Modeling Actual and Attainable Soil and Terrestrial Carbon.

Given the complexity and heterogeneity of soil-landscapes, research gaps still exists to quantify soil and terrestrial carbon and identify suitable adaptation and management options to optimize natural carbon capital in context of imposed changes, such as land use and global climate change. Our objectives were to: i) assess the spatially-explicit relationships between soil organic carbon (SOC) and environmental factors, and ii) assess actual (TerrC_actual) and attainable (TerrC_attain) terrestrial carbon consisting of below-ground (soil) and above-ground (biomass) carbon. A total of 234 soil samples in the topsoil (0-20 cm) were collected between 2008 and 2009 across the Suwannee River Basin in Florida, U.S. Biomass data was retrieved from the National Biomass Carbon Data. A comprehensive set of 172 environmental and human covariates was compiled from multiple data sources to predict and validate SOC stocks and TerrC_actual using Random Forest (RF) The STEP-AWBH conceptual model (with S: Soil, T: Topography, E: Ecology, P: Parent material, A: Atmosphere, W: Water, B: Biota, and H: Human factors), RF, and simulated annealing were used in conjunction to model TerrC_attain. In the simulation model, the STEP factors were kept constant, whereas the AWBH factors were varied by 10, 20, and 30 percent to assess TerrC_attain. Results suggest that biotic, soil, parent material, topographic, and water-related factors play crucial roles in determining SOC storage, while human factors fade from being strong predictors. Even though climatic factors were minor predictors, mean annual precipitation and monthly mean temperature in summer months were significant to explain SOC and terrestrial stocks. The TerrC_attain stocks derived by the model presented slightly larger amounts than the TerrC_actual stocks within the whole basin. The basin-wide TerrC_actual was 179.4 Tg C and the TerrC_attain was 183.7 Tg C.