note series supports the implementation be expanded in the future to include other relevant of the World Bank 1993 Water Resources Manage- categories or topics. Not all topics will be of inter- ment Policy, 2001 Environment Strategy, and 2003 est to all specialists. Some will find the review of Water Resources Sector Strategy, as well as the past environmental practices in the water sector implementation of the Bank's safeguard policies. useful for learning and improving their perfor- The Notes are also consistent with the Millennium mance; others may find their suggestions for fur- Development Goal objectives related to environmen- ther, more detailed information to be valuable; while tal sustainability of water resources. still others will find them useful as a reference on emerging topics such as environmental flow assess- The Notes are intended for use by those without ment, environmental regulations for private water specific training in water resources management utilities, inter-basin water transfers and climate such as technical specialists, policymakers and variability and climate change. The latter topics are managers working on water sector related invest- likely to be of increasing importance as the World ments within the Bank; practitioners from bilateral, Bank implements its environment and water re- multilateral, and nongovernmental organizations; sources sector strategies and supports the next gen- and public and private sector specialists interested eration of water resources and environmental policy in environmentally sustainable water resources and institutional reforms. management. These people may have been trained as environmental, municipal, water resources, ir- rigation, power, or mining engineers; or as econo- mists; lawyers, sociologists, natural resources Kristalina Georgieva specialists, urban planners, environmental planners, Director or ecologists. Environment Department 6 ENVIRONMENTAL FLOWS: FLOOD FLOWS ACKNOWLEDGMENTS The Bank is deeply grateful to the Government of Zambia: Harry Chabwela the Netherlands for financing the production of this South Africa: Charel Brewer Technical Note. Kenya: Stephen Njuguna, Clifford Mutero Pakistan: Tahir Qureshi This Note is based on research for the World Com- Sri Lanka: H. Manthrithilake mission on Dams, funded by the United Kingdom Cameroon: Daniel Ngantou, Richard Braund Department for International Development (DFID). The research was published as: This research was a collaborative undertaking between the UK core team and the case study Acreman, M.C., F.A.K. Farquharson, M.P. McCartney, partners based in various countries around the C. Sullivan, K. Campbell, N. Hodgson, J. Morton, D. world, particularly: Smith, M. Birley, D. Knott, J. Lazenby, R. Wingfield, and E.B. Barbier. 2000. Managed flood releases from Centre for Ecology & Hydrology, Wallingford: reservoirs: issues and guidance. Report to DFID and M. Acreman, F. Farquharson, M. McCartney, the World Commission on Dams. Wallingford, UK: C. Sullivan Centre for Ecology and Hydrology. Natural Resources Institute: K. Campbell, N. Hodgson, J. Morton, D. Smith At the World Bank, this Note was reviewed by Liverpool School of Tropical Medicine: M. Birley Hans-Olav Ibrekk, Alessandro Palmieri, Tor Ziegler, Gibb Ltd: J. Lazenby, R. Wingfield, D. Knott and Jean-Roger Mercier. We are grateful for their University of York: E. Barbier suggestions. Senegal: Mbarack Diop 7 WATER RESOURCES AND ENVIRONMENT · TECHNICAL NOTE C.3 INTRODUCTION Large dams provide benefits to many people around optimize storage in multi-reservoir systems, and the world by providing water for hydropower gen- from unplanned releases made to prevent dam fail- eration, irrigation, and water supply. However, with- ures. out proper provisions being made, these dams can also cause considerable damage to upstream and Managed floods are not a panacea for downstream downstream ecosystems. In a few cases, the oper- environmental problems caused by dams. However, ating rules for these dams allow for releases to they may be appropriate where downstream wet- maintain instream ecosystems. However, almost land and riparian ecosystems support livelihoods-- exclusively, these flows remain within the river particularly where alternative livelihood strategies channel. It is rare for high flows to be released to are limited--or where important biodiversity and inundate floodplain and deltaic ecosystems. Yet, bioproductivity need to be maintained. when flooded periodically, these floodplain ecosys- tems supply important products (arable land, fish- Technical Notes C.1 to C.4 deal with environmen- eries, grazing land), functions (groundwater tal flows. Although changes in flow will affect wa- recharge, nutrient cycling) and attributes ter quality--for example, by increasing or decreasing (biodiversity). For many centuries, floodplain eco- turbidity--the focus in these Notes is primarily on systems have provided economic, social, and envi- the direct effects of flow on the ecological function- ronmental security for rural communities. Floods ing of rivers and the management of water quan- are also very important for fish migration and sedi- tity. Note C.1 introduces concepts and methods for ment transport. In some cases, managed floods have determining environmental flow requirements for been proposed--and in a few places, implemented-- rivers. Note C.2 reviews some important case his- as part of a dam mitigation strategy to restore and tories. This Note describes the reinstatement of flood conserve wetland ecosystems, which help sustain releases from reservoirs for floodplain inundation. traditional livelihoods. Such releases involve a trade- It assesses the general appropriateness and feasi- off, since less water will remain in the reservoir for bility of introducing managed floods; considers the its primary design purposes, such as hydropower, design of appropriate flood flows, including the op- irrigation, or public supply (Figure 1). timum magnitude, duration, frequency and timing of managed floods (given the economic, social, and In this Note, a managed flood is defined as a con- environmental feedbacks and tradeoffs); and dis- trolled release of water from cusses implementing the a reservoir to inundate a spe- chosen managed flood op- cific area of downstream tion. It concludes with a floodplain or river delta to case study--the Waza- restore and maintain ecologi- Bank Logone floodplain in cal processes and natural re- orld Cameroon--that illustrates W, sources for dependent many of the desirable fea- livelihoods. Such releases are tures described earlier. undertaken in collaboration Carnemark Note C.4 addresses the with stakeholders. This is dis- Curt downstream social issues by tinct from planned releases to arising from changes in Photo test floodgate operations or to Okavango Swamp, Botswana flows. 