GOES Radar Estimation via Machine Learning to Inform NWP (GREMLIN)

GREMLIN (GOES Radar Estimation via Machine Learning to Inform NWP) provides synthetic radar reflectivity fields from GOES Advanced Baseline Imager (ABI) radiances and Geostationary Lightning Mapper (GLM) lightning data fusion using a convolutional neural network trained to match the Multi-Radar Multi-Sensor (MRMS) composite reflectivity product. The original use case for GREMLIN was initializing convection in high-resolution convection-allowing numerical weather prediction (NWP) models. GREMLIN radar reflectivity also assists forecasters in areas lacking good ground-based radar coverage. GREMLIN is currently being run on ABI Full Disk for GOES-16 and GOES-18.

Goals:

• To extend GREMLIN to the entire Geo-Ring of satellites (Himawari, GEO-KOMPSAT, MTG FCI) using lightning information from MTG Lightning Imager and the World Wide Lightning Location Network (WWLLN).