Evaluation of CMORPH skills in capturing rainfall extreme events: A case study in Mignone river catchment.

Rainfall data are the main input of hydrological and flood forecasting models and as such their precision is crucial. Our knowledge on precipitation, however, relies mainly on rain gauge networks, which have been declining in the past decades. In the last 20 years many satellite precipitation estimates have been developed, with different temporal and spatial resolution and with a quasi-global coverage. Thanks to their wide availability they can be a valid alternative to rain gauge observations in hydrologic applications. Although satellite performances have been tested in the last years, our knowledge on their potentialities is still incomplete. In this paper we test the skills in detecting heavy rainfall events of the Climate Prediction Center MORPHing (CMORPH) precipitation estimates over a medium-sized catchment in central Italy. We use Normalized Standard Error (NSE), Mean Absolute Error (MAE) and a scatter plot comparison to test CMORPH on a dataset of 75 severe rainfall events. Our preliminary results show that in 63% of the cases the satellite product underestimates the total rainfall depth occurred during an event and no clear trend emerged for MAE and NSE. Each point represents a precipitation event. [ABSTRACT FROM AUTHOR]