Classification and Digital Mapping of Soils in a Semiarid Region of Afghanistan

— A soil map provides information on soil spatial distribution. Such information is essential for effective use of soil resources for crop production, land evaluation, spatial planning, environmental control, and for other similar purposes. In this study, a digital soil map for a semiarid region of Afghanistan was developed. This map contains soil taxonomic information up to the subgroup level, which is the first such attempt in Afghanistan. A total of 114 soil samples were collected in and around the Khost Province through an intensive soil survey. The collected samples were classified into 14 subgroups of soils, following the USDA soil classification system. A soil land inference model (SoLIM) was applied for mapping the recognized 14 soil subgroups digitally, via an expert knowledge-based fuzzy soil inference scheme, with surface topography and other spatial data as inputs. The overall accuracies from the error matrix and Kappa statistics were 0.74 and 0.71, respectively. This map was also compared with the currently used soil map. A general agreement between the two maps was found in the spatial distribution of soil classes, at the great group level. However, the newly developed map contains more detailed information on soils, which might be useful for the advanced use of soil information, for example, to better determine the crop type for cultivation by considering the detailed soil properties. Throughout this study, 14 different recognized classes of soil subgroups were digitally mapped in the study area.