Optimizing the estimation of water storage variation in lakes with limited satellite altimetry coverage

Abstract

The empirical formula (EF) method, which do not rely on topographic data, stands as the prevailing technique for estimating lake water storage variation (LWSV). However, for smaller lakes, the sporadic monitoring frequency of satellite altimetry fails to adequately support this method, presenting a challenge in accurately gauging LWSV. Using Lake Chahannur, a lake in China with an area smaller than 50 km², as a case study, seven schemes based on the EF method and the Area-Volume-Height (A-V-H) curve method were designed to estimate the LWSV of this undersized lake. The efficacy and precision of each scheme were evaluated against field-measured elevations. Findings reveal that due to the limited satellite altimetry monitoring, both the EF method and the H-driven A-V-H curve schemes struggle to provide consistent and comprehensive estimations. In the A-driven A-V-H curve schemes, terrain data from SRTM DEM suffers from mask processing and substantial errors, with the former posing challenges for shrinking lakes and the latter significantly compromising estimation accuracy. While field-measured elevations boast high precision, the interpolation process leads to terrain maps lacking in detail, with site density becoming a crucial factor influencing the accuracy of LWSV estimation. The combination of terrain reconstruction and A-driven pattern emerges as the most promising, boasting high accuracy, rich detail, and significantly reduced reliance on satellite altimetry monitoring, making it particularly suitable for small lakes. Chahannur’s bottom elevation ranges between 1271.71 and 1273.44 m, and the lake shows a downward trend in water volume from 1991 to 2020, with fluctuations totaling approximately 35 million m³. This study serves as a vital addition to the field of LWSV estimation, potentially broadening the scope of estimation from large-scale lakes to a wider array of global surface water bodies.