The Robustness of Lightning Proxies for Numerical Weather Prediction and Data Assimilation

The Geostationary Operational Environmental Satellite-R (GOES-R) East and West satellites launched beginning in 2016 provide a high detection efficiency lightning detection system via the Geostationary Lightning Mapper (GLM). The GLM uses a hierarchy system that categorizes observations into events, groups, and flashes. Following this system, a flash is not represented by a singular location but by all the locations that that flash reaches. The number of flashes that reach a given location within a period of time is known as flash extent density (FED) which is one of several high spatial and temporal resolution products that this categorical system allows GLM to produce. While radar is a vital tool used in data assimilation in numerical weather prediction (NWP) models, limitations result in lacking data across mountains, oceans, and dead zones. Although GLM has limitations of its own, such as non-perfect detection efficiency, it avoids the coverage limitations of radar allowing it to provide observations in regions that have insufficient radar data. Lightning detections can indicate storm cores and provide insight into storm intensity, leading to the exploration of GLM’s viability in data assimilation to supplement radar. Empirical relationships between FED and ice hydrometeor-related variables simulated by convection-allowing models, such as column-integrated graupel mass, have been researched to develop potential parameterizations. This study looks at two numerical weather models, the High-Resolution Rapid Refresh (HRRR) and the Rapid Refresh Forecast System (RRFS), and explores the applicability of candidate parameterizations for these models to a diverse set of events within recent years including thundersnow, tropical cyclones, and mesoscale convective systems. Distributions of FED and minimum flash area, another GLM product, are compared across the cases. Furthermore, the relationships between model-derived vertical flux of graupel and column-integrated graupel mass data and observed FED are also compared both amongst the cases and to the proposed parameterizations. With the expected operational implementation of RRFS in 2024, this study provides insight and suggestions for improvements in the empirical relationships that model FED and in the assimilation of lightning data.