8 ENVIRONMENTAL FLOWS: FLOOD FLOWS FIGURE 1. FLOW CHART SHOWING THE TRADE OFF BETWEEN USING WATER FOR MANAGED FLOOD RELEASES AND FOR RESERVOIR BASED - - ACTIVITIES . Optimize Overall Benefit Reservoir Dependent Reservoir Dependent Flood Dependent Livelihood Livelihood Livelihood Irrigated Domestic Water Floodplain Recession Floodplain Hydropower Industry Fisheries Biodiversity Agriculture Supply Products Agriculture Gr azing Water Remaining Flood Release In Reservoir Duration,Timing, Frequency, Extent Flood Release Option Legislation Local & National Laws Development Drive Political Imperative Objective Water Availability Social & Economic Variables Hydrological Regime Engineering Conditions Livelihood Decisions Reservoir Storage AdaptationTo Dams Outlet Structures Constraints ESTABLISHING THE APPROPRIATENESS OF MANAGED FLOODS OBJECTIVES FOR FLOOD RELEASES (Box 1). The study must identify both those people who will benefit from flood releases, and those who Flows to replenish floodplains and associated eco- may suffer. For existing dams, adaptations to the systems are one of the component environmental altered hydrological regime may mean that release flows described in Technical Note C.1. As with all of managed floods will not provide significant ben- environmental flows, clear objectives for managed efits. Objectives must be defined--in the context of flood releases must be defined through a study of any river basin management plan and national/re- the ecosystem, economy, and social structure of the gional policies--and assessed alongside other op- floodplain to ensure that managed floods are com- tions for water use, including public water supply, patible with the livelihood strategies of the flood- irrigation, navigation, and hydropower. plain communities. Consequently, consultation with local people is a critical element of flood releases. ENSURING TECHNICAL FEASIBILITY Objectives can include flood recession agriculture, fishing, animal husbandry, groundwater recharge, Three factors are particularly significant for ensur- or conservation of biodiversity and bioproductivity ing the technical and financial feasibility of man- 9 WATER RESOURCES AND ENVIRONMENT · TECHNICAL NOTE C.3 BOX 1. THE SENEGAL VALLEY A SUITABLE SITUATION FOR MANAGED FLOOD RELEASES : As part of the development of the Senegal River basin, the Organisation pour la Mise en Valeur du Fleuve Senegal (OMVS) constructed two dams. At the river's mouth, the Diama dam inhibits saltwater intrusion into the river to allow its use for irrigation and to regulate water levels for navigation. In the headwaters, the Manantali dam was built to gener- ate hydroelectric power and to provide flows for navigation and irrigation in the middle valley of the Senegal River. Prior to dam construction, which began in 1986, natural inundation covered up to 250,000 hectares of the floodplain of the middle Senegal valley, and supported up to 125,000 hectares of flood recession agriculture (including maize, beans, watermelon, potatoes, and millet), grazing, forests (which provide fuelwood and construction timber), fisheries, and wildlife habitat. The estimated economic value of the floodplain was: recession agriculture, $56-136 per hectare; fishing, $140 per hectare; and grazing, $70 per hectare. It was feared that cessation of floods through operation of the Manantali dam would have a devastating effect on mid-valley livelihoods. By 1991, the reservoir had filled and the spillway was tested, but the turbines had not been installed. In subsequent years, OMVS conducted managed flood releases to inundate 100,000 hectares of land that provided a cultivation area of 50,000 hectares. This required around 7.5 billion m3 of water, yielding a value of about $2 per 1,000 m3 for agriculture, fishing, and grazing. The World Bank has agreed to help finance the installation of turbines at Manatali in 2001, provided that the principle of managed floods is recognized by OMVS as a possible long-term option. The Governments of Mali, Senegal, and Mauritania have now signed a Water Charter that includes the release of the annual flood when sufficient water is available (normally 9 years out of 10). aged floods. First, the engineering characteristics local streams entering the floodplain; rainfall; and of the dam and reservoir must be adequate. Reser- subsurface flows. If the dam has limited control over voir storage capacity must be sufficient, the outlet flooding, then managed flood releases will have little structures must be large enough to permit major impact unless the releases can be timed to augment releases, and they must be able to work operation- high flows from other tributaries. ally. For many existing dams, outlet structures may be too small, thereby making floods that adequately In addition, broad financial factors must be consid- inundate the floodplain less feasible. Where sev- ered, including the costs of installing new release eral dams occur up/downstream of each other, it structures in the dam, the opportunity costs of wa- may be necessary to coordinate operation of all dams ter releases, and the benefits to downstream liveli- to produce flood releases. In such cases, each dam hoods (Box 2). must have suitable release structures. Second, the environmental characteristics of the