REGIONAL SOIL ORGANIC CARBON STOCK ESTIMATES ARE INFLUENCED BY ASSESSMENT METHOD

National and regional soil organic carbon (SOC) inventories are indispensable to assess direction and magnitude of C fluxes between soils and the atmosphere. For these inventories SOC density [C mass/unit area] is typically determined from spatial databases of soil properties and spatially distributed to soil or soil/land use categories on a map. Conclusions based on these SOC assessments are often weakened by uncertainty given the large (10-50%) errors associated with some of the studies. The purpose of the present study was to evaluate alternative assessment methods for SOC stock, and was carried out in Flanders, one of the three federal regions of Belgium. Previous studies have examined the effect of multiple SOC estimation approaches by comparing results from various soil databases and soil maps for a given region, and have demonstrated that outcomes can significantly differ. We employed one soil database and one digital soil map but evaluated three methods to determine SOC density for soils in the spatial database and three spatial distributing approaches to assign the SOC density to map units.

Results show that calculated SOC density is significantly influenced by density estimation method. The maximum difference in total SOC stock for Flanders resulting from the use of various SOC density estimation methods is 6%. Findings indicate that horizon-specific SOC density estimation procedures are better for SOC stock determination than soil type based procedures. When a soil map alone is used as the geographical reference, little difference (1 %) in total SOC stock results from the application of various spatial distributing methods. When the soil map is combined with a land use map, which is frequently done in this type of study, results for total SOC stock are consistently lower (up to 2%) than when a soil map alone is used. The largest difference in SOC stock estimate between any combination of estimation methods is 7 %. This difference is comparable to the maximum variation in stock estimate resulting from the analysis of various soil databases and spatial aggregation methods for other regions. Results underscore the importance of complete, high-quality spatial soil datasets and standardized methods to determine SOC stocks in order for stock estimates to be useful for efforts examining the role of soils in the C cycle.