Using homosoils for quantitative extrapolation of soil mapping models

Abstract Since the early 2000s, digital soil maps have been successfully used for various applications, including precision agriculture, environmental assessments and land use management. Globally, however, there are large disparities in availability of data on which mapping (DSM) models can be fitted. Several studies attempted to transfer a DSM model fitted from an area with well‐developed database map areas low sampling density. This usually is challenging task because two hardly ever same soil‐forming factors different regions world. In this study, we aim determine whether finding homosoils (i.e., locations sharing similar factors) help transferring information by means extrapolation. We hypothesize that within world considered as homosoils, one leverage high density fit model, then extrapolated geographically little or no data. collected publicly available clay, silt, sand, organic carbon (OC), pH total nitrogen (N) our study Mali, West Africa its homosoils. regression tree between properties covariates applied Mali. calibration validation strategies were explored. also compared approach existing made at global continental scale. concluded geographic extrapolation was possible, but accuracy dramatically improved when local included dataset. The produced more accurate than products area, three (silt, pH) out six properties. would relevant very carry critical soils risk regional level. Highlights Soil area. Model Homosoil 3 6 maps. New opportunity pattern limited coverage.