Suburban Runoff Modeling for Seasonal Water Management in Rice Paddy Fields: An Elaborated Application of Unmanned Aerial Vehicle Photogrammetry

Abstract

This study developed a runoff model considering water flow conditions in paddy fields using high-resolution geospatial data obtained via unmanned aerial vehicle (UAV) photogrammetry. The model spans from the rice transplanting to the harvesting period, accounting for different water management systems. The methodology involved UAV photogrammetry to create a digital surface model (DSM), followed by a detailed catchment analysis. A 1D drainage network model and a 2D ground surface model were combined to simulate the runoff, incorporating Honma's equations for model coupling. Calibration included adjusting the initial loss, infiltration rate, and link roughness coefficients. The results showed high prediction accuracy with Nash-Sutcliffe efficiency (NSE) values of 0.92 and 0.82 for the August 30, 2019 and June 13, 2020 events, respectively, particularly under extreme rainfall (>50 mm). When the initial loss of 4 mm was set as the no water field condition, high reproducibility for initial water depth could be derived (NSE = 0.82). Further, the initial loss in paddy fields and the roughness coefficient for agricultural channels, considering the water level, significantly improved the model accuracy. The study demonstrates that UAV-derived geospatial data can enhance runoff simulations, highlighting the potential for improved water management and flood risk assessment in suburban regions with complex land-use conditions.