Scientific Data for Decision Making Toward Sustainable Development: Senegal River Basin Case Study: Summary of a Workshop (2003)

The participants at the workshop addressed a broad range of issues concerning the human, natural, and constructed environment in the Senegal River basin (SRB), with particular focus on the problems created by the Diama and Manantali dams. They identified and to some extent characterized the many different kinds of data that have either already been collected or that are still needed to support decision making toward sustainable development of the SRB. The presentations covered four broad topical areas.¹ The first three addressed environmental, health, and SRB socioeconomic issues and related data aspects, while the fourth focused on issues concerning data for decision making about dam projects.

The SRB stretches from the river’s source in mountains of Guinea to the coastal and ocean zone at Saint-Louis, Senegal. Nutrients and sediment generated in the headwaters are recycled downstream, driving plant and biotic productivity. The lateral connectivity between river and floodplain also drives river life. The appearance of aquatic weeds, sedimentation at Diama Dam, and health

problems at Richard Toll upstream from the dam underscore the need for regional understanding of these processes.

As noted in Chapter 2, the annual flood at Bakel has peaked in August and September, coming almost entirely from the upper basin. There has been a dramatic reduction of flow in the last 20 years, however, to less than half the 100-year average of 700 m³/s. The total mean annual discharge is 21,000 million³. In the middle valley there are 72 minor floodplain basins of 1,000 ha to 15,000 ha supporting traditional flood-recession agriculture and also serving as fisheries breeding areas. Erratic flows and episodic inundation in the river contributed to a wide diversity of floodplain habitats and species, which enabled a variety of food production systems by middle-valley communities. The high variability of rainfall, however, prevented the sustained use of basin resources. The OMVS, therefore, was created in 1972, proposing 375,000 ha of pumped irrigation, navigation from the ocean to Kayes, 800 GWh/yr electricity 9 years out of 10, flood mitigation, maintenance of flood recession agriculture in transition to irrigated agriculture, and control of saltwater intrusion at Diama.

The OMVS program resulted in construction of the Diama Dam by 1986, the Manantali Dam by 1987, and inauguration of the Manantali hydropower unit and transmission line to Bamako by 2001. Improvement of navigation facilities has not been implemented. Irrigation development has been slower than planned, with only 131,000 ha irrigated by 1998, and only half of that area being cropped on average. The major beneficiary of the Manantali and Diama dams was the development of irrigated sugarcane at Richard Toll by the Senegalese Sugar Company, based on a steady water supply above Diama Dam.

Inadequate agricultural and health planning, however, resulted in a major crisis in health and nutrition beginning in 1987, especially the schistosomiasis epidemic at Richard Toll, which is discussed in more detail below. The Manantali and Diama dams also reduced the variety of ecosystems in the valley, benefiting monocultures such as sugarcane but also resulting in aquatic weed nuisances and water-associated disease vectors. Embankments on both sides now control flooding along the Diama Reservoir up to Dagana. Irrigation systems are labor intensive, however, leaving little time for traditional crops that previously were the major nutrition source for most households. Government-supported rice production has used up farmers’ cash income, which also had provided variety in diet. Malnutrition is most noticeable among women and children, as well as ethnic minorities. Overall, the basin ecosystems and production systems are now threatened by decreasing productivity because of inadequate resource management, including deforestation, soil erosion, overgrazing, and desertification. Species diversity has been reduced, along with elimination of wetlands by diking and expansion of irrigated areas. There are enormous social costs including malnutrition, disease, civil unrest, and social conflicts. Morbidity and death are increasing as a result of malaria, schistosomiasis, and diarrheal diseases, which were formerly held in check by the annual dry season.

Opportunities exist for capacity building and regional cooperation in monitoring of riparian ecosystems and agricultural productivity. Consequent improvement in environmental management is the key to maintaining political and economic equilibrium in the SRB. Action should be initiated on (1) improving regional cooperation on water resources and environmental management, (2) data collection on hydrology and sediment regimes, (3) developing small hydro-dams and irrigation systems, and (4) wildlife and biodiversity conservation. Available data are either confined to national sub-basins or to single resources. Historical records on flow and rainfall exist from Saint-Louis up to Bakel, but recent changes in rainfall and flow are not yet included in the planning process. General ecological data for the entire watershed are needed, which would make use of remote sensing, cartography, and ground monitoring, especially in the upper basin. Further work should be focused on the needs for development of the SRB riparian communities.

