The GLM ground flash fraction retrieval algorithm: improvement, testing, and demonstration the focus of this project is to extend, optimize, and rigorously evaluate an existing algorithm that…

The GLM ground flash fraction retrieval algorithm: improvement, testing, and demonstration the focus of this project is to extend, optimize, and rigorously evaluate an existing algorithm that retrieves the ground flash fraction in a set of n flashes observed by the geostationary lightning mapper (GLM). The ground flash fraction (which provides the ratio of the # of cloud flashes to the # of ground flashes, or so-called z ratio ) has significant utility in a variety of scientific areas such as: lightning nitrogen oxide (NOX) production for air quality & global chemistry/climate modeling, severe weather studies, lightning-convection relationships, the contribution of lightning to the global electric circuit, and cross-sensor validation. The existing algorithm was developed by previous R3 efforts. IT employs a bayesian retrieval methodology. IT also employs a mixed exponential distribution model for describing the statistical nature of cloud-top lightning optical characteristics. The retrieval algorithm was tested using simulations, and was directly applied to optical transient detector (OTD) data. The retrieval results from the OTD study were compared with the ground flash fraction inferred from national lightning detection network (nldn) data to assess retrieval errors. The overall results were encouraging, but more work is needed to fully optimize the algorithm and to build the highest level of confidence in its operational application. In the proposed project, four tasks will be performed. First, the existing bayesian retrieval algorithm will be improved by employing a parameter search initialization that is based on the grobner-basis solution scheme; improvements will be quantified using simulations and the previous OTD/nldn datasets. Second, the grobner-modified algorithm will be tested and evaluated by applying IT to lightning imaging sensor (LIS)/nldn datasets. Third, a journal paper will be completed by the end of year 2 that fully documents the retrieval algorithm and that serves as the foundation to an algorithm theoretical basis document (atbd). Fourth, as part of the optional third year, a visual demonstration of the algorithm product within the proving ground will be provided. In particular, IT will be shown how ground flash fraction information can be handed off to the NASA msfc lightning nitrogen oxides model (lnom) to improve the operational accuracy of the community multiscale air quality (cmaq) model.