A new method for modelling key hydrological processes in paddy-dominated watershed based on water balance and remote sensing

This study addresses the need for accurate runoff data for sustainable water resource management in paddy fields, focusing on China, where agriculture consumes more than 62% of freshwater, and paddy rice is the most water-intensive crop. Given the risk of nitrogen and phosphorus loss through runoff, accurate models are crucial for enabling improved irrigation management and assessing agricultural non-point source pollution.

Modern hydrological models range from semi-empirical models, which are deficient in describing the growth phase of paddy, to process-based models that span either single large-scale paddy fields or the entire watershed. However, variations in historical models—and specifically, models such as SWAT-Paddy—indicate significant uncertainties due to the uniform application of irrigation date, amount and drainage outlet height.

This study introduces a novel method that synthesizes the spatial distribution patterns of drainage outlet height and irrigation information (date and amount), while integrating different irrigation and drainage management protocols across various phenological periods. This method uses Google Earth Engine to build a continuous spatiotemporal resolution evapotranspiration model based on multiple-source remote sensing satellites. It also leverages the water balance equation to automatically identify spatiotemporal patterns of runoff at the field scale.

We anticipate that this inclusive, accurate, and automated method will not only facilitate accurate quantification and assessment of paddy runoff but also provide critical data for studying agricultural non-point source pollution. These findings contribute to the existing body of knowledge on paddy water cycle dynamics and highlight the potential of remote sensing technology in addressing data scarcity challenges.