The International Program of the U.S. Geological Survey’s Earth Resources Observation Systems (EROS) Data Center participates in several research and development activities in the countries of the Senegal River basin. The Famine Early Warning System Network (FEWS NET) program, discussed in further detail below, has secured and makes available from its Africa Data Dissemination Server several core data sets for the countries. These are maintained in the EROS Data Center archive, are freely available, and are routinely updated. The program also generates several dynamic data sets that are presented in near real time and that include spatially distributed rainfall estimates, water runoff estimates, a normalized difference vegetation index, and crop-water saturation indexes. These are also accumulated in the archive and are available for some online access, processing, and downloading.

A current project on carbon sequestration in Senegal is generating estimates of carbon stocks and fluxes in three defined geographic areas: Velingara, Bambey, and Podor. These are defining the biophysical potential for sequestration as impacted by climate and land-use management, and the socioeconomic incentives for various management practices.

Extensive work in Senegal has provided documented information on natural resources, land cover, and land-cover change over periods covering as much as 60 years. These data are both in the form of reports and digital maps. Conversions of gallery forests, for example, along the Senegal River are clearly docu

mented. Techniques to evaluate changes in the performance of land cover (e.g., greenness level, start of growing season) at the pixel (1 km) level have been developed. These identify temporal trends in land-cover performance associated with land degradation (e.g., near boreholes or large cities) or spatial anomalies, which suggest degradation or improvement, associated with conserved areas and other management practices. These anomalies are verified with high-resolution satellite imagery or ground observations.

The EROS Data Center has developed clearinghouse systems for the identification, access, and distribution of spatial data. In the SRB countries the Centre de Suivi Ecologique has already received the appropriate software and is initiating this system consistent with the standards established by the International Organization for Standardization.

Weather and climate are essential concerns of human existence. Data about weather and climate describe and quantify variable resources such as wind, rain, and temperature that are shared by everyone and are sometimes contested. The outcome of policies adopted by governments as well as individuals in managing food production, water supplies, natural ecosystems, and health services can be heavily influenced by the quality of available weather and climate information. As decision making and concepts of sustainable development look further into the future, the assembly and analysis of weather records and the creation of information products become more complex and more important. Weather and climate information can help in monitoring and assessing many environmental parameters and problems, including agricultural production and fresh water supplies. Global and regional data provide means to check the quality of local data, place local climate patterns into a larger context, and lend confidence and credibility to analyses that are important to environmental effects in the Senegal River region.

Observations and records of weather variables important to Senegal extend across the country and well beyond its borders. Accordingly, while observation and data management systems within the country should be improved, they must also comply with regional and international collaboration. New sources of data are being added all the time, such as the ones at the EROS Data Center. However, these sources must be integrated into the existing methods of analysis, and the influence of these new data sources on the traditional analysis schemes must be understood well enough to engender confidence by those who would apply them.

Meteorological and climatological data are available on the World Wide

Web from many sources. For example, weather data can be obtained from Web Africa, the U.S. National Weather Service, and the European Centre for Medium-Range Weather Forecasts. Information relating to climate impacts can be obtained from the Intergovernmental Panel on Climate Change, which was established by the World Meteorological Organization and the U.N. Environment Programme, as well as from the U.S. National Weather Service’s Climate Prediction Center, which hosts the Global Data Display System. Seasonal forecasts for Africa can be obtained from the International Research Institute for Climate Prediction.

The SRB may be a good candidate for the International Watershed Research Network. This network, which is in its conceptual phase, has as its objective to advance understanding of watershed processes, and draws upon shared expertise from watersheds around the world.

