Linking land use changes to variation in soil properties in a Mediterranean mountain agroecosystem.

Abstract Several decades of intensive rainfed farming in Mediterranean mountains and later land abandonment has led to rapid land use and land cover changes. During recent centuries, the conversion of rangelands into croplands has increased the surfaces prone to erosion. In the southern Pre-Pyrenees, the process was reversed during the middle of the twentieth century, allowing the recovery of vegetation and subsequent variation in land cover. This work aims to assess how land use changes after generalised land abandonment affect some major soil properties related to soil quality. For this purpose, 98 replicate bulk soil samples were collected in a 23 km2 catchment that was mostly cultivated at the beginning of the last century. Soil samples were distributed over areas representing the main land uses (agricultural land, natural forest, pine afforestation and scrubland). Bulk density, stoniness, grain size, pH, carbonates, electrical conductivity, soil organic carbon (SOC), total nitrogen (TN), water retention capacity and magnetic properties (low frequency magnetic susceptibility (LF) and frequency dependence (FD)) were analysed in the samples from different land use areas. A past scenario was recreated using estimated data from the SPEROS-C model in order to evaluate changes in SOC over time. Furthermore, a multitemporal analysis of the Normalised Difference Vegetation Index of Landsat images was performed between 1972 and the present in order to assess the dynamics of revegetation. After land abandonment, 16.5% of the area remained as croplands, but afforestation and natural revegetation occupied 83.5% of the catchment. The highest mean value for SOC was found in the pine afforested area and the highest TN mean value was found in the natural forest. The lowest mean values for SOC and TN were recorded on the agricultural land. These results show the impact of soil changes produced by land use changes in fragile Mediterranean mountain agroecosystems. Graphical abstract Unlabelled Image Highlights • Changes in land use have resulted in differences in soil properties. • NDVI allows identifying vegetation cover variation developed after land use changes. • Spatial patterns of soil properties are highly conditioned by land use distribution. • Modelling SOC stocks along time is possible applying SPEROS-C model. • NDVI is directly related with the distribution of soil nutrients in the catchment. [ABSTRACT FROM AUTHOR]