Precipitation retrievals from the SSMIS using the PRPS scheme: formulation, validation and intercomparison.

The US Department of Defense (DoD) Meteorological Satellite Program (DMSP) has provided a long-term record of passive microwave observations from the Special Sensor Microwave/Imager (SSM/I) and the Special Sensor Microwave Imager/Sounder (SSMIS). These observations, available from 1987 to the present, provide the backbone of data used for global precipitation measurements. The SSM/I and SSMIS instruments have similar lower frequency channels (19.35-85.0 GHz vs 19.35-91.655 GHz), with the SSMIS having higher frequency channels at 150 GHz and three around 183.31 GHz.

The Precipitation Retrieval and Profiling Scheme (PRPS) is a retrieval scheme designed to be efficient and avoid the use of any external dynamic data sets, such as model information. This is particularly important for a truly independent data product that can be used for evaluating model performance. The PRPS was originally designed for use with cross track sounding instruments but has been adapted to other passive microwave sensors: here it has been adapted to utilise the SSMI and SSMIS Fundamental Climate Data Records generated by the EUMETSAT CM SAF. The PRPS-SSMIS relies upon an observational a priori database derived for each sensor paired with a database index file to provide a computationally efficient retrieval scheme.

This poster will present an outline of the PRPS-SSMIS scheme together with the validation and intercomparison of the resulting precipitation products. At present the databases for the retrieval scheme are based upon 7 years of observations (2016-2022) from SSMIS sensors on the F16, F17, F18 DMSP satellites, matched to co-incident and co-temporal measurements of precipitation from the Global Precipitation Measurement (GPM) mission’s Dual frequency Precipitation Radar (DPR). Comparisons are made at a number of scales: ‘climate’ scale comparisons are made against the GPCP v3.2 global precipitation product, through to instantaneous precipitation retrievals which are compared with surface radar over the US and Europe. In addition, comparisons are made with the Ferraro and GPROF precipitation products to assess consistency with other estimates. Overall, the PRPS-SSMIS retrievals tend to underestimate the precipitation, primarily due to the internal assumptions in the retrieval scheme as a result of the skewed distribution of precipitation occurrence and may easily be corrected. Correlations between the PRPS-SSMIS products and the GPCP are similar to those of the GPROF-SSMIS products, particularly when a comparable spatial resolution is used. Both the GPROF and PRPS scheme outperform the Ferraro precipitation product in terms of bias and correlation and are more consistent over time.