Bayesian Analysis of the Detection Performance of the Geostationary Lightning Mappers

Abstract Geostationary Lightning Mappers (GLMs) aboard Geostationary Operational Environmental Satellites ( GOES-16 / GOES - 17 / GOES - 18 ) have collectively provided over 5 years of continuous observations of total lightning for the Western Hemisphere. Quantifying GLM’s detection performance is a necessary step toward accurately interpreting the data and analyzing it with lightning detections from other sensors and networks, including the new Lightning Imager (LI) on Meteosat-12 . This study compares lightning observations from the GLM sensors with each other and with ground-based reference sensors including Earth Networks Total Lightning Network (ENTLN) and Global Lightning Dataset (GLD360). ENTLN and GLD360 detection performance throughout the GLM field of view is both imperfect and imperfectly known; thus, we adopt a Bayesian approach that estimates the upper limit of the absolute detection efficiency (ADE) of each system being analyzed. The results illustrate the geographical pattern and diurnal cycle of GLM performance. GLM ADE is consistently high near the center of its field of view but decreases steadily beyond ∼4000 km from the satellite subpoint. ADE is also lower in areas with frequent inverted polarity thunderstorms. Aggregate ADEs (averaged over all flashes) are 82.6% (GLM16), 72.3% (GLM17), and 82.5% (GLM18). GLM18 is more sensitive than GLM16/17 throughout its field of view, but GLM17/18 aggregate ADE is biased lower because ∼90% of the lightning they observe is more than 4000 km from the satellite subpoint, where GLM performance decreases. Each GLM exhibits a pronounced diurnal cycle, with ADE ∼11%–16% lower during the day.