KF-MFWL: A High-Resolution Time Series Construction Algorithm for Lake Water Levels Based on Multi source Altimeter Satellites and Meteorological Data Fusion

Changes in lake water levels are closely related to climate change and can also reflect information about local human activities. Therefore, obtaining high temporal resolution time series of lake water levels is necessary for accurately analyzing hydrological changes. However, the existing methods mainly focus on the long-term changes in lake water levels, with less attention paid to short-term changes in lake water levels. In this paper, we proposed a new method to construct high temporal resolution lake water level time series by fusing multi-source altimetry satellite data based on Kalman filtering and using the MissForest algorithm to combine meteorological data (Kalman Fusion-MissForest water level, KF-MFWL). The accuracy of KF-MFWL was validated using gauge data , as well as compared with HYDROWEB and DAHITI. Finally, a dataset of daily lake water level time series for the Qinghai-Tibet Plateau from 2019 to 2021 has been compiled, and the driving factors influencing water level changes were analyzed. Our result shows that the KF-MFWL time series is comparable to that of HYDROWEB and DAHITI, but with a much higher temporal resolution. The annual rate of water level change for 264 lakes in the Qinghai-Tibet Plateau is 0.021m/y. Among them, the water level of 82 lakes has significantly increased with an average annual change rate of 0.171m/y, while that of 55 lakes exhibits a remarkable decrease with an average annual change rate of -0.145m/y. This study can provide an important data basis for water resource management in the Qinghai-Tibet Plateau region.