reservoir and its catchment must be appropriate. Thermal stratification in many reservoirs leads to poor water quality--low oxygen, high concentrations of hydrogen sulphide, iron, and manganese--at depth, which is inappropriate for release (see Note G.2). In rivers with a high sediment load during floods, sediment will need to be passed through the dam with the flood release, which may be very dif- ficult and expensive, although some solutions have been developed. Senegal, OM Third, the extent to which a dam controls flooding ORST by in target ecosystems will also depend on flow con- Photo tributions from tributaries downstream of the dam; Satellite image, Senegal River below Manatali Dam 10 ENVIRONMENTAL FLOWS: FLOOD FLOWS BOX 2. MAHAWELI: A RIVER WHERE FLOOD RELEASES ARE NOT APPLICABLE The Mahaweli Scheme is a comprehensive multipurpose water resource development designed to harness the hydroelectric and irrigation potential of the Mahaweli Ganga, Sri Lanka's largest and most important river. The river has been utilized and diverted for irrigation purposes since at least the 1st century BC. King Parakrama Badhu 1 (AD 1153- 1186) is said to have constructed or restored 165 dams, 3,910 canals, 163 major tanks and 2,376 minor tanks, making extensive use of the water resources of the Mahaweli Ganga. Today only a fraction of these constructions are still functioning, but they are nevertheless indicative of the old and extensive water diversion and use systems in the Mahaweli Ganga. The current Accelerated Mahaweli Development Programme (AMDP) includes the development of 320,000 acres of new irrigated land. Due to their complexity, large and integrated water-use structures such as seen in the Mahaweli basin may not be suitable candidates for managed flood releases. The experience gained from sediment flushing trials conducted in Rantembe reservoir by the Mahaweli Authority and Ceylon Electricity Board and Irrigation Department is indicative of some of the problems that could be expected if controlled flood releases were to be used in managing the Mahaweli system. Rantembe reservoir is silting fast, at about 6 percent per annum, and flushing has been considered as a technical option. Flushing of sediment has been carried out twice with a three-year interval. A substantial portion of the outflow from Rantembe reservoir moves along irrigation canals downstream. A third flushing event planned for 1999 was canceled due to problems resulting downstream from siltation in these canals from the earlier flushings. Both old and new weirs for diversion of water into these irrigation systems were constructed without any regulatory facilities. As a result, sediment-laden water cannot be confined to the main river. Any managed flood release from upstream reservoirs can be expected to have a similar effect downstream. The cost of converting weirs to regulate flows, together with the extensive logistical and coordination problems associated with managing the releases, have not been calculated, but are likely to be prohibitive. DESIGNING FLOOD RELEASES If preliminary screening determines that there could Many of the techniques commonly included under be significant benefits from managed flood releases, the label of "Participatory Rural Appraisal" (such and it is technically and financially feasible, then as village mapping and ranking exercises) may be possible flood releases need to be designed. useful here, but what is important is a genuine com- mitment to participation, with adequate time and STAKEHOLDER AND TECHNICAL resources (Box 3). Where necessary, the process PARTICIPATION should include building and strengthening stake- holder organizations, such as local NGOs, to en- able them to participate fully. The aim is to ensure A key element in defining appropriate flood releases that managed floods are compatible with the liveli- is to incorporate the expectations of stakeholders for hood strategies of the floodplain communities. If lo- floodplain inundation. This should begin with iden- cal social structures allow, participation can be tification of the various stakeholders, such as dam structured through the establishment of a Water operators/owners, local authorities, electricity com- Committee (Box 4), which can help in deciding on panies, farmers, and--most importantly--downstream the magnitude, frequency, duration, and timing of farmers and communities. Downstream communi- floods, as well as in informing local communities ties--floodplain farmers and residents, including when the flood is coming. pastoralists and fisherfolk--should be regarded as subject to significant risk. Getting them involved re- quires a specific and innovative effort. Their cur- At an early stage, an equally important action is to rent livelihood strategies, assets, and aspirations initiate data collection to improve understanding should be mapped to establish the population at risk. of the physical, chemical, biological, and socioeco- 11 WATER RESOURCES AND ENVIRONMENT · TECHNICAL NOTE C.3 BOX 3. PARTICIPATORY RURAL APPRAISAL IN THE WAZA-LOGONE PROJECT, CAMEROON Participatory Rural Appraisal (PRA) methods were used in the Logone Floodplain in Northern Cameroon to investigate the socioeconomic impacts of declining floods. PRA was chosen because of the difficulties of collecting complex information on interactions between society and the environment in a participatory way. It was also felt that conven- tional survey methods would fail to capture seasonal variations in an area that was inaccessible by road in the dry season. Several PRA methods were used, including: I Participatory mapping was used with both men and women to gather information on land-use, the floodplain environment and the operation of the village I Timelines identified key dates in floodplain history; these illuminated the issue of conflicts among farmers, herders, and fishers I Transect walks were used outside the village to examine natural resource issues, and inside the village to look at building materials and living conditions I Ranking was used for problems and productive