Scientists map West Africa’s dams and reservoirs in unprecedented detail

Across West Africa, thousands of dams and reservoirs regulate water for drinking, irrigation, and hydropower, yet the information describing this infrastructure has long been scattered, inconsistent, or simply missing. A team of researchers from Benin, Côte d’Ivoire, and Germany has now compiled the most complete harmonized dataset of dams and reservoirs ever assembled for the region, and tested it against observations collected on the ground. The work, published in the journal Earth System Science Data, catalogues 1,429 georeferenced dams and 1,258 reservoirs, more than any existing global database reports for West Africa, and reveals both the promise and the persistent blind spots of the world’s water infrastructure maps.

The motivation for the study stems from a fundamental problem in water science: global databases disagree wildly about how many dams and reservoirs exist in West Africa, and none of them has been rigorously validated with field data from the region. Existing inventories suffer from inaccurate georeferencing, inconsistent accessibility, heterogeneous records, and a lack of validation against observations. Because researchers and practitioners in data-scarce regions often rely on global datasets, these shortcomings propagate directly into hydrological models, climate impact assessments, and investment decisions. The consequences are not academic. Inaccurate estimates of storage capacity or reservoir distribution can misguide water allocation decisions, fuel conflicts among users, and undermine projections of future water availability under climate change.

To build the new dataset, the team first conducted a systematic literature review using scholarly databases, identifying nineteen datasets containing West African dam and reservoir information. From these, they selected twelve open-access sources with standardized file formats for compilation, including well-known global products such as the Global Reservoirs and Dams database (GRanD), the Food and Agriculture Organization’s AQUASTAT, the Global Dam Tracker (GDAT), GeoDAR, HydroLAKES, and the Global Dam Watch database. The researchers extracted all West African entries, merged dam points and reservoir polygons separately in a geographic information system, and overlaid the results on high-resolution satellite imagery from ESRI and Google Earth to verify coordinates and digitize missing features. Natural lakes and rivers were manually removed so that only man-made reservoirs with dams were retained.

Duplicate records posed a particular challenge, since many global datasets inherit information from earlier inventories or share common sources. The team flagged records with identical or nearly identical coordinates, then compared attributes such as name, country, river basin, reservoir area, storage capacity, and construction year to confirm whether records referred to the same infrastructure. When duplicates were confirmed, a single representative record was retained, with consistent attribute information merged to improve completeness. The final product contains 38 attributes, ranging from geographic coordinates and country identifiers to reservoir surface area, storage volume, dam height, and construction year. All dam entries have known coordinates, roughly 88 percent are linked to a delineated reservoir polygon, and 80 percent include estimated reservoir surface areas.

The numbers reveal a striking geography of water infrastructure. Burkina Faso leads the region with 489 recorded dams, followed by Ghana with 236, while Guinea-Bissau stands out as the only country with no recorded dam or reservoir. The total estimated reservoir surface area amounts to 14,038 square kilometers, with cumulative storage capacity of approximately 283 billion cubic meters. That storage, however, is far from evenly distributed: Ghana alone accounts for more than half of the regional total, thanks to the enormous Akosombo Reservoir, which covers over 6,000 square kilometers. Nigeria follows with the Kainji Reservoir and about 15 percent of regional storage, while Côte d’Ivoire and Mali hold smaller shares. Intriguingly, Burkina Faso, despite having the most dams, represents only 2.4 percent of total storage capacity, a contrast that reflects the abundance of small structures in the semi-arid north versus the few but massive reservoirs of the humid south.

The dataset also captures a clear latitudinal pattern tied to West Africa’s climatic gradients. Dam construction peaked in 1985, with the oldest recorded structure dating back to 1881 in Ghana. In the humid Guinean coastal zone between 6 and 8 degrees north, dams are relatively few but associated with large reservoirs. Between 8 and 14 degrees north, spanning the Sudanian savannah and semi-arid Sahel, dam density increases markedly, with most reservoirs small to medium in size, reflecting decades of effort to buffer rainfall variability for agriculture. North of 14 degrees, near the Sahel’s desert edge, dams thin out to only a few isolated structures. This spatial signature matches the long-standing strategy of using small reservoirs to stabilize food production in regions where precipitation swings dramatically from year to year.