The data set available at NOAA for climate monitoring over the Senegal River basin includes in-situ and satellite-derived data and information. The in-situ data are received daily through the World Meteorological Organization’s Global Telecommunications System (GTS). These data are fed into NOAA’s databases to construct climatology information based on several parameters, including rainfall, temperature, and wind, and to derive near-real-time anomalies for those parameters at time scales ranging from 10 days to monthly and seasonal. Satellite-derived rainfall estimates are produced using the newly developed NOAA satellite rainfall estimation technique (RFE). RFE is primarily based on (1) GTS rain gauge measurements; (2) the Geostationary Operational Environmental Satellite (GOES) Precipitation Index, a technique that derives rainfall estimates from fractional cloud coverage colder than 235 K obtained from Meteosat infrared data; (3) Special Sensor Microwave/Imager (SSM/I) rainfall estimates; and (4) Advanced Microwave Sounding Unit (AMSU) rainfall estimates. RFE is produced operationally at NOAA’s Climate Prediction Center for the USAID Famine Early Warning System (FEWS) in support of weather- and climate-related hazard assessments in Africa.

Difficulties are encountered in obtaining local in-situ data, primarily because of telecommunication problems. NOAA’s database contains long-term historical data for only six rain gauge stations over the SRB. A new approach for more dense data coverage and to address the telecommunication issues is being developed.

Correlations between Sahel rainfall and global SST are being developed for

predictive purposes. Other prediction methods are also being investigated at NOAA and many other world centers for climate forecasting. This has led to the development of Climate Outlook Forums,⁶ which were initiated by NOAA in 1997. These forums have brought together both forecasters and users of climate information not only to bridge the gap between them but also to issue a readily available consensus forecast, which can be interpreted easily by the users. Forecasts are given in above normal, normal, and below normal probabilistic categories, and various types of decision makers are now involved in the process.

Worldwide, flooding causes loss of life and extreme damage to property on an almost annual basis, and Africa is no exception. In 1999 floods in the Senegal watershed and many other parts of West Africa killed several hundred people and left thousands homeless. The February-March 2000 floods in the Limpopo watershed cost more than 900 lives and caused extensive damage to infrastructure, while the 2001 floods in the Zambezi watershed were responsible for more than 100 deaths. Although intense flooding often causes much damage, floods also provide the main source of livelihood to many people through flood-recession cropping, which is how the Senegal River valley below Bakel was farmed prior to the construction of the Manantali Dam. With such a paradoxical scenario the ability to tell in advance the nature of impending floods becomes necessary, so that any required preparatory action can be taken. Thus, the ability to accurately predict the timing, spatial extent, and volumes associated with a flood event is a fundamental aspect of floodplain management. Toward this end, FEWS NET through its implementation partners has developed a flood modeling and prediction tool, the FEWS Stream Flow Model (SFM), to allow the timely provision of early warning information on flooding in Africa. This tool is now fully operational for the Limpopo watershed and is currently being calibrated for other watersheds in Africa.

While the SFM provides valuable estimates of stream flow (and hence river stage), a second crucial component of flood information pertains to the spatial patterns and extent of flooding, an issue that can be addressed using fine-resolution satellite images. The Global River Floodplains (GRF) project has developed

a database identifying remote sensing images for many large rivers in the world. The National Aeronautics and Space Administration Moderate-resolution Imaging Spectrometer (MODIS) and Landsat images identified through the GRF database have been processed and analyzed to derive maps showing the extent of inundation and patterns of sediment concentration for significant flood events in the Zambezi and Limpopo watersheds. Analysis of these satellite images shows where local water that accumulates on a floodplain before overbank flooding occurs contributes significantly to the patterns and extent of inundation, thus helping to improve the accuracy of the predictions from the FEWS SFM model.

Initial work on the Senegal watershed involves an analysis of the flood events over the past 25 years, the morphometric characteristics of stream order, and the drainage density patterns within this watershed. In particular, the characteristic temporal and spatial patterns of drought and flood within the Senegal watershed were derived using results from U.S. Geological Survey climate modeling and are compared to the selected remote-sensing images.

Environmental models require a large number of data sets to run and test the models. The most critical data include (1) soil physical data, (2) atmospheric weather data, (3) vegetation characteristics, and (4) current and historical land-use data. Although the data needs for the environmental models are model specific, there are standard data sets used by most environmental models. The time step of the models has a big impact on the data requirements of the models. Monthly time-step models typically require (1) monthly total precipitation and monthly averaged daily maximum and minimum temperatures as atmospheric drivers, (2) soil texture, bulk density and field capacity, and wilting point as required soils data, and (3) a general description of land cover and land-use management. Daily time-step environmental models require more detailed information about the atmospheric driving variables (daily total precipitation, daily maximum and minimum air temperature, wind speed, relative humidity, and solar radiation) and soils data (e.g., saturated and unsaturated hydraulic conductivity, and water potential versus soil water content).

