A revised methodology for estimation of forest soil carbon from spatial soils and forest inventory data sets

Soil organic carbon (SOC) represents the largest constituent of the global C pool and is used by researchers in C cycling, global climate change, and soil quality studies. Spatial, pedon, and soil interpretation record databases are widely used to estimate regional SOC. This study compared published SOC estimates with estimates of mass SOC to 2 m in Maine and Minnesota using STATSGO data tables edited and filled by automated software scripts. Valid STATSGO soil property data were used to produce replacement values for invalid or missing data after grouping by soil order, MLRA, layer number, and texture. Area-weighted mass SOC was calculated using log-transformed data. Between 30% and 54% of the large rock fragment data were invalid, and between 18% and 48% of the missing OM values were replaced. The log-transformed area-weighted mass SOC to 2 m was 7.88 kg/m(2) (SD = 9.24 kg C/m(2) CV = 117.2%) for Maine and 17.38 kg/m(2) (SD = 15.30 kg C/m(2) CV = 88.1%) for Minnesota. These values were lower than earlier estimates because of the log-transformation and because our error checking increased the volume of rock fragments. The FIA database was merged with STATSGO to produce mass SOC by forest-type group. The elm-ash-cottonwood (7.22 kg C/m(2)) and the spruce-fir (17.73 kg C/m(2)) forest-type groups had the highest SOC (to 1 m depth) in Maine and Minnesota, respectively. The methods and scripts used in this study can be easily adjusted, and as they are improved, they in turn can improve the quality of data in STATSGO tables.