A Spatially Explicit Global Soil Carbon Map

Background: Previous studies have shown biofuels can be an attractive substitute for fossil fuels because biofuels can achieve significant carbon reductions. However, several recent studies have shown that increased demand for some biofuels can lead to net greenhouse gas (GHG) emissions when indirect land use change (ILUC) emissions were included in the emissions accounting. ILUC occurs when land for biofuel feedstock displaces existing crop production in productive land, which then needs to be displaced, including converting new lands, elsewhere, potentially resulting in significant GHG emissions. Yet, there are multiple uncertainties in estimating ILUC emissions, from the types of land conversion, soil and biomass carbon stocks, and emissions factors associated with these conversions. The purpose of this research is to update the carbon stock data used in ILUC analysis by the California Air Resources Board (CARB), which is based on the data from the Woods Hole Research Center (WHRC) and averages for 10 global regions in the form of a look-up table. This study intends to improve the spatial resolution and the land categories of existing values by creating a global spatially-explicit estimate of soil carbon stocks for forest, cropland and pasture lands.   Methods: CARB used the Global Trade Analysis Project (GTAP) model to project the global land use conversion patterns associated with biofuel policies. The GTAP model consists of economic regions and Agro-ecological zones (AEZ). GIS software and the best available maps to date were used to estimate the weighted averages of soil carbon by GTAP AEZ region and land cover types. The Harmonized World Soil Database (HWSD), the most recent spatially explicit global soil map, was used to create spatially-explicit global soil carbon values at the 30 cm and 100 cm depth. This soil carbon map was then overlaid with the forest, cropland, and pasture land cover maps and the GTAP AEZ regions map to estimate the weighted average soil carbon value. Results: The estimates show that forest soil carbon had the highest estimates for both 30 cm and 100 cm, respectively (~360 t C ha^-1, ~1000 t C ha^-1); cropland and pasture had similar high values (cropland: 280 t C ha^-1, 820 t C ha^-1; pasture: 270 t C ha^-1, 770 t C ha^-1). For a majority of the regions, the soil carbon stock at the 100 cm depth was roughly double the soil carbon stock at the 30cm depth. In most cases the cropland and pasture soil carbon values were similar or the cropland values were slightly higher at the 30 cm depth; but there were some significant differences when compared at the 100 cm depth. Conclusions: The soil carbon estimates provided in this dataset are a vast improvement from the previous data used for global land use emission analysis because the best, up to date, and spatially explicit soil and land cover maps were used. However, there is room for improvement such as, incorporating regional soil datasets including: STATSGO for the US, CSIRO for Australia, and NSDB for Canada.