How Data from the Geostationary Lightning Mapper Can Affect Model Output for Thunderstorms (credit: NASA and American Meteorological Society)
One of the first things to come to mind with the word “thunderstorm” is lightning. As is the case with most thunderstorms, the presence of lightning is dictated by the level of instability in the atmosphere and, more intensively, the intracloud interactions of graupel, or supercooled water droplets that rime together but are not quite as solid as hail stones. Keep in mind that these processes occur well above the surface of the earth (approximately anywhere from 35,000 to as high as 70,000 feet above the surface) and are captured by the Geostationary Lightning Mapper (GLM) onboard the Geostationary Operational Environmental Satellite (GOES). These lightning observations can provide crucial evidence into the development/maturity of a thunderstorm over the course of its lifetime. But this information is most important for the forecasters who ultimately digest the information outputted by the models to come up with as highly accurate a public forecast as possible.
To understand the importance of how assimilating lightning data models can potentially enhance a forecast, we must first understand how models ingest this lightning data. Now, data assimilation is a complex topic that, for brevity, I will summarize in just a few sentences. The general idea of assimilating data into a model is to be able to provide extra information otherwise unknown to the model so that it can compute a forecast (known specifically as a posterior). A weather model is only as good as the assumptions it has at the “analysis” time, or start time, so more data could give the model more to consider as it integrates through time. The data that goes into these models must be quality controlled as the amount of noise that the raw data can contain can alter the model’s perception of the state of the atmosphere. In other words, if a model feeds on bad data, it will produce a bad forecast, and this occurs in a compounding loop until the quality of the data is improved on. Regarding assimilating lightning data, the model can use the available data to help it evolve thunderstorms in the short-term. Whether they could intensify or weaken could be highlighted in clues within the storm’s total lightning count (more precisely, the flash extent density or FED). The FED is the most common GLM-derived product to be assimilated as it serves to show convective intensity and is often correlated against graupel interactions within the cloud, so it serves as a good example of an intensifying or weakening storm.
Studies completed over the last few years have shown an appreciable gain of ingesting quality-controlled lightning observations from the GLM. For instance, a study by Fierro et al. (2019) examined the benefits of assimilating GLM lightning-derived water vapor data and radar data during short-term forecasts (usually no further than 6 hours in time). They found modest improvements in the accuracy of the forecasts. Similar studies of organized supercell cases (e.g., Fierro et al. 2016; Mansell 2014) have also demonstrated improvements of forecast accuracy and statistical gains for success scores. In short, there is much to be learned (and possibly gained) from assimilating GLM lightning observations!
A great article to understand more about the lightning process is given here by the folks at the National Weather Service: https://www.weather.gov/jetstream/lightning
To learn more about other weather research topics, be sure to click here!
Image credit above: NWS Jetstream
© 2020 Meteorologist Brian Matilla