A New Lightning Simulation Scheme that Incorporates GOES-16 Data for Implementation in the NASA GMAO GEOS Model.

Lightning parameterization schemes in Earth System Models (ESMs) are relatively under-developed and many of these schemes rely on proxy variables, such as cloud top height. Recent studies show that cloud top height is not a good predictor of the amount and location of lightning and their associated lightning-produced nitrogen oxides (NO_x), particularly in a changing climate. Our goal is therefore to develop a new lightning simulation algorithm that uses infrared (IR) channels from the NOAA Geostationary Operational Environmental Satellite’s (GOES)-16 Advanced Baseline Imager (ABI) and implement it in the NASA Global Modeling and Assimilation Office’s (GMAO) Goddard Earth Observing System (GEOS) ESM (hereafter the “GEOS model”). The aim of this new scheme is to improve the location and amount of simulated lightning flashes compared to current lightning parameterization schemes in the GEOS model.

As such, IR brightness temperatures from the GOES-16 ABI instrument were co-located with the latitude and longitude location of the first lightning group per flash from the GOES Geostationary Lightning Mapper (GLM) during the months of March to October 2020. Statistical information from these data points were then obtained and used in developing the new GMAO GOES lightning simulation (GGLS) algorithm.

Two different GGLS algorithms were created, one that covers only the contiguous United States (CONUS) region, and one that covers the Full GOES-16 domain (hereafter FULL). The reason for developing two different algorithms was that the CONUS algorithm would eventually be incorporated into the “stretched” grid of the GEOS model over the CONUS region, with the resolution varying between 2 and 12 km, while the FULL algorithm would be applied to the full GOES-16 domain in the GEOS model with a 50 km resolution. The GGLS algorithm was then evaluated by comparing the results to the GLM observations for three different geographical domains: (1) CONUS, (2) FULL, and (3) South America (SAM), which incorporated the same algorithm as for the FULL domain and was therefore just a subset of the FULL domain.

The results showed that the new GGLS algorithm outperformed the current lightning parameterization schemes from the GEOS model for at least 75% of the cases over the CONUS domain, for 100% of the cases for the FULL domain, and for 95% of the cases over the SAM domain. Furthermore, there was an average maximum improvement for all domains (CONUS, FULL, and SAM) of over 30% for all statistical tests, while the largest improvements occurred over the SAM domain.The advantage of our algorithm is that it limits the lightning that is simulated over the tropical oceans and does not take cloud top height into account.

Our next step is to implement this new GGLS algorithm in the GEOS model and test its impact on the NO_x simulations. We plan to extend the algorithm to regions covered by GOES-18 and EUMETSAT’s Meteosat Third Generation, expanding the use of satellite IR fields to simulate lightning in the GEOS model.