Knowledge is power: Where geopedologic insights are necessary for predictive digital soil mapping

Abstract

Much of current predictive digital soil mapping (PDSM) practice relies on terrain, climate, and remote sensing-derived covariates. These are easy to obtain and can serve as proxies to soil forming factors and from these to soil properties. However, mapping of soil bodies, not properties in isolation, is what gives insight into the soil landscape. A naïve attempt at correlating environmental covariates from current terrain, vegetation density, and surrogates for climate will not succeed in the presence of unmapped variations in parent material, soil bodies, and landforms inherited from past environments. Geopedology integrates an understanding of the geomorphic conditions under which soils evolve with field observations. Examples where simplistic DSM would fail but geopedology would succeed in mapping and, even better, explaining the soil distribution are shown: exhumed paleosols, low-relief depositional environments, and recent post-glacial landscapes.