River discharge monitoring using remote-sensing and hydrological modeling: the case of Golina River, Danakil Basin, Ethiopia

Abstract

Reliable discharge data remain scarce in arid and semi-arid basins, limiting water balance assessments and management. Although remote sensing indices and large-scale hydrological models have been applied, their suitability for small, data-scarce catchments such as the Golina River (Danakil Basin, Ethiopia) has not been systematically evaluated. This study directly compares the multiple pixel ratio (MPR) approach, which links Landsat 8 near-infrared reflectance to observed discharge with Global Flood Awareness System (GloFAS) reanalysis products. Using 251 Landsat scenes (2013–2024), we calibrated discharge estimates at different channel lengths and widths near the Addis-Alem station with in-situ data (2022–2024). Results show that MPR, especially when applied close to the gauging station, outperforms GloFAS in capturing local discharge variability. This demonstrates a transferable satellite-based methodology for ungauged rivers, offering a robust alternative for monitoring discharge in data-scarce basins.