The Impact of Single Group Flashes on Geostationary Lightning Mapper Data

Abstract

AbstractThe Geostationary Lightning Mapper (GLM) is an instrument on the NOAA Geostationary Operational Environmental Satellites‐R (GOES‐R) Series. The GLM data are processed by the Lightning Cluster Filter Algorithm which filters the GLM data and clusters it into a series of events, groups, and flashes. One of the current filters in the GLM algorithm (as of June 2023) removes all flashes that have only a single group. Combining the flashes (including the flashes with only single group) from two of the GLMs (the GLM on GOES‐16 and the GLM on GOES‐17) in the region where they overlap produces a unique data set of coincident flashes (detected by both GLMs) and non‐coincident flashes (detected by only one GLM). Coincident flashes detected by both GLMs are very unlikely to be from noise sources. This data set allows us to estimate the impact of the single group flashes on the Detection Efficiency (DE) and the False Alarm Rate (FAR). We find that the single group flash filter does greatly improve the FAR but does slightly decrease the DE. Subsequent analysis shows that some of the removed flashes with single groups are from true lightning. A simple modification to the single group flash filter to only remove flashes with a single event increases the FAR without a corresponding improvement in DE. A more complex algorithm is needed to recover the single group flashes that are likely from lightning without adding back single group flashes that are from noise sources.