Scale-Specific Prediction of Topsoil Organic Carbon Contents Using Terrain Attributes and SCMaP Soil Reflectance Composites

There is a growing need for an area-wide knowledge of SOC contents in agricultural soils at the field scale food security and monitoring long-term changes related to soil health climate change. In Germany, maps are mostly available with spatial resolution 250 m 1 km2. The nationwide availability both digital elevation models various resolutions multi-temporal satellite imagery enables derivation multi-scale terrain attributes (here: Landsat-based) reflectance composites (SRC) as explanatory variables. example Bavarian test about 8000 km2, relations between 220 content samples well different aggregation levels variables were analyzed their scale-specific predictive power. generated by applying region-growing segmentation procedure, prediction was realized Random Forest algorithm. doing so, established approaches (geographic) object-based image analysis (GEOBIA) machine learning combined. modeling results revealed differences. Compared attributes, use SRC parameters leads significant model improvement field-related levels. joint resulted further improvements. best variant characterized accuracy R2 = 0.84 RMSE 1.99.