activities I Wealth ranking enabled sampling of households for interview, estimation of human and livestock numbers, and understanding of local notions of wealth I Semi-structured interviews I Venn diagrams described relations between groups and institutions within and outside the village I Resource flow diagrams were used on common property issues, trade, and service delivery I Feedback to villagers Despite a relatively short time-scale and constraints such as the lack of an educated woman on the team, the experience generated a great deal of quantitative and qualitative information. The organizers were also able to set up simple environmental and socioeconomic monitoring systems that could be run by villagers. BOX 4. COMMUNITY-BASED INSTITUTIONS IN THE PHONGOLO RIVER In the late 1960s, the Pongolapoort dam was constructed on the Phongolo River in northeast South Africa near its borders with Swaziland and Mozambique. The reservoir was filled in 1970. It was intended to irrigate 40,000 hectares of agricultural land for white settlers, with no provision for hydropower generation. There were no assessments of the impacts of the impoundment on the downstream floodplain, where 70,000 Tembe-Thonga people were dependent on recession agriculture, fishing, and other wetland resources, or on the biodiversity of the Ndumu game reserve. When constructed, the dam changed the whole flooding regime of the river, which had significant negative impacts on agriculture and fisheries. In addition, few settlers used the irrigation scheme. In 1978, a workshop was held on the Phongolo floodplain to review the future of irrigation and to minimize the negative impacts on the floodplain. This led to a plan for controlled releases to rehabilitate the indigenous agricultural system and the wildlife. However, initial releases of water from the dam were made at the wrong time of the year, and crops were either washed away or rotted. In 1987, the South African Department of Water Affairs and Forestry and the local authorities agreed to experiment with community participation. As a result, water committees were established, repre- senting various user groups, including fisherfolk, livestock keepers, women, and healthworkers (both new primary health care workers and traditional herbalists and diviners). They were given the mandate to decide when floodwaters should be released. These committees successfully implemented people's views, which in turn led to river basin management that benefited the floodplain users. nomic aspects of the reservoir (see Note G.2) and zations need training and awareness building to floodplain and wetlands (see Note G.3). All reports understand and use floodplains wisely. This requires and data should be made publicly available; estab- strengthened technical skills in ecology, hydrology, lishing a local information center is an excellent engineering, health, economics, social anthropol- way to do this. In addition, implementing organi- ogy, and law. 12 ENVIRONMENTAL FLOWS: FLOOD FLOWS LINKS BETWEEN FLOODS AND FIGURE 2. RELATIONSHIP BETWEEN FLOODED AREA AND FISH CATCH THE ECOSYSTEM 180000 Indicators such as the frequency, duration, timing, 160000 and discharge of floods can be used to identify 140000 linkages between the products, functions, and at- g/yr) 120000 tributes of the floodplain and flooding and its so- (k 100000 cial implications. In particular, the flooding tch 80000 ca necessary to maintain those parts of the ecosystem 60000 required to fulfill the agreed objectives needs to be ishF 40000 quantified. For example, some wetlands may be 20000 important for recession agriculture or fisheries. This will aid the definition of an optimum, targeted flood 0 200 400 600 800 1000 1200 1400 1600 1800 2 release. Thresholds of flooding, below which a Flooded area (km ) floodplain ecosystem ceases to function adequately, Source: Welcomme, R.L., 1976. may exist, but in most cases they cannot be easily identified. sediment down the Indus. During the period 1960- Instead, available data often exhibit smooth or lin- 71, freshwater inflows were only 43 billion m3, and ear relationships between flood extent and ecologi- the Indus Water Accord only provides 12.3 billion cal productivity (Figure 2) without clear critical m3 per year. This occurs mainly in the period June- points. Consequently, considering how much wa- August, with little or no flow in other months. Eco- ter the ecosystem needs is not appropriate; instead, logical studies by the Sindh Forest Department the more relevant question is how much water is estimated that each 40 ha of mangrove forest re- needed to meet specific objectives, such as fish pro- quires 0.028 m3 sec-1 during July and August to re- duction targets. main healthy and to support the associated fisheries. For the estimated 260,000 hectares of mangroves, a For example, the Indus River dominates the land- total volume of 33.3 billion m3 would be needed. A scape and economy of Pakistan, providing water for typical flood hydrograph shape would suggest a flow the world's largest irrigated area. The lower Indus peak of 5,000 m3 sec-1. The floodplain forests need supports a large floodplain that provides timber to be inundated at least twice in 5 years to enable products, especially pit-props for the mining indus- saplings to become established. try. Where the river joins the sea, an extensive delta has formed, stabilized by mangroves. These pro- DEFINING FLOOD RELEASE OPTIONS vide camel fodder and fuelwood; support an exten- sive fishery, which earned $100 million in foreign Typically, a small number (3-5) of flood-release exchange in 1997; and provide habitat for many rare options will be identified and quantified, taking into species. Fresh groundwater within the delta sup- account the objectives, technical and financial con- ports fishing communities and camel herders. straints, and ecosystem requirements. Assessments should include descriptions of the magnitude, tim- Records from the 19th century suggest that fresh- ing, frequency, and duration of flood releases from water flows