Data-Driven Assessment of Climate and Topographic Integration With Satellite Imagery for Improved Surface Water Detection Under Agricultural Vegetation

Accurate detection of surface water in rice paddy fields is essential for optimizing irrigation practices and evaluating greenhouse gas emissions. While optical satellite data offer valuable insights into field conditions, their effectiveness is reduced under dense vegetation where surface water is often obscured. To address this limitation, we present a multi-source synergistic framework that integrates Earth observation data (e.g., Landsat, Sentinel) with climate variables and topographic features. This integration captures key environmental drivers, including antecedent climate conditions, soil moisture dynamics, and terrain-driven runoff potential. The framework was evaluated through data-driven supervised classification using insitu water level measurements as ground truth. Results show that incorporating climate and topography data significantly enhances detection performance, improving surface water presence detection accuracy by 49% and increasing prediction reliability by 22% compared to using satellite data alone. The framework also maintains high performance under vegetated conditions, demonstrating its robustness across varying crop stages. This study highlights the potential of combining satellite and environmental data to more accurately monitor surface water dynamics in rice paddies at a high spatial and temporal scale. The improved detection supports more informed irrigation scheduling and enhances the capacity to assess the environmental impacts of water management strategies, including those related to greenhouse gas emissions.