The most revealing part of the study, however, is its ground-truthing exercise. The researchers selected the Upper Bandama watershed in northern Côte d’Ivoire, a 14,500-square-kilometer area in the Sudanian zone upstream of the Kossou hydroelectric dam, and compared the compiled dataset against field observations. During the dry season of December 2024, they consulted national institutions and conducted group interviews in ten villages across four districts, engaging traditional authorities, dam management bodies, user cooperatives, women’s and youth associations, and smallholder farmers. Using structured questionnaires and GPS devices, they documented 192 dams with matching reservoirs, recording construction years, uses, maintenance status, surface areas, and user numbers.

The comparison was sobering. The compiled dataset contained only 62 of the 192 dams documented in the field, capturing just 32 percent of the infrastructure, though every one of those 62 entries was confirmed as real. It accounted for about 62 percent of the total reservoir area and 59 percent of the total storage volume reported from the field campaign. The spatial accuracy of the entries it did contain was excellent, with a very low root mean square error in dam latitudes, and the latitudinal distribution peaked consistently around 9.5 degrees north in both datasets. But the field data showed a systematic underrepresentation of small and recently built infrastructure: existing datasets biased toward older dams by roughly eight years on average, and strongly underestimated small and medium reservoirs in both surface area and volume, with correspondingly high error values. The field survey also revealed much smaller dams, in terms of height, than the compiled records suggested.

The reasons for these gaps are structural. Many small dams in West Africa are built informally at the community level for irrigation or livestock watering, without official documentation, making them invisible to curated inventories. Meanwhile, moderate-resolution satellite imagery of around 30 meters, used by most remote-sensing products, struggles to detect small reservoirs at all. Curated inventory datasets tend to capture large, officially reported dams, while remote-sensing datasets capture smaller waterbodies but cannot always distinguish man-made reservoirs from natural lakes. The compiled dataset inherits biases from both worlds, even as it reduces the missing entries and bias of any individual source by explicitly integrating both large reservoirs of 10 square kilometers or more and smaller ones below that threshold.

The implications reach well beyond cartography. Hydrological models that omit or underrepresent reservoirs can compensate through parameter adjustments, hiding structural deficiencies that surface only when models are applied to climate change projections, where flow estimates diverge substantially from reservoir-explicit simulations. In semi-arid regions where communities depend heavily on reservoirs, inaccurate storage estimates can misguide allocation and planning, as past disputes in the Volta Basin have shown. The authors argue that closing the gap will require open data policies, official publication of national inventories, data-sharing requirements in donor-funded projects, and participatory mapping that captures informal community-built reservoirs. Combining field data with high-resolution satellite imagery, drone surveys, and machine-learning detection could further sharpen the picture. The dataset itself is freely available under a Creative Commons license, offering researchers, planners, and investors across West Africa a regionally optimized foundation for water, food, and energy security decisions, while making clear that the smallest and newest pieces of the region’s water infrastructure still await discovery.

Subject of Research: Harmonization and field validation of dam and reservoir datasets in West Africa

Article Title: A harmonized dataset for dams and reservoirs in West Africa

Article References: Kouassi, V. B. S., Yao, B. A. F., Soro, G. E., Goula, B. T. A., Kelome, N. C., & Klaus, J. (2026). A harmonized dataset for dams and reservoirs in West Africa. Earth System Science Data, 18(10), 7227-7252. https://doi.org/10.5194/essd-18-7227-2026

Image Credits: AI Generated

DOI: 10.5194/essd-18-7227-2026

Keywords: dams, reservoirs, West Africa, hydrology, water security, dataset harmonization, remote sensing, ground-truthing, irrigation, hydropower, Sahel, water management