Enhancing reservoir water level time series in the Mekong river basin by improving area-elevation curves and integrating multi-source satellite data

Abstract

Study region Mekong River Basin. Study focus Accurately constructing long-term, high-frequency reservoir water levels is challenging due to limited altimetry data, frequent cloud contamination of optical imagery, and large biases in area-elevation (AE) curves from the current DEM. We presented a framework for enhancing reservoir water levels by improving AE fitting curves through integrating multiple altimetry and optical data. First, an improved satellite-derived AE curve was proposed by integrating the altimetry level and the corrected water surface area. Besides, two improved DEM-derived AE curves were proposed by adding satellite-derived points or reservoir parameters. Then, performances of the proposed AE curves were evaluated and compared. Finally, the optimal curves were implemented at 32 reservoirs to construct long-term and high-frequency water levels based on the availability of altimetry data. New hydrological insights The three proposed curves improved the accuracy of inverted water levels with an average declined RMSE by over 2 m. The improved satellite-derived AE curve declined RMSE by 2.03 m. The DEM-derived AE curve with satellite-derived matched points declined RMSE by 2.85 m. The joint use of reservoir parameters also decreased RMSE by 2.14 m. The average monthly frequency of water levels was enhanced by 79.14% among the 32 reservoirs. The study proposed a potential and reliable framework for constructing long-term and high-frequency water levels in ungauged reservoirs.