Analysis and Validation of Water Surface Data in Two Reservoirs of Bogotá’s Water Supply System Using SWOT and Sentinel-2 Satellite Products

Bogotá, the capital of Colombia, relies on one of the largest and most complex urban water supply systems in the country. Drinking water is primarily sourced from high-altitude reservoirs located within the Chingaza National Park, situated in the Andean páramo ecosystems surrounding the metropolitan area. Beginning in late 2023 and continuing into 2024, severe drought conditions linked to the El Niño phenomenon led to historically low storage levels in the Chingaza system. As a consequence, in April 2024, Bogotá implemented water rationing for the first time in recent decades.

In this context, this study evaluates the potential of satellite-based observations to support reservoir monitoring and water resources management in Bogotá.  We analyze water surface elevation (WSE) and water surface area (WSA) for the Chuza and San Rafael reservoirs—both part of the Chingaza system—by integrating in situ observations with products from the Surface Water and Ocean Topography (SWOT) mission and Sentinel-2 imagery.

First, WSE estimates from SWOT Level-2 products (LakeSP and Raster) were validated against observed WSE between September 2023 and February 2026. Over the same period, WSA derived from SWOT and Sentinel-2 was validated against in situ WSA estimated from reservoir hypsometric curves.

Results show an excellent agreement between SWOT LakeSP WSE and observations for the Chuza reservoir, with a Pearson correlation coefficient of 0.9934. In contrast, performance for the San Rafael was substantially lower, with a correlation of 0.5971 and a systematic negative bias of −8.83 m, indicating reduced accuracy of SWOT-derived WSE in this smaller reservoir. Similar patterns were observed for the SWOT Raster product with moderate correlation for Chuza (0.7138), and large discrepancies for San Rafael. In contrast, WSA time series derived from the Normalized Difference Water Index (NDWI) from Sentinel-2 exhibited acceptable performance for San Rafael, with a correlation of 0.8535 and a bias of −0.37 km², and 0.6207 correlation and -0.61 km² bias for the Chuza reservoir.

Overall, the results demonstrate the strong potential of SWOT and Sentinel-2 products for monitoring strategic water supply reservoirs in the Colombian Andes, while also highlighting pronounced performance differences linked to reservoir size, shape and surrounding topography. The findings emphasize the usefulness of these satellite products for hydrological applications in the Colombian Andes but also underscore challenges such as persistent cloud cover and the need for more robust algorithms that integrate the high precision of SWOT with complementary datasets to improve area estimation.