Abstract:
Satellite monitoring can obtain the dynamic changes in the water body area of reservoirs, thereby addressing the shortage of traditional ground observation data.This study employs a water body pixel recovery algorithm to correct contaminated images from optical satellites such as Landsat 8, Landsat 9, and Sentinel-2, and to conduct automated water body area extraction.Additionally, Sentinel-1 synthetic aperture radar satellite data is introduced, with mountain shadow correction applied to further enhance the temporal and spatial resolution of water body area monitoring.Taking the Biliu River Reservoir as an example, the results show that:①The water body classification enhancement algorithm can effectively correct cloud contamination and increase the utilization rate of optical images. The availability of optical images for the Biliu River Reservoir increased from 36% to 60%, with an overall classification accuracy as high as 98.58%; ②After mountain shadow correction, the overall classification accuracy of radar images is 93.20%, and the combination of optical and radar remote sensing images can basically achieve 2~7 d high-frequency observation of water body area; ③Over the nearly 40 years from 1984 to 2023, the water body area of the Biliu River Reservoir has shown an increasing trend, with a rate of change of 0.23 km
2/a.The results of the study can improve the accuracy and spatial and temporal resolution of the dynamic monitoring of the water body area in the area lacking observational data, and provide a scientific basis for the management of water resources in the basin and the formulation of scheduling plans.