GOES GLM, biased bolides, and debiased distributions

The large combined field of view of the Geostationary Lightning Mapper (GLM) instruments onboard the GOES weather satellites makes them useful for studying the population of other atmospheric phenomena, such as bolides. Being a lightning mapper, GLM has many detection biases when applied to non-lightning and these systematics must be studied and properly accounted for before precise measurements of bolide flux can be ascertained. We developed a Bayesian Poisson regression model which simultaneously estimates instrumental biases and our statistic of principal interest: the latitudinal variation of bolide flux. We find that the estimated bias due to the angle of incident light upon the instrument corresponds roughly with the known sensitivity of the GLM instruments. We compare our latitudinal flux variation estimates to existing theoretical models and find our estimates consistent with GLM being strongly biased towards high-velocity bolides.

Automated summary

The Geostationary Lightning Mapper (GLM) instruments, with their large combined field of view, are useful for studying the population of atmospheric phenomena like bolides. However, there are biases when using GLM for non-lightning purposes, which need to be studied and accounted for before precise measurements of bolide flux can be obtained. A Bayesian Poisson regression model was developed to estimate instrumental biases and the latitudinal variation of bolide flux concurrently. The estimated bias corresponds to the known sensitivity of the GLM instruments, and the latitudinal flux variation estimates are consistent with a strong bias towards high-velocity bolides, as compared to existing theoretical models.