Density of soil observations in digital soil mapping: A study in the Mayenne region, France

The density of soil observations is a major determinant digital mapping (DSM) prediction accuracy. In this study, we investigated the effect sampling on performance DSM to predict topsoil particle-size distribution in Mayenne region France. We tested two algorithms, namely ordinary kriging (OK) and quantile random forest (QRF). study area ~5000 km2 with highest field France (1 profile per 0.64 km2). number training sites was progressively reduced (from n = 7500 400, corresponding 1 0.7 13 km2) simulate different observations. For OK QRF, subsampling for splitting data into testing datasets using k-fold cross validation. QRF also conditioned Latin hypercube based point coordinates or covariates. results indicated that, increasing observations, performed as well even better than depending fraction. silt prediction, systematically QRF. However, intervals were much larger did not seem estimate uncertainty correctly. Overall, indicators increased threshold at about 2 which suggests that main limitation accuracy amount collected field, type calibration strategy. Future activities should focus gathering more