THE H SAF EPS-SG MICROWAVE IMAGER AND MICROWAVE SOUNDER GLOBAL PRECIPITATION PRODUCTS

P. Sanò¹, D. Casella¹, A. Camplani¹, L. P. D’Adderio¹, S. Sebastianelli¹, G. Panegrossi¹

¹CNR-ISAC

Precipitation is an essential element of the global hydrological and energy cycles, and its measurement is of great importance in a variety of research areas, such as climate studies, management of water resources, natural hazards and hydrology. A significant effort in the study of the precipitation is carried out in Europe within H SAF (Satellite Application Facility on Support to Operational Hydrology and Water Management), a EUMETSAT program designed to deliver satellite products (precipitation, soil moisture and snow parameters) for hydrological applications to research and operational users worldwide. H SAF precipitation products are based on the exploitation of present and future Low Earth Orbit (LEO) satellites carrying cross-track and conically scanning passive microwave (PMW) radiometers orbiting around the globe. The development of precipitation retrieval techniques, as well as the quality assessment of satellite precipitation estimates around the globe, can benefit from the availability of unique spaceborne cloud and precipitation radar observations: the NASA/JAXA GPM Core Observatory Dual-frequency Precipitation Radar (DPR) , available since March 2014, and the W-band Cloud Profiling Radar (CPR), onboard the NASA CloudSat (from 2006 to 2023) and currently the ESA/JAXA EarthCare missions. The DPR and CPR have demonstrated their complementarity in precipitation monitoring with the DPR more suitable for medium to intense precipitation regimes, and CPR for snowfall.

Within H SAF, two new machine learning (ML) based precipitation retrieval algorithms for the MicroWave Imager (MWI) and MicroWave Sounder (MWS) radiometers on board the European MetOp Second Generation satellites have been recently developed. The two algorithms are designed to be used operationally as the EPS-SG MWI and MWS surface precipitation rate Level 2 day-1 products. Moreover, they will complement the H SAF Level 2 precipitation rate products for all current and future passive microwave radiometers with precipitation sensing capabilities which are used as input in H SAF multisensor MW-based and MW/IR products.

The algorithms have been developed by using observational databases built coupling surface precipitation rate estimates from radar profiles with coincident (in time and space) radiometer brightness temperatures (TB) measurements. Separate modules for detecting and estimating snowfall and mixed phase/rain precipitation have been developed considering CPR and DPR as reference, respectively, in order to fully exploit the strengths of each radar. Regarding the PMW observations, the observational datasets have been built by using the GPM Microwave Imager (GMI) and Advanced Technology Microwave Sounder (ATMS) measurements, which are the two radiometers having similar characteristics in terms of channel frequencies to MWI and MWS. In addition to the observed TBs, the algorithm uses ECMWF model-derived variables as additional inputs. Different ML architectures were tested in order to select the most suitable one for optimising the algorithm’s performance.

The details concerning the coincidence datasets creation, the design of the ML modules and results of both MWS and MWI algorithms testing are presented.