Improving GOES derived Rain Areal Coverage Estimate by using a Hybrid WSR-88D and GOES Classification for Deep Convective Systems

GOES satellite date have been playing an increasingly important role in supporting timely weather forecasts for the public, and the cloud products derived from these data have also been widely used in studying our climate systems. While progress has been made for improving the quality of the GOES derived product in precipitating areal coverage estimate and severe weather forecast, significant uncertainties remain in the presence of deep convective systems (DCS) when passive optical sensor cannot penetrate deep into these clouds.

We have developed a hybrid classification technique to combine WSR-88D ground radar network and GOES satellite data to separate various precipitating deep cloud and non-precipitating anvil regions of a DCS. This technique builds upon a 3-D radar classification algorithm that separates convective and stratiform rain, transition deep clouds, mixed and ice phase anvils, and other thinner cirrus clouds. Preliminary results show that current GOES cloud retrieval algorithm over all precipitating deep clouds and non-precipitating thick anvils (mixed and ice phase) have similar signatures in Infra-Red (IR) and Visible channels, which can lead to errors in GOES satellite instantaneous precipitation retrievals. Other GOES microphysical retrievals, such as ice water path and optical depth, show more substantial differences between ice anvils and precipitating clouds. This suggests possible improvements in identifying precipitating clouds by including additional retrieved microphysical parameters for GOES cloud classification. Further, the hybrid classification can continue to provide active-sensing-based cloud classification for evaluating future precipitation and cloud product improvements in GOES-R satellite.