Simulating field-scale soil organic carbon dynamics using EPIC

Simulation models integrate our knowledge of soil organic C (SOC) dynamics and are useful tools for evaluating impacts of crop management on soil C sequestration; yet, they require local calibration. Our objectives were to calibrate the Environmental Policy Integrated Climate (EPIC) model, and evaluate its performance for simulating SOC fractions as affected by soil landscape and management. An automated parameter optimization procedure was used to calibrate the model for a site-specific experiment in the Coastal Plain of central Alabama. The ability of EPIC to predict corn (Zea mays L.) and cotton (Gossypium hirsutum L.) yields and SOC dynamics on different soil landscape positions (summit, sideslope, and drainageway) during the initial period of conservation tillage adoption (5 yr) was evaluated using regression and mean squared deviations. Simulated yield explained 88% of measured yield variation, with the greatest disagreement on the sideslope position and the greatest agreement in the drainageway. Simulations explained approximately 1, 34, and 40% of the total variation in microbial biomass C (MBC), particulate organic C (POC), and total organic C (TOC), respectively. The lowest errors in TOC simulations (0-20 cm) were found on the sideslope and summit. We conclude that the automated parameterization was generally successful, although further work is needed to refine the MBC and POC fractions, and to improve EPIC predictions of SOC dynamics with depth. Overall, EPIC was sensitive to spatial differences in C fractions that resulted from differing soil landscape positions. The model needs additional refinement for accurate simulations of field-scale SOC dynamics affected by short-term management decisions.