Modelling soil detachment rates in rainfed agrosystems in the south-central Pyrenees

Distributed erosion models are potential tools for identifying soil sediment sources and guiding efficient Soil and Water Conservation (SWC) planning. However, the uncertainty of model predictions has yet to be resolved. Splash erosion is one of the most important mechanisms in soil loss. In this study, monthly splash detachment rates were predicted using the Morgan, Morgan and Finney (MMF) empirical erosion model and the more complex Revised Morgan, Morgan and Finney (RMMF) erosion model. These two models were used to assess active and abandoned fields in the Spanish Pyrenees. Land uses were barley fields, pasture, recently and old abandoned fields. Input parameters assessed were rainfall characteristics, soil properties, land forms, and land cover. The splash detachment rates predicted by the MMF and the RMMF models were higher for barley fields than for pasture and abandoned fields. However, the more complex RMMF model predicted lower splash detachment rates, especially in pastures. In contrast, runoff detachment was highest in old abandoned fields although rates were much lower than those of splash detachment. Moreover, soil detachment by runoff was low or equal to zero from November to May for the different land uses since the soil remained unsaturated during this period as a consequence of low rainfall intensities and soil surface roughness. Monthly values for total detachment were highest in barley fields, reaching a maximum of 17.2 and 16.9 mg ha-1 in September and October. The mean annual detachment rates for barley, pastures and recently and old abandoned fields were 81.1, 0.8, 61.8 and 22.3 mg ha-1, respectively. The splash and runoff detachment rates of the RMMF model appeared to be sensitive to land cover factors, rainfall intensity and soil micro-topography, thus it is a better model for assessing soil detachment for various land uses. The comparison of erosion rates between the 137Cs and the MMF and RMMF models shows that the models predict lower erosion rates due to the low estimated rates of the runoff transport capacity. However, the estimated and measured rates are in close agreement and are under the limit of the tolerable soil loss for soils under Mediterranean conditions.