Assessment of the Potential of Medium‐Resolution Satellite Images for Estimating Rice Yields in Semiarid Irrigation Systems in Niamey

Rice underpins global food security; however, in semiarid regions such as Niger, production is limited by climate variability, arable land and urban expansion, making reliable yield estimates essential for guiding agricultural policies and enhancing food security. This study aims to evaluate the effectiveness of remote sensing in estimating rice yields in the Niamey region during four dry seasons (2022–2025). Landsat imagery was used to derive the normalised difference vegetation index (NDVI), which was correlated with yield statistics from ‘Office National des Aménagements Hydro‐Agricole’. Rice fields were mapped using a support vector machine classifier, and exponential regression models were developed. The results show a good NDVI‐yield relationship, with R 2 exceeding 0.91 across all seasons. Estimated yields ranged from 5.61 to 6.15 t/ha, closely matching reported values of 5.32–5.85 t/ha and achieving accuracies of 94.05%–94.87%. However, cultivated areas were underestimated by 12–40 ha due to mixed pixels, fragmented boundaries and heterogeneous plot configurations. Area estimation showed moderate reliability ( R 2  = 0.58); production performed strongly ( R 2  = 0.93), whereas yield ( R 2  = −0.90) exhibited poor explicative power due to flooding impacts and management variability. Medium‐resolution satellite imagery enables rapid, cost‐effective rice monitoring, irrigation management support, food‐security planning and land‐use policy development across the Sahel.