Long-Term Assessment of the DPR Rainfall Products in the Mediterranean Area According to the H-SAF Validation Protocol

The H-SAF project (Satellite Application Facility on support to Operational Hydrology and Water Management, funded by EUMETSAT) is aimed at retrieving key hydrological parameters such as precipitation, soil moisture and snow cover. Within the H-SAF consortium, the Product Precipitation Validation Group (PPVG) evaluate the accuracy of instantaneous and accumulated precipitation products with respect to ground radar and rain gauge data adopting the same methodology (using a Unique Common Code) throughout Europe. The adopted validation methodology can be summarized by the following few steps: (1) ground data (radar and rain gauge) quality control; (2) spatial interpolation of rain gauge measurements; (3) up-scaling of radar data to satellite native grid; (4) temporal comparison of satellite and ground-based precipitation products; and (5) production and evaluation of continuous and multi-categorical statistical scores for long time series and case studies. The statistical scores are evaluated taking into account the satellite product native grid.

With the recent advent of the GPM era starting in march 2014, new global precipitation products are available. The validation methodology developed in H-SAF can be easily applicable to different precipitation products. In this work, we have validated the precipitation rate products estimated by the DPR (Dual-frequency Precipitation Radar) instrument onboard of the GPM-CO (Global Precipitation Measurement Core Observatory) satellite. More specifically, we have analyzed the near surface and estimated precipitation fields collected in the 2A-Level for 3 different scans (NS, MS and HS).

The Italian radar mosaic managed by the National Department of Civil Protection, available every 10 minutes, was used as ground reference. Accurate post-processing of the ground-radar estimates was devoted to filter out extemporaneous anomalies not managed by the real-time processing system. Furthermore, the estimated ground-data quality was used to constraint the validation on the most reliable reference samples.

The assessment analysis, carried out on a two-year period, highlighted the capability of the DPR to properly identify the precipitation areas. Generally, the higher accuracy was found in stratiform precipitation (especially for the HS product), whereas an underestimation of the rainfall rate was found for some convective events.