What's the deal with dam data?

This blog is the first in a series on Global Data on Dams and Reservoirs.

Societies have relied on dams for centuries. 

Storing water is a critical part of water security and has been a societal response to hydrological variability for millennia. Humans have been supplementing natural water storage with dams and reservoirs since as early as 3000 BCE.

Today, dams and reservoirs represent a significant share of the available built storage, enabling greater control over surface water resources and driving social and economic development. They provide significant benefits to society, allowing water to be stored for potable supply, irrigation, power generation, and flood attenuation. They also embrace a broad range of complex social, environmental, and political choices, fundamentally altering the rivers along which they are developed.

After dams are constructed, they are often very long-lived infrastructure. A well-designed dam, with good maintenance and periodic rehabilitation, can function for a hundred years or longer. Some, built centuries ago, are still in operation. Examples include the Proserpina Dam in Spain (1st to 2nd century CE) and the Lake Homs Dam in Syria (3rd century CE), both constructed by the Romans. As countries continue to build, the number of dams in existence continues to grow, year after year.

Just how many dams are there in the world? 

Truth is, we don’t know. The World Bank commissioned a review in 2023 that identified at least 15 global and 14 regional/national datasets on dams and reservoirs, each with differing numbers of records. Among the most well-known and trusted of these datasets is the World Register of Dams, curated by the International Commission on Large Dams (ICOLD), which captures “large dams” meeting ICOLD’s definition and includes more than 62,000 records. The National Inventory of Dams, published by the U.S. Army Corps of Engineers in cooperation with the Association of State Dam Safety Officials, includes more than 90,000 dams in the U.S. and its territories. Other well-known datasets include those curated by Global Dam Watch (GDW), including its new GDW Consensus Global Database with 41,145 barriers and dam locations and 35,295 associated reservoirs, with a cumulative storage capacity of 7,405 cubic kilometers. 

Still, relatively little is known about the distribution of smaller dams, which are thought to number in the millions (Lehner et al. 2011) and typically experience higher rates of failure than large dams (World Bank 2021). Most national inventories of dams and reservoirs give superior coverage to large dams or dams for specific purposes, such as hydropower generation. 

Knowledge gaps pose several challenges to ensuring the safe and sustainable development and operation of dams. Among them is ensuring that all dams are subject to the appropriate level of oversight, understanding their influence on the water system, and fully assessing the cumulative impacts of their construction on local communities and the environment. Yet, surveying dams, particularly small ones, is a time-consuming and costly undertaking, especially in large countries and in remote or difficult-to-access locations, and where the dams are distributed across different sectors, owners, and jurisdictions.

Technology as a game changer

Technological developments, namely in earth observation and machine learning, are beginning to make this task easier by enabling the initial identification and mapping of dams and reservoirs across large scales. This includes first-order estimations of important characteristics like dam height, length, storage capacity, and reservoir area. These technologies are even being used to detect small movements and deformations, which may indicate structural issues on a dam. And the proof is in the data. 

In Australia, remote detection yielded 202,119 farm dams that were previously undocumented, adding to the 1,694,675 farm dams captured in regional and federal data (Malerba et al. 2021). At the global level, Donchyts et al. (2022) processed nearly 35 years of satellite data to identify 71,208 small to medium-sized reservoirs. These technologies also allow us to look at how water bodies, including artificial reservoirs, change across seasons and over the years, as shown by Druce et al. (2021) in their mapping of surface water dynamics in mainland China.

The World Bank has also been supporting the deployment of these tools in its client countries, and the results are striking. In 2011, a World Bank-supported remote sensing analysis identified 1,022 reservoirs (0.5 hectare or larger) in Zambia’s Southern Province, which previously recorded a total of 293 dams. In the Nile Basin, a 2023 analysis detected more than 800 reservoirs (1 hectare or larger). Both initiatives made significant contributions to national inventories, adding dams that were not previously recorded in government repositories. Similar analyses are underway in other parts of the world, including Cambodia, Central Asia, and the Euphrates and Tigris basins. 

Various researchers across the globe are now applying these techniques at scale. The result is a rapidly growing wealth of data, and this exponential increase highlights the need for consensus on a global protocol to improve the acquisition, accuracy, assurance and archiving of such data. With greater collaboration, these outputs can unlock the enormous analytic potential of more complete, more accurate, open data on dams and reservoirs, and water resources more broadly.  

Towards this, the World Bank is working with stakeholders to promote collaboration in data gathering and good practices in the use of these emergent technologies. As we harness the power of innovation and partnerships, we mark a new chapter in our enduring relationship with these ancient structures, ensuring that they are well accounted for into the future. This initiative is crucial not only for safeguarding downstream communities, but also supporting the planning and sustainable management of water resources.