Predicting the Intrinsic Capacity of Terrestrial Carbon Storage Using an Integrative Approach.

Title: Predicting the Intrinsic Capacity of Terrestrial Carbon Storage using an Integrative Approach Authors: P. Chaikaew1,2, S. Grunwald1 1Soil and Water Science Dept, University of Florida, 2181 McCarty Hall A, PO Box 110290, Gainesville, Florida, 32611, USA 2Department of Environmental Science, Chulalongkorn University, 254 Payathai Rd., Pathumwan, Bangkok, 10330, Thailand Organic carbon is a key component of the terrestrial system that affects the physical, chemical, and biological processes. Changes in the terrestrial carbon storage occur due to interactions of natural ecological process and anthropogenic activities. Given the heterogeneity of soil-landscapes and the complexity of ecosystems, research gaps to quantify soil and terrestrial carbon still exist. Carbon storage quantification between the actual and attainable carbon sequestration states helps identify suitable adaptation and management alternatives to optimize natural carbon capital in the context of regional imposed changes, such as land use and climate change. Our objectives were to i) assess the spatially-explicit relationships between soil organic carbon (SOC) and environmental factors, and ii) assess actual and attainable (TerrCactual, TerrCattain) terrestrial carbon capital considering below-ground (soil) and above-ground (biomass) carbon. We collected 234 soil samples in the topsoil (0-20 cm) in 2008 and 2009 across the Suwannee River Basin in Florida, U.S. based on the random design stratified by land cover/land use and soil suborder classes. For above-ground carbon assessment, we derived data from the LANDFIRE project which provided a high-resolution map of Year-2000 baseline estimates (in kg C m-2). A comprehensive set of 172 environmental and human covariates was assembled from multiple data sources to predict and validate SOC stocks and TerrCactual 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) provided the conceptual modeling framework that was implemented using RF and simulated annealing in combination to model TerrCattain. In the simulation, the STEP factors were kept constant. The AWBH factors were varied by ±10, ±20, and ±30 percent. The combined factors which amount to the highest terrestrial carbon stocks were postulated to equal the attainable terrestrial carbon stocks. Results suggest that the TerrCattain stocks derived by the model showed slightly larger amounts than the TerrCactual stocks across the basin. The TerrCactual was 179.4 Tg C and the TerrCattain was 183.7 Tg C. Biotic, soil, parent material, topographic, and water-related factors played important roles in determining SOC storage, while human factors muted from being strong predictors. Mean annual precipitation and monthly mean temperature in summer months were significant to explain both SOC and terrestrial carbon stocks, even though long-term climatic factors showed minor influence on carbon storage. The land use/land cover variables were the strongest factors predicting soil and terrestrial carbon stocks. Land use adaptions have much potential to reach TerrCattain, specifically conversions from cropland to systems with larger net primary productivity. Bare soils, which represent marginal soils, also bear potential to elevate carbon storage through management adaptions that would not compete with other land uses.