to the lower Indus were around 185 a reservoir to produce a target flood extent on a flood- billion m3 per year. Some flow occurred all year plain, as well as the impact on water reserves re- round, with higher flows starting in March, peak- tained within the reservoir for other purposes. The ing in August, and declining in November. Since overall aim is to optimize benefits from reservoir- then, construction of numerous dams and barrages based and flood-based livelihoods (Figure 1). This for irrigation has reduced the flows of water and step will almost certainly require a hydrological/ 13 WATER RESOURCES AND ENVIRONMENT · TECHNICAL NOTE C.3 hydraulic model of the floodplain, reservoir, main tion are greater than the flows that would occur channel, and catchment (Box 5). The "no-flood" op- naturally. This increase, combined with storage in tion should always be included in the analysis. Kafue Gorge reservoir, has created a larger perma- nently flooded area. The annual minimum flooded Appropriate flows during the non-flood season must area has increased from 300 km2 to about 1,500 km2. also be addressed. For example, the Kafue Flats The decrease in difference between dry and wet floodplain in Zambia is unusual in being directly season inundation has reduced the flood recession downstream of one reservoir (Itezhi-tezhi) and di- zone and also resulted in the loss of grazing, tradi- rectly upstream of another (Kafue Gorge). The joint tional flood recession gardening systems, and a de- operation of the dams significantly affects flooding cline in fisheries. in the flats. Under natural conditions, the flats used to empty almost completely by the end of the dry This highlights both the need for full understand- season, creating a large recession zone that was ing of the interaction between the entire hydrologi- important for agriculture, fishing, grazing, and cal regime and the ecological functioning of biodiversity (especially providing habitat for the rare the floodplain and, where more than one dam ex- lechwe antelope). Managed flood releases from ists, the importance of coordinating their opera- Itezhi-tezhi have meant little change in the maxi- tion to ensure that flows--including periods of no mum area flooded each year. However, dry season flow--achieve the objectives agreed between the releases from Itezhi-tezhi for hydropower genera- parties. BOX 5. FLOOD RELEASE OPTIONS FOR THE ANA T RIVER, KENYA The Tana River is the largest river in Kenya. It rises in the highlands of the Aberdares and Mt. Kenya, and then flows for more than 250 km over the flat, dry coastal plain to the Indian Ocean. The Tana has an extensive floodplain and delta, which support the livelihoods of over half a million people through agriculture, fishing, livestock rearing, and horticulture. To satisfy the increasing electricity demand within Kenya, planning is under way to dam the Tana up- stream of the floodplain at Mutonga (and Grand Falls) to supplement a cascade of five existing upstream dams. Recognizing the local and national value of the floodplain and delta, planners and developers have agreed that floods should continue. Consequently, the feasibility study designed a dam that will store enough water to produce a flood downstream through short-term high-flow releases, as well as produce the necessary electricity. Furthermore, recognizing that silt is as important as water for the maintenance of the productivity of the floodplain, the dam designers will try to release silt together with the flood water, although this is technically very difficult and may 1200 not be feasible. A major challenge is to release the Type A Desired Flood correct amount of water, such that when it combines Type B Type D with rainfall and tributary inflows between the reservoir 1000 and the floodplain, the desired level of floodplain inundation is achieved. 800 Using data from historical and recent floods, a model- /sec) 3 ing study is currently underway to address these (m limitations. The figure shows three possible flood release 600 hydrographs (A, B & D ­ broken lines) that vary in ge volume. Depending on the hydrological conditions in 400 the rest of the catchment, these can all produce the desired flood downstream (solid line). To ensure the Dischar optimum release, regional rainfall, soil moisture, and 200 flows from tributaries must all be monitored. Despite this complexity, it is anticipated that the final design will result in a dam that meets the demand for electricity 0 0 1 2 3 4 5 6 7 8 9 10 and respects the needs of its floodplain and depen- Time (days) dent livelihoods downstream. 14 ENVIRONMENTAL FLOWS: FLOOD FLOWS IMPACTS OF FLOOD OPTIONS influx of people into restored floodplain ecosystems. The adequacy of current laws can be evaluated Since a final flooding regime will inevitably be a though a legal impact assessment. Managed floods compromise between different objectives, it is im- may have positive or negative impacts on nutrition, portant to assess the impacts of the various flood- drinking water quality, injury, stress, communal vio- ing options, both for floodplain-dependent lence, and communicable diseases such as malaria, livelihoods and for reservoir-dependent livelihoods. schistosomiasis, and arboviruses. Thus, the health A range of impact assessments must therefore be impacts of flood options should also be considered initiated at an early stage, so that scoping of issues as part of the assessment. and collection of data can begin and possible mea- sures to mitigate negative impacts can be defined. The assessment must equally include the impacts In most countries, a formal EIA is a legal require- on the reservoir and its dependent activities. Re- ment. Impact assessment can take many forms, but leasing water to create a managed flood will nor- environment, health, and law are key areas. There mally mean less water for intensive irrigation, should also be direct and participatory assessment hydropower generation and/or water supply. These of impacts on livelihoods, using a broad definition impacts must be quantified. In