Relationships among the physical and chemical properties of soil, vegetation and land degradation in semi-arid environments

This study examines the spatial patterns of soil parameters to test the hypothesis that shrub encroachment initiates a change in scale of soil heterogeneity, which consequently influences a landscape's biotic and abiotic interactions and thus the susceptibility of soil to erosion. Grassland, shrubland and badland sites were established in two semi-arid environments; the Karoo, South Africa and the Sevilleta National Wildlife Refuge, New Mexico, U.S. 108 soil samples from each of the eleven 60m x 60m plots were analysed for bulk density, shear strength, texture, aggregate stability, organic matter content, pH, conductivity and available sodium, calcium, magnesium, potassium and phosphorus content. Geostatistical analyses determined that, at a scale representative of the vegetation community, the grassland landscape appeared relatively homogeneous in its distribution of soil parameters. Shrublands, however, demonstrated an increase in heterogeneity of all soil parameters. Periodicity in the semi-variograms indicated that regular patterns across the landscape were evident for all parameters and thus likely to represent the differences between shrub and intershrub regions. Due to the complex plant-soil interactions, and the interactions amongst the soil parameters themselves, the cyclic patterns represent areas of high and low erodibility. More pronounced patterns were identified in the badlands. This indicates that, if the conditions are right, changes in plant-soil interactions caused by soil parameter redistribution in shrubland landscapes can exacerbate erosion leading to further degradation in the form of badlands. Comparisons between the two semi-arid regions suggest that although local variations in soil type and different species of vegetation will affect the intensity of the spatial response, the underlying patterns are similar at both locations and hence, potentially, at a global scale.