Abstract. Soil loss is commonly estimated using the Revised Universal Soil Loss Equation (RUSLE). Since RUSLE is an empirically based soil-loss model derived from surveys on plots, the high spatial and temporal variability of erosion in Mediterranean environments and scale effects mean that it is necessary to evaluate the model in other spatial units such as the microcatchment. In this study, a series of topographic and soil surveys was carried out on a microcatchment of 6.7 ha in a mountainous area under no-tillage farming with bare soil in order to examine spatial and temporal results produced by RUSLE. GPS measurements of the microrelief height differences were used in a control area in the microcatchment to compare observed erosion and deposition with RUSLE predictions. Erosion points located in certain areas correlate very closely with RUSLE predictions, while the distribution of deposition points showed no correlations with RUSLE predictions. Secondly, a time series of daily rainfall data was used to calculate annual erosivity values, which were fitted to an appropriate distribution function. It was determined that the rainfall distribution best fitted the Pearson type III distribution function. Next, efforts were made to quantify the long term erosion and to check the suitability of the land-use and management under different thresholds of tolerance. It was found that values of erosivity in the study area with a return period of 10 years generate a mean annual erosion of 5 t ha⁻¹ yr⁻¹. On the study scale, RUSLE allowed us to locate the most erosive areas and to combine the suitability of the soil land-use and the management with the frequency of the annual erosivity. In addition, an annual sediment delivery ratio of approximately 47% was estimated for the period 2005–2006.