Predicting dissolved reactive phosphorus in tile-drained catchments using a modified SWAT model

Abstract

Phosphorus (P) is mainly leached by subsurface transport pathways in tile-drained landscapes. In this study, we modified the SWAT model (SWAT-P) by incorporating dissolved reactive phosphorus (DRP) losses from drainage water and the deep aquifer. SWAT-P was tested in a tile-drained lowland catchment using a multi-site calibration and validation approach. SWAT reached a good statistical performance regarding discharge for all sub-catchments and a daily time step. As discharge was dominated by subsurface flow, we optimized DRP concentrations for the drainage water (SWAT-P), the shallow aquifer (SWAT, SWAT-P), and the deep aquifer (SWAT-P) and left other P-related parameters at their default settings, since they did not influence the model output. DRP losses were simulated at a monthly time step using SWAT and SWAT-P. The predictive power was weaker compared to discharge for both SWAT and SWAT-P. Nevertheless, SWAT-P performed considerably better than SWAT. Additionally, calibrated DRP concentrations were unrealistically high for SWAT, whereas calibrated DRP concentrations reflected the prevailing conditions in the region using SWAT-P. The results indicated improved prediction accuracy for DRP losses into streams by using SWAT-P, as well as a roughly realistic estimation of DRP losses from tile drainage water. Further research is necessary to account for the temporal DRP concentration dynamics in drainage water. SWAT-P is ready to use after defining DRP concentrations in tile drainage water and the deep aquifer in the SWAT-P input files. In addition, the model output was extended in SWAT-P to visualize DRP losses from drainage water and from the deep aquifer.