GFDL Atmospheric Model flash frequency parameterization insights from the Geostationary Lightning Mapper

To quantify seasonal lightning flash frequency (ff), we compare the novel GOES-16 (G16, 130-20°W and ±55°N/S) and -17 (G17, 160°E-70°W and ±55°N/S) Geostationary Lightning Mappers (GLMs) ff (flashes/km2/second) observations to the GFDL Atmospheric Model version 4 (AM4) flash frequency output. Through these comparisons we quantify model parameterization and observational biases. These biases will inform model development of ff parameterizations, lightning-nitrogen oxide parameterizations, and subsequently refine estimates of tropospheric ozone and methane within the GFDL’s AM4. Observation and model comparisons suggest that observed ffs over both GLM field of views (FOVs) are 2-3 orders of magnitude greater than the cloud top height (CTH) ff parameterization estimates in the GFDL AM4. Correlations were greater than 0.50 for both the land model grids and for all of the model grids in DJF and JJA over both GLM (FOVs). Comparing G16 GLM ff to the multiplication of ERA5 convective available potential energy (CAPE) and total precipitation rate developed by Romps et al. (2014) produces correlations greater than 0.6. Using Global Precipitation Mission IMERG total precipitation observations increased JJA correlations to greater than 0.70. Comparisons of G16 GLM ff to MERRA2 3D flux of ice convective precipitation suggest correlations are as high as 0.81 in SON over land. Correlations are generally poor (> 0.50) over the G17 GLM domain for these ff parameterizations. Experiments for cloud ice water path (Han et al. 2021) and environmental variable (Stolz et al. 2017 and 2021) ff parameterizations are ongoing.