Influence of the Rainfall Time Step on the Thresholds for Separating Erosive and Non-erosive Events

In previous work [1] Todisco et al. (2019), Todisco et al. (2019) analyzed 522 rainfall events from 2008 to 2017 at the Masse experimental station (central Italy) to define and evaluate several thresholds of rainfall characteristics able to classify non-erosive and erosive events. The analysis was performed based on a 5-min rainfall dataset. In this study, working on the same dataset, we evaluated how the value and performance of the thresholds change when their determination is made based on rainfall records at different time steps. In particular, the original 5-min data were aggregated at 30 min, one of the typical timesteps of the data provided by the Hydrographic Services and that required for calculating the USLE erosivity factor. The results indicate that some rainfall characteristics maintain effectiveness, passing from a 5-min to a 30-min rainfall dataset. However, only the total event rainfall has a threshold (~ 15 mm) that is almost independent of the resolution, while for other effective variables (e.g., Maximum duration of an individual run and the Maximum rainfall amount in a burst), the thresholds change significantly. Moreover, some variables dependent on the number of runs or showers during the event become entirely ineffective when switching to the 30-min resolution due to the consequent flattening of the internal dynamics of the hyetographs.