some cases, legisla- of assets, including social capital. The criteria to tion designed to protect downstream communities evaluate livelihood outcomes should also be de- may prevent those responsible for the dam from fined with stakeholder participation. making managed flood releases. In addition, down- stream communities may have rights to use the Additionally, laws may need to be established to water, which would require compensation. regulate natural resource use, as there may be an SELECTING THE BEST FLOOD OPTION For each option, the monetary and non-monetary costs and benefits and their distribution among all stakeholders should be estimated (Box 6). National or local economic benefits should not be the only criteria on which to make decisions about managed floods. It is also vital to consider other political, so- cial, historical, and ecological issues, particularly the rights and welfare of those affected. In many Bank cases, the link between flooding and livelihoods will orld W, be imprecise and understood in only a qualitative Hirji way. A decision should then be made on the most afikR by appropriate option, using a transparent method in which stakeholders participate and understand the Photo Nyumba ya Mungu Dam, Tanzania results. 15 WATER RESOURCES AND ENVIRONMENT · TECHNICAL NOTE C.3 BOX 6. VALUATION IN THE HADEJIA-JAMA'ARE RIVER BASIN, NORTHERN NIGERIA In Northern Nigeria, an extensive floodplain exists where the Hadejia and Jama'are Rivers converge. The floodplain provides essential income and nutrition benefits in the form of agriculture, grazing resources, non-timber forest prod- ucts, fuelwood, and fishing for local populations. In addition, it helps to recharge the regional aquifer and serves as an essential groundwater source. The maximum extent of flooding has declined from 300,000 ha in the 1960s to around 70,000 to 100,000 ha more recently, following the construction of the Tiga and Challawa Gorge dams upstream. There are plans for a new dam at Kafin Zaki. Economic analysis of the Kano River Project (a major irrigation scheme benefiting from the upstream dams) and the floodplain showed that the net economic benefits of the floodplain (agriculture, fishing, fuelwood) were at least $32 per 1000 m3 of water (at 1989 prices). However, the returns from crops grown in the Kano River Project were at most only $1.73 per 1000 m3. When the operational costs are included, the net benefits of the project are reduced to $0.04 per 1000 m3. A combined economic and hydrological analysis was conducted to simulate the impacts of these upstream projects on the flood extent that determines the downstream floodplain area. The economic gains of the upstream water projects were then compared to the resulting economic losses to downstream agricultural, fuelwood, and fishing benefits. Given the high productivity of the floodplain, the losses in economic benefits due to changes in flood extent for all scenarios are large, ranging from $2.6 million to $24 million. As expected, there is a direct trade-off between increasing irrigation upstream and impacts on the floodplain downstream. Full implementation of all the upstream dams and large-scale irrigation schemes would produce the greatest overall net losses, around $20 million. These results suggest that the expansion of the existing irrigation schemes within the river basin is uneconomic. The introduction of a regulated flooding regime would reduce the scale of this negative balance substantially, to around $15 to $16 million. However, the overall combined value of production from irrigation and the floodplain would still fall well below the levels experienced if the additional upstream schemes were not constructed. IMPLEMENTATION DESIGNING ENGINEERING STRUCTURES example, to make the most of the flooding that does occur in the Hadejia-Jama'are River Basin (Box 6), The dam outlet structures will need to be designed a series of small embankments have been built by and constructed (for new dams) or may need to farmers to control the level of water in different parts be adapted (for existing dams) to allow necessary of the floodplain. Water movement is controlled by and feasible flood releases and sediment delivery. small, hand-operated sluice gates. Early in the flood- The position of outlets will need to take into ing season, mesh "gates" are installed, which al- account any thermal and associated chemical low water in but keep out fish that would eat the stratification in the reservoir. Embankments young growing plants. Once the plants are estab- may need to be constructed on the floodplain to lished, the mesh is removed to allow fish to spawn protect infrastructure--such as houses, roads, and and fry to grow in the flooded areas. As the flood factories--from flooding or to enhance, control, or subsides, wooden boards may be put in if flooding reduce inundation in selected areas. However, in- needs to be prolonged. frastructure is still likely to be affected by large natural flood events, irrespective of any managed Further associated structures may also be required, flood releases. such as fish ladders, sluice gates, or adaptation of the spillway to minimize disease vectors. Local Some infrastructure may also be needed to maxi- health centers may need to be stocked with drugs mize the benefits from the controlled releases. For before the beginning of each flood season. 