Modeling evapotranspiration from rice paddies with variable water depths

Accurate estimation of actual crop evapotranspiration (ET c,act ) and its components is essential for irrigation management. In this study, the dual source Shuttle-Wallace (SW) model was applied to estimate ET c,act under variable ponding water depths in flooded rice paddies. To better represent the underlying biophysical processes, the SW framework was enhanced through two revised parameterizations: (1) a canopy resistance model accounting for water depth and phenology, and (2) a ground heat flux sub-model derived from surface energy balance under different ponding depths. Field measurements of stomatal conductance, ET c,act , and evaporation (E) collected during 2021–2023 growing seasons were used to calibrate and independently evaluate the proposed model. Results showed substantial improvements in model performance across different management scenarios, with the RMSE for ET c,act decreasing from 1.00 to 0.52 mm d −1 , and that for E decreasing from 0.41 to 0.26 mm d −1 . Sensitivity analysis indicated high model sensitivity to canopy resistance, highlighting the importance of its accurate parameterization. Model simulation revealed that E decreased with increasing water depth, whereas transpiration (Tr) initially increased and then declined, peaking at a ponding depth of 6.5 cm. ET c,act followed a similar pattern to Tr, with the depth corresponding to maximum ET c,act shifting as LAI increased. Whole-season ET c,act was highest under the 5–7 cm treatments, with relatively small differences among treatments (15 %).