Impact of GOES-16 Multi-spectral Satellite Observations on the Identification of Precipitation Typology using the Multi-Radar/Multi-Sensor System

Abstract:

The unprecedented high spatial, temporal, and spectral resolutions from the new generation of geostationary Earth orbit (GEO) satellites provide opportunities to improve our understanding of precipitation processes. The objective is to evaluate the impact and advantages of the multi-spectral observations from the ABI sensor onboard the GOES-16 satellite in identifying precipitation types from the Ground Validation Multi-Radar/Multi-Sensor (GV-MRMS) system. A machine learning (ML) based classification is developed with several categories of predictors, such as ABI brightness temperatures (Tb) spectral channel differences and textures, and environmental variables from the Rapid Refresh numerical forecast model (NWP). Experimental setups are designed to understand the impact of each category of predictors and various ABI channels. The ML estimates show promising results for separating broad precipitation types categories such as convective, stratiform, and no-precipitation, and specifically for identifying hail or cool stratiform types. Convective types are better detected using GOES-16 derived predictors and the detection of stratiform types is significantly improved with the addition of NWP predictors. Simple Tbs detect no-precipitation and hail types correctly, whereas Tb textures significantly contribute to the classification accuracy of warm stratiform and convective precipitation types. The analysis also shows that, by adding one water vapor absorption (6.2 μm) channel to the heritage 11.2 μm channel on geostationary satellite platforms, the overall classification accuracy improves from 39% to 54%; it improves further to 76% with 5 channels (6.2μm, 11.2μm, 12.2μm, 7.3μm and 8.5μm). This study sets up a benchmark on the accuracy that can be achieved with historical GEO sensors operating with one or two channels w.r.t. the new generation satellite.