Simulating crop yield, surface runoff, tile drainage and phosphorus loss in a clay loam soil of the Lake Erie region using EPIC

Abstract Useful in evaluating best management practices and nutrient management planning, the prediction of phosphorus (P) transfer from agricultural lands to water bodies via surface runoff and tile drainage remains as a challenge, as few models can provide reasonably accurate P loss simulations. The EPIC (Environmental Policy Integrated Climate) model was firstly applied to simulate crop yields, surface runoff, tile drainage, and dissolved reactive P (DRP) losses under a corn-soybean rotation grown on a “cracking” Brookston clay loam soil (Vertisol) in the Lake Erie basin, Ontario, Canada. Different potential evapotranspiration and curve number equations were compared to determine the most suitable equations (Penman-Monteith equation and variable Daily Curve Number with soil moisture index) for this region. A crack flow coefficient was used to deal with inflow partitioned to cracks. A soil layer below tile drain with low saturated hydraulic conductivity was employed in simulating the experimental site, where most water was leaving the field through tile drain. Lateral subsurface flow was used to substitute for drainage. Annual simulations of crop grain yield, cumulative surface runoff, and cumulative drainage closely matched observed data. Over shorter periods (months), surface runoff (NSE = 0.78), tile drainage (NSE = 0.57), and relevant DRP loss (NSE > 0.5) simulations were satisfactory, except for two periods of DRP loss in surface runoff, where most DRP moved downward through lateral flow and deep percolation due to limitations in the crack flow coefficient. For this vertic soil, EPIC generally simulated crop yields and flow volumes well, while DRP losses were only adequately simulated.