Fortunately a large number of environmental data sets are available at the local, regional, and national level for the important atmospheric driving variables, soil description data, and land-cover data. For instance, global data sets at the 0.5 * 0.5 degree scale are available for daily and monthly atmospheric driver data, soil-texture data, and land-cover and use data. More detailed soil physical data can be derived from the soil-texture data, and a coarse historical analysis of

changes in land use has recently been developed. The atmospheric driving data, soil characteristics, and land-use data are available on a finer spatial scale (e.g., 30 * 30 m², 5 * 5 km²) for select regions of the world.

Regional data sets are available to describe the spatial patterns of the major environmental variables for the SRB. These data sets describe land use, climatic variables, and soils data needed to drive ecosystems models. The data sets have been interpolated to the 0.5 * 0.5 degree resolution at the global scale and are available from the Century Ecosystem Modeling group:⁹ (1) 1900-1998 monthly precipitation, (2) 1900-1998 monthly average maximum and minimum air temperature, (3) potential natural land-use patterns, and (4) soil-texture description (e.g., sand, silt, clay content). These driving variables have been used to simulate the global patterns of the major ecosystem variables using the Century ecosystem model.¹⁰ The patterns of the Century model output variables have been simulated for the past 100 years using weather data sets that are interpolated to the 0.5 * 0.5-degree global scale. There are over 300 output variables from the Century model available to represent the status of the major ecosystems variables. Some of the important ecosystems variables include annual and average total plant production, soil carbon and nitrogen levels (0-20 cm soil depth), plant production for the tree and grass components of the ecosystem, stream flow, actual evapotranspiration, and soil nitrogen mineralization rates. These output variables are available at monthly, annual, or long-term averaged time scales for the SRB region for the last 100 years.¹¹

Since the great West African drought that ended in 1974, concerns about health impacts of water resources development have been a high priority in the SRB. The first dam constructed in the basin was at Foum Gleita in Mauritania on the Gorgol River, the only tributary of the Senegal River in Mauritania. Health data on water-associated diseases were collected specifically for this dam by Blue Nile Associates in 1974 prior to construction, and a few health surveys were carried out several years after construction.

Planning for the Manantali Dam on the Bafing River tributary in Mali and

Diama Dam near the delta included several health surveys conducted for OMVS at the request of the World Bank, which was financing both projects initially. The largest study was sponsored by USAID in 1978 and conducted by Gannett Fleming. There were pre- and post-construction surveys at Manantali by the Institut Nationale de Recherche en Santé Publique, or the National Institute for Public Health Research, of Bamako, as well as surveys by Blue Nile Associates and the Water and Sanitation for Health Project of USAID. A limnology unit of the OMVS stationed at Manantali Dam started collecting data on disease vectors in the reservoir soon after construction. Disease epidemics around Diama Dam began the year after the dam was completed, and a large number of surveys on water-associated diseases have been conducted around Diama by local and foreign universities and agencies almost every year since then. The Onchocerciasis Control program of the WHO collected data on river blindness and its spread by blackflies in the upper basin until the disease was controlled, which occurred around about 1990. USAID sponsored a study on AIDS in the valley in 1995. A large amount of clinical data on all diseases is found in ministries of health units of all three countries along the river, with the most extensive being in the Saint-Louis and Rosso areas. The Senegalese Sugar Company maintains its own clinic and medical records in Richard Toll as well. However, records found in various ministry offices in capital cities are of little use in epidemiologic evaluations of disease impacts of the river development. Finally, the WHO has sponsored a multitude of disease surveys along the river, starting with Watson’s survey of 1970. These may be found at the WHO country offices but are no longer in the central library in Geneva.

Schistosomiasis control has been implemented successfully in certain areas and in many countries over the past 20 years. However, the number of people infested has not changed substantially and disease burden remains high. The potential for transmission has increased due to water resource developments, with many more people living in endemic areas. About 600 million people live in schistosomiasis areas, of whom approximately 200 million are infested, 120 million symptomatic, and 20 million with severe disease; several hundred thousand die each year from urinary and intestinal complications.

