Using Magnetic Susceptibility Mapping for Assessing Soil Degradation Due to Water Erosion.

This study focused on developing a method for estimating topsoil organic carbon content from measured massspecific magnetic susceptibility in Chernozems heavily affected by water erosion. The study was performed on a 100 ha area, whereby 202 soil samples were taken. A set of soil samples was divided into 3 subsets: A (32 samples), B (67 samples), and C (103 samples). The mass-specific magnetic susceptibility using low (χ_lf) and high (χ_hf) frequency, and organic carbon content were measured at all soil samples. The contents of iron and manganese, extracted with a dithionite-citrate solution (Fe_d, Mn_d) and ammonium oxalate (Fe_o, Mn_o), were quantified in A and B samples. Models for predicting organic carbon content from magnetic susceptibilities were designed as follows: (1) subset A was used as the training set for calibration, and subsets B and C were used as the test sets for model validation, either separately (subset B only), or together (merged subsets B and C); (2) merged subsets A and B were used as the training set and subset C was used as the test set. Results showed very close correlations between organic carbon content and all measured soil properties. Obtained models relating organic carbon content to mass-specific magnetic susceptibility successfully predicted soil organic carbon contents. [ABSTRACT FROM AUTHOR]