Multi-Source and Multi-Temporal Radar Data for Hydrological Monitoring of Precipitation and Surface Water in Northwestern South America

Water-related ecosystems provide services that are essential for the subsistence of life on Earth. Global initiatives such as Agenda 2030 and the SDGs defined goals to ensure the sustainable use of these ecosystems. However, a lack of information on the hydrology of these ecosystems poses a challenge to reach the goals. Satellite remote sensing present an opportunity to address this limitation and are the only source of information in ungauged regions. In this study, we developed strategies to monitoring hydrological processes using multi-source and multi-temporal radar satellite data. We evaluated precipitation, surface water flow and water level change as descriptive elements of surface water. This evaluation was done in Northwestern South America, representing an excellent opportunity due to its climatic and ecosystem diversity. The dataset included precipitation products from the TRMM and the GPM Missions over the entire region. We evaluated the accuracy of the products to estimated rainfall amounts. The precipitation data were combined with Interferometric Synthetic Aperture Radar (InSAR) observations from PALSAR-1 and Sentinel-1 to analyse surface water flow and water level change in different ecosystems. Evaluated ecosystems included lakes of the Cajas Massif National Park in Ecuador and low land floodplains with vegetated wetlands in the Atrato River, Colombia, an essential fluvial system in the region. First, we found that satellite-based precipitation estimation in the region is a challenge due to the complex topography, rain, and precipitation mechanisms. Although errors in the amount of rainfall reached 50%, the products represented well the spatiotemporal distribution of precipitation. Second, we found that changes in the water level of the floodplains were explained in 56% by discharge in the main river and precipitation in its river basin, thus demonstrating significant hydrological connectivity. Thirdly, water level changes were related to precipitation and lake area, with correlations reaching R2 values of 0.65. This study implemented methods and datasets that had not been explored before in this region and provided elements to improve precipitation and surface water monitoring.