16 ENVIRONMENTAL FLOWS: FLOOD FLOWS RELEASING THE WATER In many cases, it will be necessary to liberate sedi- ment from the reservoir as part of the flood re- Where possible, pilot flood releases of various sizes leases to maintain the structure of downstream can be made over a period of 2­3 years to determine rivers and deltas and the fertility of floodplains. the response of the ecosystem and of local commu- This will also help to prolong the life of the reser- nities. If project funds permit, it is best to install real- voir by sustaining its storage capacity (Box 7). time monitoring of both precipitation and river flows, However, the technique will be of limited value and link these data to flood forecasting models. This with hydropower dams, because it is not possible helps define the appropriate release to create the to pass high silt loads through hydropower tur- required inundation, thus optimizing water resources bines. allocation between reservoir-dependent livelihoods and flood-dependent livelihoods. In many cases, it MONITOR, EVALUATE, AND ADAPT will be possible to create the desired floodplain in- RELEASE PROGRAM undation by adding water from a reservoir to a natu- ral small flood rather than trying to create the entire flood from managed reservoir releases. Adequate ecological and socioeconomic monitor- ing will help to assess the effectiveness of flood re- Floodplain residents need advance warning about leases in relation to the objectives. It is preferable planned flood releases to enable them to move live- to decide at an early stage on the indicators of stock and other vulnerable assets to safe areas. Ide- whether the objectives are being attained or not. ally, the coordination mechanisms established Monitoring should be designed to differentiate be- during the planning phases to elicit stakeholder tween the changes to social and environmental con- views can be used to disseminate flood warnings. ditions caused by flood releases from those arising In addition, radio broadcasts, posters, and other from other factors. Since the time-scales for eco- media may be required to reach all floodplain com- logical and social change may be many decades, munities. Further awareness development and ca- monitoring should be continued for long periods pacity building exercises can be used to demonstrate with adequate funding. The flood release program the impacts of the new flood regime. When the flood can be modified once the effectiveness of the re- release notification program is completed, full re- leases becomes apparent. leases can be made. BOX 7. RELEASING SEDIMENT FROM ARBELA T DAM, PAKISTAN Pakistan's Tarbela Dam, completed in 1974 as a component of the Indus basin scheme, provides for both irrigation and hydropower. Of the 240,000 tons of sediment entering the reservoir annually, 80 percent is trapped. Unless action is taken, the steadily declining volume of the reservoir will mean that the facility has a severely diminished value by 2030. A technique to flush the sediment through the dam by lowering the reservoir level and releasing high flows for a sustained period has been tested recently. These studies have shown that it is possible to maintain a substantial long- term live storage with only a small annual reduction in stored water. The process works effectively when a low reservoir level is maintained and the period of flushing is several weeks. The live-storage volume at Tarbela would eventually settle at about half its original value. To use this flushing technique to extend the reservoir's life, it may be necessary to increase the capacity of the low-level flow release facilities at the dam. 17 WATER RESOURCES AND ENVIRONMENT · TECHNICAL NOTE C.3 A PRACTICAL EXAMPLE -- THE WAZA-LOGONE FLOODPLAIN The Waza-Logone floodplain in Cameroon is eco- The SEMRY rice scheme was not making full use logically and socially suited to flood releases. In this of water stored in Lake Maga. Using the water to example, flood releases from an existing dam were rehabilitate the floodplain, while retaining enough negotiated in a way that satisfied downstream com- to maintain rice production, was a development op- munities dependent on floodplains, as well as those tion. Hydrological, ecological, and socioeconomic groups who used the stored water for irrigation. studies suggested that such managed flood releases could improve fisheries, agriculture, and herding, The Logone River drains from the Adamoua pla- though much would depend on the available flood teau in central Cameroon. Before it joins the Chari discharges and the response of local people and their River, flood water from the river inundates annu- environment to renewed flooding. However, re-es- ally a large floodplain, originally around 6,000 km2. tablishing these flood flows could attract a large This wetland has a high biodiversity with large herds influx of people, who would over-exploit the area of giraffe, elephant, lions, and various ungulates. and thus degrade the rehabilitated resources of the Part of the floodplain is designated as the Waza floodplain. National Park, which attracts around 6,000 tourists per year. The Waza-Logone Conservation Project, coordinated by IUCN, undertook a wide range of activities in In the flood season, the entire floodplain becomes the floodplain and surrounding area over 8 years a vast fish nursery. Up to the 1960s, fishing was that highlighted many of the issues discussed in the primary economic activity among the local these Notes. The results of these activities are used Kotoko people, who could earn $2,000 in four to exemplify the design and implementation of man- months. The Fulani name for the floodplain is aged flood releases. yaérés, which means "dry season pasture." Annu- ally, some 300,000 cattle and 10,000 sheep and goats PRELIMINARY STEPS grazed on the floodplains. Pastures become acces- sible when surrounding savanna grasses withered and their protein content was depleted. The carry- The objectives were to restore biodiversity and the ing capacity has been estimated at 1­2 cattle per livelihoods of communities--who depend on natu- ha, compared with 0.2 for surrounding savannah. ral resources of the floodplain--through managed Floating rice was the main arable crop of the flood- flood releases, while retaining sufficient water in plain, since it has low labour demands and fits well the reservoir for intensive rice irrigation. These into the fishing calendar. Yields were not high, but objectives were defined through analyses of the deg- enough land was available to ensure self-sufficiency radation of the floodplain ecosystem and of the ef- in rice. ficiency of the SEMRY irrigation scheme. The area inundated has been reduced since 1979. Technical studies were undertaken