Tradeoffs in Model Performance and Effort for Long‐Term Phosphorus Leaching Based on In Situ Field Data

Core Ideas Numerical models were calibrated with in situ infiltration and leaching data. We evaluated how increasing model complexity improves predictions. Critical parameters included field‐measured hydraulic conductivity and dispersivity. A one‐dimensional dual‐porosity model was adequate for long‐term simulations. Phosphorus and N are critical nutrients for agriculture but are also responsible for surface water enrichment that leads to toxic algal growth. Although P loading to surface waters has traditionally been thought to occur primarily in surface runoff, contributions from subsurface transport can also be significant. The primary objectives of this research were to evaluate several methods of representing macropore flow and transport in a finite element model using plot‐scale infiltration and leaching data and to compare several models of various levels of complexity to simulate long‐term P leaching. To determine flow and transport parameters, single‐ and dual‐porosity models in HYDRUS‐2D were calibrated with infiltration, Cl−, and P data from a 22‐h plot‐scale leaching experiment on a silt loam mantle with gravel subsoil. Both homogeneous and heterogeneous gravel profiles were simulated. The dual‐porosity model with heterogeneous hydraulic conductivity best matched experimental data, with physical nonequilibrium (dual porosity) being more important than two‐dimensional (2D) heterogeneity. Long‐term (9 yr) P leaching to the water table (3 m below the soil surface) at the field site was simulated with both one‐dimensional (1D) and 2D models using the calibrated parameters. There was little difference between analogous 1D and 2D models, suggesting that HYDRUS‐1D may be sufficient to model long‐term P leaching. Overall, the most important elements for accurately simulating P leaching in this silt loam and gravel soil profile were found to be (i) field‐measured hydraulic conductivity of the limiting soil layer, (ii) calibrated dispersivity, and (iii) dual‐porosity, in some circumstances.