Combining multiple methods for automated soil delineation: from traditional to digital

Abstract

Context: soil maps are a fundamental tool for agriculture development and for land management planning. Digital soil mapping (DSM) consists of a group of techniques based on geotechnologies and statistics/geostatistics that helps soil specialists to map soil types and properties. Aims: four DSM strategies were applied in south-east Brazil. The goal was to visually delineate soil polygons with support of different strategies. Methods: the delineation started with aerial photographs, followed by a bare soil image composition. Afterwards, it was added layers with landscape characterisation derived from digital terrain covariates and clustering analysis. Finally, digital clay content map from A and B horizons were used to produce a soil texture gradient raster (clay content increasing in depth). Key results: the increasing number of polygons proved that the addition of covariates increased the detail level of the soil map, enhancing visualisation of the landscape variation, resulting on a map that substantially improved both national and state soil inventories. Conclusions: we concluded that combining simple geotechnological tools might be of great utility for increasing detailed soil information proper for farmers and decision making. Implications: therefore, new soil information will be available for end users, supporting land management, food production sustainability, and soil conservation.