Predicting soil erosion with RUSLE in mediterranean agricultural systems at catchment scale
Accurate assessment of soil loss is essential for sustainable agricultural production, management and conservation planning, especially in productive rain-fed agro-ecosystems and protected areas. The European Union considers soil as a non-renewable resource and identifies that soil degradation has strong impacts on soil and water resources. In this work the Revised Universal Soil Loss Equation model was applied within a geographic information system in the Estaña catchment (Spanish Pre-Pyrenees) as representative of a Mediterranean agro-ecosystem to elaborate a map of soil erosion at high spatial resolution (5 x 5 m of cell size). The soil erodibility factor (K) is calculated from three different approaches to evaluate the importance of spatial variations in soil texture, field infiltration measurements (Kfs) and amount of coarse fragments. The average value of estimated soil loss for the whole study area is 2.3 Mg ha-1 yr-1 and the highest rates are estimated in crops in steep areas (5.8 Mg ha-1 yr-1) and trails (18.7 Mg ha-1 yr-1). Cultivated soils with high soil erosion rates (higher than 8 Mg ha-1 yr-1) represent 20% of the cultivated area. The average value of soil loss in areas with human disturbances (4.21 Mg ha-1 yr-1) is 4.4 times higher than that estimated for areas with natural vegetation (0.96 Mg ha-1 yr-1). Field validation with 137Cs shows that the estimated value of soil loss in barley fields with the K-Kfs-rocks factor improves the model predictions in comparison with those obtained with the K-texture and K-Kfs factors. The RUSLE model predicts a decrease in soil erosion in fields in accordance with the increase of the age of abandonment. Predicted values of soil erosion and measured soil organic matter and stoniness in old abandoned fields agree with those in areas of natural forest and indicate the recovery of the original conditions of the soil. Statistical analysis highlights that the C factor contributes most of the variability of the values of predicted soil erosion, the K and LS factors contribute in a similar way and the P factor contributes least to the variability of soil erosion. Cultivated soils developed over clay materials in high slope areas are the most susceptible to soil degradation processes in comparison with soils developed over limestones in gentle and medium slope areas. The recovery of terraces in steep fields and conservation of crop residues are proposed as soil conservation practices to reduce the magnitude of soil loss in the study area. ; This research was financially supported by the following project: "Soil erosion and carbon dynamic in Mediterranean agroecosystems: radioisotopic modelling at different spatial and temporal scales" (MEDEROCAR, CGL2008-00831/BTE) funded by the Spanish Ministry of Science and Innovation. ; Peer reviewed