Digital Updating of Traditional Soil Maps

Abstract The new approach for updating traditional paper soil maps is presented. It is based on the imitation modeling of traditional expert-based methods of soil mapping. The approbation results of our approach to soil mapping are exemplified by a test plot in the center of European Russia. The open-source script program IMSOIL has been developed for updating soil maps, using the R programming language. Contrary to other methods, our approach enables to use the expert-based knowledge on soil geography in combination with the results of the statistical analysis of data. The updated soil map better corresponds to actual information about soil-forming factors as compared to traditional paper soil maps. The proposed approach to the imitation of soil mapping in geoinformation system includes the use of qualitative information about soil geography and quantitative rules of soil mapping. The models in the form of decision trees can be analyzed by an expert, and the improvements to the covariates and obtained models may be made. The form of the decision trees of the soil mapping rules provides saving traditional qualitative description of the soil geography in the formalized form and to integrate it with unambiguous quantitative rules of soil mapping. They are based on satellite data and thematic maps. Such a method enables automatic extraction of information concerning the relationships of soils and soil-forming factors. The updating of soil maps is accompanied by the probability map.