“…For instance, Landsat, WorldView, SPOT, KOMPSAT, and IKONOS images have been effectively used for this purpose, for different soil layers [16]. In homogeneous or bare soils, simple linear regression models and remote sensing band values are usually used to estimate SOC and STN stocks across space [12,17]. To date, various DSM techniques have been used for the mapping, such as multiple stepwise regression (MLSR) model [13], regression Kriging [18], ordinary Kriging [19], random forest model [20], boosted regression trees (BRT) [21], geographic weighted regression (GWR) [22], and principal component regression [23].…”