The Soil and Water Assessment Tool (SWAT) is a widely used eco-hydrological model for evaluating the land and water resources sustainability in agricultural watersheds. However, when it comes to paddy fields, which are prevalent in Asian countries and contribute to over 90% of global rice production, the standard versions of SWAT are unable to simulate the realistic paddy water dynamics. Previous studies attempting to represent paddy fields as potholes/depression areas of conical shape have been inadequate, as the actual hydrological processes of paddy fields differ from those of potholes, and the complex water management practices in paddy fields cannot be captured. Additionally, the irrigation algorithms and management options in SWAT do not account for water impoundment conditions and water conservation irrigation practices, such as Alternate Wetting and Drying (AWD) and Flush Irrigation, commonly found in irrigated lowland paddy fields. While some recent studies have made modifications to the pothole module within SWAT, it still falls short in realistically representing sub-surface processes during ponding and drying phases of paddy, as well as various management practices like puddling, irrigation, drainage, and the presence of shrinkage cracks during paddy drying periods, all of which have significant impacts on hydrology and crop growth. To enhance hydrological modeling in paddy-dominated watersheds, a more accurate water balance model is required to realistically represent various components, including irrigation, lateral flow, seepage and percolation losses, runoff, and evapotranspiration.
This study proposes integrating irrigation and soil water routing algorithms from field-scale models like ORYZA and CSM-CERES-Rice into SWAT to improve hydrological simulations in river basin analyses. The analysis, conducted using HYDRUS-1D as a pseudo simulator, indicates that these rice production models outperform SWAT in simulating runoff and evapotranspiration, with simulated water balance components aligning well with HYDRUS simulations. The enhanced SWAT model will be evaluated in the Indian river basins where rice paddy is a major crop, to develop best water management strategies for the region. Multi-site and multi-variable calibration using remote sensing data is recommended to improve predictions for various hydrological components in heavily managed river basins. This approach provides better spatial and temporal constraints on water flow processes, enhancing model accuracy. Parameter uncertainty and sensitivity analysis, along with stability tests, are also advised to assess model performance under different calibration schemes.