Reservoirs are changing: What Landsat data reveal about water loss and gain

Reservoirs are changing: What Landsat data reveal about water loss and gain

Reservoirs serve as vital resources for communities around the globe, providing essential services such as drinking water, hydroelectric power, and irrigation. However, these freshwater sources are subject to both seasonal and long-term changes. Water levels can decline during hot summer months or extended droughts, and conversely, they can surge following severe storms. Despite their significance, there remains a substantial lack of information concerning reservoir structure and dynamics. Recent studies published in Scientific Data utilize Landsat data to bridge this knowledge gap.

Researchers from the University of Southampton leveraged Landsat data to map areas of water advancement and retreat from 1984 to 2022, developing the first global dataset that accurately identifies the year of permanent surface water alterations—such as the formation of a reservoir or the drying out of a stream. This comprehensive study can track changes in water bodies as narrow as 30 meters and as small as 900 square meters. Additionally, a separate study by researchers from Texas A&M University generated a global bathymetry dataset named "3D-LAKES," enabling water managers to better estimate reservoir storage capacities.

An output from the 3D-LAKES dataset illustrates the underwater geography of reservoir areas, with green indicating shallow waters and purple indicating deeper regions. This imagery is superimposed on data collected by the Thematic Mapper (TM) on Landsat 5 as of July 31, 1985, combined with a Copernicus Digital Elevation Model (DEM) and bathymetric data from the same dataset.

The animation highlights the Amistad Reservoir, located on the Texas-Mexico border. It employs a natural-color Landsat image from 1985, overlaid with a Copernicus DEM and the 3D-LAKES bathymetric data. The vertical relief in the image has been exaggerated by a factor of four to enhance the visibility of topographic features. The reservoir is managed collaboratively by the U.S. and Mexico through the International Boundary and Water Commission (IBWC) for purposes including flood control, recreation, and hydroelectric power generation.

Despite its crucial role for both nations, the Amistad Reservoir is experiencing gradual shrinkage. The surface water transitions dataset indicates a decline in water levels over recent decades, with notable reductions occurring between 2012 and 2016. The 3D-LAKES dataset further provides insights into the underwater structure of the reservoir, complementing the in situ data on water levels and conditions collected throughout the year.

The relationship between human communities and water is a dynamic interplay; communities modify rivers, construct reservoirs, and create artificial islands, while natural factors such as storms, river shifts, and rising sea levels continually reshape water bodies and coastlines. With satellite data emerging as a vital resource for understanding ecosystem dynamics, researchers have begun to compile a deeper, more global perspective on the presence of water and the changes it undergoes over time. The research team from the University of Southampton honed in on long-lasting alterations in lakes, rivers, coastlines, and other aquatic environments worldwide.

Investigating long-term changes in surface water is instrumental for scientists aiming to discern the underlying factors driving these transformations. Establishing the timeline for when a lake begins to recede enables water managers to explore the causes of such declines, whether due to drought, irrigation, or other factors.

Scientists, policymakers, and water managers are encouraged to utilize the interactive dataset created by Nagel and his team, which allows for the visualization of both local changes and significant global shifts, such as the desiccation of the Aral Sea, newly formed lakes from glacial melt in Tibet, and the creation of the Palm Islands in Dubai.

Evaluating long-term changes in surface water is inherently complex, given the highly dynamic nature of these water bodies. Seasonal variations and climatic influences continuously alter rivers, lakes, and coastlines.

To isolate permanent changes in water bodies while filtering out seasonal variations, the researchers employed two algorithms. The first utilized the Modified Normalized Difference Water Index (mNDWI), which harnesses shortwave-infrared (SWIR) data instead of near-infrared (NIR) band data, to determine whether the water body was advancing or retreating over the given study period. The second algorithm applied the Green_Red Normalized Difference Water Index (grNDWI)—an innovative index proposed by the research team—to pinpoint the precise year when the water body experienced a transition. A change was classified as "permanent" if the water body did not revert to its previous state throughout the study timeline from 1984 to 2022.

"The dataset is showing, for every location on the planet, areas where water advanced or retracted and the year of that change," stated Nagel.