Over the same period, the UNDP/World Bank/WHO Special Programme for Research and Training in Tropical Diseases (TDR) invested in research to develop tools for the control of schistosomiasis and other tropical diseases. This re

search ranged from basic sciences to the “proof of principle.”¹⁴ TDR and its partners evaluated the effectiveness of drugs, validated a range of diagnostic techniques, and investigated measures of schistosomiasis morbidity. TDR has also built capacity for scientists from disease-endemic countries to participate in this research. The reintroduction of a disease perspective and the strengthening of implementation research in the strategy of TDR has resulted in the reassessment of the research needs in schistosomiasis. While new tools for the control of schistosomiasis are necessary, more work needs to be done to optimize the use of existing tools in resource-limited environments and to formulate new strategies for control. The TDR strategic direction on schistosomiasis research is focused on this objective until an extensive review by a Scientific Working Group in 2005. WHO also supports related data collection, database development, and data access.

The officially commissioned Environment and Health Impact Analysis of the planned Senegal River dams did not anticipate the schistosomiasis problem. This was because the people who did the assessment were experts in snails and schistosomiasis but not dams. Neither were other dams (e.g., Aswan, Upper Volta) considered in the assessment. By 1994 a significant proportion (more than 90 percent) of the population living along the Diama reservoir was infested. Several surveys conducted in communities near Richard Toll showed that virtually everyone above five years of age had S. mansoni infestation. Schistosomiasis has been controlled in China, Brazil, and North Africa but not in the Sahel.

Malacological fauna, especially snails, are the intermediate host of human and animal trematodosis (a parasitic flatworm). This has been studied in the SRB, particularly in the delta and Lac de Guiers areas. Surveys conducted from 1978 to 1980 identified the potential of natural snails in the transmission of trematodosis. Among these snails are five species of Bulinus: B. truncatus, B. forskalii, B. senegalensis, B. umbilicatus, B. globosus; one species of Biomphalaria: Biomphalaria pfeifferi; and one species of Lymnaeidae, Lymnaea na talensis. Bulinus truncatus was the most widespread and abundant snail. The B. senegalensis and B. forskalii snails were less extensive and had lower densities. B. umbilicatus had restricted distribution in the middle valley, whereas

Biomphalaria pfeifferi, Lymnaea natalensis, and B. globosus were found very sparsely distributed in some sites with small populations.

Ecological changes, particularly physical and chemical changes in the water environment, resulted from the opening of the Diama and Manantali dams. The more important changes for the ecology of the snails are the prevention of the intrusion of sea water into the Senegal River, the regularization of the level of water in the river and other effluents in the delta, and the pH of the water becoming more alkaline. The combination of these factors has provided new favorable conditions facilitating the growth, spread, and increase of snails.

Malacological surveys that were carried out in the delta and Lac de Guiers from 1989 to 1996 revealed very large changes in the distribution and abundance of snails. The main signs of these changes were the rapid proliferation of Biomphalaria pfeifferi and Lymnaea natalensis molluscs, which seemed to have disappeared in 1977, while populations of Bulinus globosus had colonized more sites in the delta and Lac de Guiers with high densities. The other Bulinus had remained stable. These changes in ecological conditions have allowed these species of snails to flourish by colonizing more habitats, increasing substantially their population size, and extending their distribution. Another remarkable finding in the recent studies was the colonization of the Senegal River by the snails Biomphalaria pfeifferi, Bulinus globosus, and Lymnaea natalensis. Generally, these intermediate host snails are reported to occur only scarcely in large rivers like the Senegal River.

Because the optimal ecological conditions favorable to the growth and increase of snails have been created, it is easy to imagine the consequences of this situation in the epidemiology of snail-borne diseases in the SRB. The first outbreak of intestinal schistosomiasis in humans was in the delta town of Richard Toll in 1988-1989.

