Enhancing Chashma Barrage water level estimations with sentinel 3 radar altimetry

Abstract

Satellite Radar altimetry has emerged as a powerful tool for monitoring inland water bodies including rivers, lakes, and coastal regions. However, challenges persist regarding its accuracy, particularly over complex terrains. This study addresses these challenges by enhancing water level estimation accuracy in complex environments. Although the scope of our study may appear limited in terms of temporal and spatial scales, it aims to contribute additional validation results for a comprehensive water-level database utilizing satellite radar altimetry data. In this paper, we applied two pulse filter criteria on Sentinel-3A radar data—Pulse Peakiness (PP) and Misfit (Mf)—to enhance the selection of altimetry waveforms. After applying the filters, the estimated water levels show improved accuracy when validated against in-situ observations. Pulses from non-water surfaces were excluded using riverbank boundaries to ensure higher-quality pulses. A total of six scenarios were considered where the PP and Mf criteria were used to optimize water level estimates. The results indicate that Mf < 1.5 performs well (RMSE = 0.27 m, R = 0.93), and PP > 0.3 yields even better outcomes (RMSE = 0.28 m, R = 0.95). These findings underscore the significance and efficacy of filtering in enhancing the accuracy of water level estimation. The dataset correlation (R) increased substantially, from 0.088 to 0.935–0.953, with the implementation of filters. Additionally, the RMSE significantly improved, reducing from 4.2 m to 0.26–0.27 m with filter parameters. As anticipated, filters removed data points but vastly improved results. Integrating multi-mission satellite data can compensate for data loss and augment overall quantity and quality.