Tuesday, November 6, 2007
204-11

C-FARM: a Simple Model to Evaluate the Carbon Balance of Soil Profiles.

Armen R. Kemanian, Blackland Research and Extension Center, Texas AgriLife Research, 720 East Blackland Road, Temple, TX 76502, Claudio Stockle, Washington State University, Washington State University, Biological Systems Eng. Department, Pullman, WA 99164-6120, and David Huggins, USDA-ARS, USDA-ARS Washington State U. ty, 215 Johnson Hall, Pullman, WA 99164.

The principles for modeling soil carbon dynamics have been established during the last decades and have been compiled in several simulation models. Early models consisted of one carbon pool and one residue pool that evolved to multi-compartment pools that represent explicitly the microbial pool and separated residues and soil carbon in several compartments. Keeping the simple structure of one-pool models, we developed C-FARM, a simple soil carbon model designed to compute decomposition rates for soil profile on a layer by layer basis. No calibration is required for using C-FARM. Both the apparent soil carbon decomposition rate (k) and the residues humification rate (h) are a function of the soil temperature and moisture, and a function of the current soil carbon content of the layer under consideration. The maximum k and minimum h occurs for soil with high carbon concentration. In a prototype Visual Basic for Applications version of the C-FARM estimates of aboveground and belowground carbon inputs are obtained from crop growth simulations based on generalized parameters for several major crops to define radiation interception, transpiration, and biomass accumulation. Tillage effects on k are accounted for with an empirical factor that varies from 1 to ~ 2.5, and depends on the tool utilized, number of passes and soil texture. We are currently testing this model against several long-term data sets. Future developments before software release include erosion and denitrification estimates. These modifications will be incorporated preserving the robust structure, minimum inputs, and no calibration requirements of the current version.