Improving soil maps of rangelands through identification of patterns in soil variability imposed by geomorphic processes

Soil maps of rangeland soils which make up approximately 80% of the landsurface in the US show that all the soils have been mapped and often have very detailed soil analyses associated with them. However, published soil maps consist mainly of poorly characterized soil mapping units which contain one or more soil taxonomic units making it difficult to locate and identify soils in the field. This is a result of the lack of a perceived economic benefit associated with more accurate soils data However, with the impacts climate change being acknowledged, there is increasing need to better understand how soil landscapes will respond to changing environmental conditions. Detailed soil maps are logistically expensive to produce with the current soil mapping programs. More accurate soil data will require a combination of remote sensing, digital soil mapping, machine learning and an understanding of the controls of pedogenic processes. Soil landscapes reflect the actions of geomorphic processes (past and present) that develop distinctive patterns of soil variability and the proxy data that can be used to identify them. These patterns should be used to inform the remote sensing and digital soil mapping approaches to developing new soil maps of nonagricultural areas.