After Diama Dam became operational and many hydro-agricultural activities were set up, an increase in animal trematodosis was observed in the SRB as well. It concerned in particular Fasciola gigantica, Schistosoma bovis, S. curassoni, and Paramphistomum sp. The cattle trematodosis epidemiology showed itself in the disease prevalence increase in existing foci (Richard Toll, Ross Bethio, Mbane, and Keur Momar Sarr). Infestation rates in cattle increased from 11 percent to 27 percent, 20 percent to 30 percent, and 15 percent to 27 percent for fasciolosis, paramphistomosis, and schistosomiasis, respectively. In small ruminants, which seemed spared, 2 percent to 62 percent fasciolosis prevalence rates were recorded, whereas they were 25 percent to 30 percent for paramphistomosis. In a parallel manner new trematodosis foci appeared, starting in 1989 to 1990:

1. At the delta area in Tilene, Pont Gendarme, and Takhembeut with 3 percent to 20 percent, 4 percent to 20 percent, and 5 percent to 36 percent prevalence rates for fasciolosis, schistosomiasis, and paramphistomosis, respectively; and

2. At the Lac de Guiers area in Temeye, Thiago, and Senda with 5 percent to 86 percent, 5 percent to 11 percent, and 5 percent to 33 percent prevalence rates for fasciolosis, schistosomiasis, and paramphistomosis, respectively. In these new foci 2 percent to 55 percent and 5 percent to 25 percent prevalence rates were recorded in small ruminants for fasciolosis and paramphistomosis, respectively. Schistosomiasis was not as frequent, with 2 percent to 4 percent prevalence rates. This new trematodosis epidemiology after the opening of dams was remarkable by its very high infestation rates and parasite burdens, and by a polyparasitism that combined Fasciola gigantica, Schistosoma bovis, S. curas soni, and Paramphistomum sp.

This atlas presents the results of a temporal and spatial analysis of the relations between population factors, agricultural land use and performances, nutrition, and land degradation. The geographic area included in this project covered four CILSS member countries: Burkina Faso, Mali, Niger, and Senegal. Using a GIS, this study shows the interplay of these indicators at different scales (sub-regional, national, and local [first administrative level]). It represents a contribution to structural analysis of vulnerability and is a first step toward a decentralized (at the district or at village-clusters levels) decision-support system for an integrated strategic plan linked to an early warning system targeting poverty alleviation in the SRB.

The methodology for developing the atlas was based on the integration of multiple sources of data on population, including censuses, West African survey on migrations and urbanization (1988-1992), agriculture (national statistics services and projects, AGRHYMET), infant malnutrition (Demographic and Health Surveys, United Nations Children’s Fund), and land degradation (Global Assessment of Human Induced Soil Degradation Web site). The database was developed using Excel and mapping with GIS software called ArcInfo. However, substantial difficulties arose in obtaining some of the data.

The atlas produced a number of usual and also innovative indicators for the 1984-1985 to 1996-1997 period, although not systemized for the SRB. These data included:

1. density levels and migratory rates and surface of cultivated land (per rural habitant, rural worker); production (per rural habitant, rural worker) of the different crops; and land degradation.

2. migratory rates and level of intensification per worker according to the type of rain-fed crops.

3. length of the growing season and carrying capacity (1985-1997 period) per hectare-surface; and rural worker.

4. comparisons of the rate of coverage of the household’s caloric needs by the local production and actual level of stunting and wasting.

The atlas still needs additional data and analytical improvements. Data collection needs include a survey on individuals’ life-cycle events (family, activity/employment, land cultivation and tenure, migration); village community surveys on land-cover/land-use issues, history of production and social infrastructures, and markets; and remote sensing (1-m resolution) on land use and land quality (wind and water erosion sensitivity, salinization) changes. Information technology needs include Web-based research and mapping.

The atlas was first disseminated at the Conference of CILSS ministers and summit of CILSS heads of state at Bamako, Mali, in November 2000. An atlas Web page is also under construction (see www.insah.org).

The World Commission on Dams (WCD) was established in 1997 under the auspices of the World Bank and the World Conservation Union with a three-pronged mandate of (a) undertaking a rigorous and independent review of the development effectiveness of large dams; (b) assessing alternatives; and (c) proposing practical guidelines for future decision making. The WCD carried out its activities from May 1998 to November 2000.¹⁸

Regarding the first prong, the WCD conducted the first comprehensive study on the impacts of dams, but was unable to declare a “final verdict.” Lack of both baseline and impact data was one of the major constraints to arriving at a definite

conclusion on the development effectiveness of large dams. The same applies to the second prong of the WCD mandate—the non-dam options for delivering water and energy services. Regarding the third prong, the WCD recommended that the planning process related to water and energy development be moved upstream, starting with a needs assessment and/or validation and followed by an options assessment. These two critical stages should take place before a dam emerges as the preferred option. A good quality information base needs to be established at the national and river basin level as a long-term effort, whether a dam is planned or not.

