Magnetic susceptibility as a proxy for soil contamination: cross-variogram analysis across contrasting land use areas

The paper presents a study on spatial cross-correlations between soil magnetic susceptibility and the concentration of selected elements in anthropogenically polluted soils. Determining spatial cross-correlations is crucial for accurate estimation and validation of pollution patterns identified via rapid magnetometric screening. Geostatistical methods based on cross-correlations can reduce the need for costly geochemical analyses and make it possible to analyze also the spatial aspect of phenomena that is not possible with classic metrics such as the Pearson correlation coefficient. The analysis was conducted in areas of southern and southwestern Poland with diverse land cover types and varying degrees of anthropogenic pressure. The study confirmed spatial correlations between magnetic and chemical measurements, especially in soils that have remained relatively undisturbed over time. The spherical model with a nugget effect best described the variograms and cross-variograms of magnetic susceptibility and concentrations of As, Cd, Cr, Co, Cu, Hg, Ni, Pb, Se, Sb, Tl, and Zn in soil. The results showed significant differences in cross-variogram parameters depending on land cover. Observed cross-correlation ranges differed depending on land type and were equal up to 550 m (average 280 m) in industrial and post-industrial areas, up to 300 m (average 170 m) in arable fields, and up to 145 m (average 67 m) in forests and parks.