Procedures for Initializing Soil Organic Carbon Pools in the DSSAT-CENTURY Model for Agricultural Systems

Process-based soil organic C (SOC) models are widely used for simulating, monitoring, and verifying soil C change. In such models, determining the initial distribution of SOC across multiple pools is oft en not well defined, yet pool initialization can strongly influence the subsequent SOC dynamics. We developed a model-based procedure to initialize SOC fractions that uses site-specific soil, climatic, and land use history information, along with measured initial total SOC, to estimate SOC distribution across pools. The procedure consists of creating sets of SOC fractions for scenarios including different field histories, soil texture, and management practices for medium-and long-term simulations to provide a broad spectrum of conditions, using data from an experimental site in Georgia. Tables created using this procedure can help model users select the SOC fractions needed to properly initialize the model. The model is executed for the duration of the prior land use history time period, and the simulated final total SOC is compared with the soil C measurement at the beginning of the subsequent experimental time period. If these values are equal, the final SOC fractions from the land use history simulations are used with the measured soil C to start the simulation of the cropping system being studied. If these values are not equal, then the procedure is repeated iteratively until the measured value of soil C is adequately predicted. We demonstrated the improved prediction accuracy by using the proposed procedure in simulating soil C for a long-term field experiment in Michigan.