Linking rocky desertification to soil erosion by investigating changes in soil magnetic susceptibility profiles on karst slopes

Karst rocky desertification and soil loss has long been regarded as a major agricultural, ecological and environmental problem in many regions around the world, especially in Southwest China. Although many researches have discussed the causes of karst rocky desertification and its driving mechanisms, and established the relationship between the development rate of rocky desertification and soil erosion intensity, almost all these studies only examine changes in area and lack supporting data from changing soil profiles. Fortunately, magnetic susceptibility provides us with a credible way to link rocky desertification to soil erosion by identifying soil profiles and analyzing pedogenic processes, soil redistribution and surface characteristics. In this study, three sampling areas, including farmland, grassland and forest, were established on a karst slope in Southwest China, and the surface characteristics (slope gradient, soil thickness and rock distribution) were measured and analyzed. Soil cores were collected along multiple transects within each sampling area, and the samples were evaluated for magnetic susceptibility and particle size composition. The results showed that the soil magnetic susceptibility (chi(1f) and chi(fd)%) in farmland, grassland and forests were markedly different, and the magnetic susceptibility of regions with natural vegetation cover was higher than that of agricultural regions. The similar chi(1f) profiles indicated that entire slopes shared the same soil development environment and pedogenic process, and the chi(fd)% profiles indicated that the loss of fine soil grains from the grassland and forest slopes were severe. Based on these chi(1f) patterns, soil erosion or sedimentation on the slopes could be effectively identified. The differences in the chi(1f) values at a depth of 0-40 cm reflected soil movement on the slopes. Overall, the soil redistribution along the slopes showed obvious heterogeneity. The chi(1f) values were strongly linked to karst slope characteristics. In the regions where bedrock was widely distributed and the soil layer was thinner, the chi(1f) values were clearly smaller. In contrast, in the regions with little or no exposed bedrock and a thicker soil layer, the chi(1f) values were significantly higher. Although tillage and soil redistribution had an impact on the relations between chi(1f) values and slope surface characteristics, the results indicated that exposed bedrock and thinner soil in the karst slope were primarily caused by soil loss. This study provides valuable research information on the relationship between rocky desertification and soil erosion based on data obtained directly from soil profile changes rather than comparisons of surface area changes. In future research, however, distinguishing changes in soil magnetic susceptibility caused by erosion of those caused by pedogenesis and further relating the changes to soil loss or sediment deposition rates quantitatively are the key challenges for developing an effective tracing method using the magnetic susceptibility properties.