to assess the This is partly due to climatic factors, but primarily capacity of the outlet structures of the dam, the chan- due to construction of a dam across the floodplain, nels leading from it, and the river embankments. which created Lake Maga. The lake supplies the Because the reservoir is shallow, thermal and SEMRY II irrigation project, where rice is grown. chemical stratification was not a problem. In addi- (The dam is adjacent to the river and thus does not tion, the reservoir is "off-line" and does not control stop all flow). There is insufficient flooding to grow the entire flow of the river, so sediment release was floating rice in large areas of the floodplain, and not an issue. Natural inundation of the floodplain fish yields fell by 90 percent. resulted from a series of three processes: (1) rain falling directly onto the floodplain; (2) runoff from 18 ENVIRONMENTAL FLOWS: FLOOD FLOWS local streams entering the floodplain; and (3) flows ply (wells), sanitation (latrines), and hygiene aware- in the Logone River exceeding the capacity of its ness programs, which led to reduced diarrheal dis- channel. eases. The team also assessed the impact of re-flooding on the operation of the SEMRY rice DESIGNING FLOOD RELEASES scheme. Further studies of the health and environ- mental impacts of the full-flood option are planned. The project developed a participatory process that The team also conducted an economic analysis of involved relevant national and local authorities, various flooding options. Re-flooding of 90 percent floodplain communities (using Participatory Rural of the floodplain yielded annual values of $550,000 Appraisal), and SEMRY management. Local inter- for fisheries, $930,000 for herding, $31,500 for re- est groups--concerned with, for example, fisheries cession agriculture, and $60,000 for tourism. The and small-scale rice farming--were established net present value over 30 years was 15,000 million within local communities. The next step was to es- CFA (about $20 million), assuming a discount rate tablish a water committee with local community of 4 percent. representatives, the SEMRY rice scheme manag- ers, Waza National Park staff, and local authorities. Technical expertise was developed in local institu- IMPLEMENTING THE CHOSEN FLOWS tions by involving their staff in the project. Changes to the embankments along the main river The project team conducted detailed studies of the and channels feeding the floodplain with water were interaction between vegetation and flooding. These planned and implemented in collaboration with showed that there was a change from annual to local communities. Modifications to the outlet struc- perennial grasses with inundation over a three-year tures of the dam were not required. No embank- period. The surveys were continued as part of the ments were necessary to protect the SEMRY rice pilot releases. scheme from flooding. A hydrological model of the reservoir, floodplain, Pilot flood releases were conducted in 1994 and 1997 main river, and local tributaries was constructed to determine the response of the ecosystem and of and used in conjunction with ecological and socio- local communities. Further test releases from the economic models to define flooding options under dam are planned. Hydrological, ecological, socio- dry, wet, and very wet conditions. logical, and economic monitoring was established for the pilot releases to assess the impacts of flood- A preliminary Environmental Impact Assessment ing and the results were used to refine the models of all flood options was undertaken. Studies of wa- and flooding options. ter-related diseases focused attention on water sup- 19 WATER RESOURCES AND ENVIRONMENT · TECHNICAL NOTE C.3 CONCLUSION Most river systems throughout the world will have in developing countries and sport fishing in devel- some element of their ecosystem--such as fish spe- oped countries. In principle, therefore, most reser- cies that breed on the floodplain or vegetation that voirs have a potential for managed flood releases. requires periodic watering--that has evolved to ben- efit from floods. In some cases, the species or land- The appropriateness of managed flood releases will scapes (particularly those that are rare or endan- depend on the relative merits of using water for flood gered) may be worth conserving for their own sake. releases compared to storing it in the reservoir for In many other cases, local communities will have irrigation, public/industrial supply, or hydropower. developed livelihoods that depend on these floods. This decision can only be made on a case-by-base These dependencies include subsistence agriculture basis. FURTHER INFORMATION The following reports provide background infor- Articles on the subject include: mation on managed flood releases: Junk, W.J., P.B. Bayley, and R.E.Sparks. 1989. "The flood Acreman, M.C. 2000. Managed flood releases from reser- pulse concept in river floodplain systems." In: voirs; issues and guidance. World Commission Dodge, D.P., ed. Proceedings of the International on Dams Thematic Review 11.1 Dams Ecosys- Large Rivers Symposium. Honey Harbour, tem functions and environmental restoration. Ontario, September 1986. Canadian Journal of World Commission on Dams. Down-loadable Fisheries and Aquatic Science, special publica- from www.dams.org or on CD from Earthscan, tion, 106: 110­127. London, www.earthscan.co.uk McCartney, M.P., and M.C. Acreman. 2001. "Managed flood Acreman, M.C., and G.E. Hollis. 1996. Water management releases as an environmental mitigation option." and wetlands in sub-Saharan Africa. Gland, Swit- International Journal of Hydro-power and Dams. zerland and Cambridge, UK: IUCN. 8 (1) : 74­80. Petts, G.E. 1984 Impounded rivers. Chichester, UK : John Scudder, T. 1991. "The need and justification for main- Wiley & Sons. taining transboundary flood regimes: The Africa case." Natural Resources Journal, 31: 75­122. Welcomme, R.L. 1976. "Some general and theoretical con- siderations on fish yields of African rivers." Jour- nal of Fisheries Biology, 8: 351­364. 20