Once an intervention is selected the subsequent decision-making stages are project planning, construction, and operation. At these stages the WCD recommended systematic baseline feasibility and impact studies, with an emphasis on such parameters as costs (financial and economic, environmental, social, health, cultural heritage), benefits, and distributional aspects that deal with the sharing of impacts.

Overall, the WCD report proposed a highly data-intensive, decision-support system. Many of the 26 WCD guidelines relate to data collection and analysis: stakeholder analysis, strategic impact assessment, project-level impact assessment, multicriteria analysis, life-cycle assessment, assessment of greenhouse gas emissions, distributional analysis, valuation of social and environmental impacts, risk assessment, baseline ecosystem surveys, baseline socioeconomic surveys, and environmental flow assessment. Because of its heavy reliance on a good information base, some critics have questioned the adaptability of the WCD proposed decision-making framework to the context of developing countries. The question however is: Is there an alternative to a good information base for informed decision making?

As a follow-up to the WCD report the UNEP Dams and Development project “promotes dialogue on improving decision making, planning and management of dams and their alternatives based on the WCD core values and strategic priorities.”¹⁹

Despite the growing literature on water and conflict in international river basins, little empirical work has been done to bolster common conclusions that

are so widely reported. In order to address this gap Wolf, Yoffe, and Giordano assessed all reported events of either conflict or cooperation between nations over water resources over the last 50 years and used these events to inform the identification of basins at greatest risk of dispute in the near future (5 to 10 years). The study was divided into two components:

1. Compilation and assessment of relevant biophysical, socioeconomic, and geopolitical data in a global geographic information system, and use of these factors to determine history-based indicators for future tensions along international waterways; and

2. Using these indicators, identification of basins at risk for the coming decade.

In general the study found that most of the parameters regularly identified as indicators of water conflict are actually only weakly linked to dispute. The institutional capacity in a basin, however, whether defined as water management bodies or treaties or generally positive international relations, are as important, if not more so, than the physical aspects of a system. It turns out, then, that very rapid changes, either on the institutional side or in the physical system, are at the root of most water conflict, as reflected in two sets of indicators: (1) “internationalized” basins (i.e., basins that include the management structures of newly independent states) and (2) basins that include unilateral development projects and the absence of cooperative regimes. By taking parameters of rapid change as indicators—internationalized basins and major planned projects in hostile and/or institution-less basins—the study was able to identify the basins with settings that suggest the potential for tensions in the coming 5 to 10 years. The study then identified “red flags,” or markers, related to these indicators, so that monitoring in the future might continue to help identify targeted regions for cooperation.

The lack of comprehensive scientific data and access to existing data is only one barrier to responsible decision making in formulating a sustainable development plan for the SRB. As noted earlier, the objectives of the OMVS in building the two dams on the river was to improve navigability, provide irrigation water

for more intensive and reliable agriculture, and produce hydroelectricity. Since the dams’ construction these goals have proved more difficult to attain than originally anticipated. Many negative effects of the dams’ construction were either unanticipated or underestimated. In the SRB “sustainable development” was heavily weighted toward technological improvements that would help propel Senegal, Mali, and Mauritania into the first world, with relatively less regard for the impact on indigenous social values and structures. As new data have emerged on the ecological, health, social, and cultural effects of the dams’ creation and operation, effective policy making and reevaluation remain encumbered by the original paradigm—modernization through technological development. An optimal decision-making process ideally would be inclusive of all uses of the river, from headwaters to the ocean, and fully take into account the input of all stakeholders. The challenge of sustainable development of natural resources is to implement adaptive management, responding to changing conditions, priorities, and information. That such changes have not yet occurred in the SRB’s management illustrates one of the difficulties of natural resource management: Decisions are rarely made on the basis of data alone.