Assimilation of Lightning Data Using a Nudging Method Involving Low-Level Warming

This paper uses the Weather Research and Forecasting (WRF) model to assimilate lightning observations from the Earth Networks Total Lightning Network at 9 km grid spacing. This resolution mimics the grid spacing of the Geostationary Lightning Mapper (GLM) that will be onboard GOES-R. The assimilation method warms the most unstable low levels of the atmosphere at locations where lightning is observed but deep convection is not simulated based on the absence of graupel. Results from the assimilation (3 km grid spacing) are compared with a control simulation and a simulation employing a lightning assimilation method developed by Alexandre Fierro and colleagues. Their method increases relative humidity according to a nudging function that depends on the intensity of observed lightning and simulated graupel mixing ratio. Results will be presented from three separate severe thunderstorm cases across the Central and Eastern United States during 2011. Both subjective comparisons and objective statistical metrics are used to compare simulated precipitation with hourly NCEP Stage IV observations. Results show that both assimilation methods improve the simulated precipitation during the assimilation period and for a short forecast period afterwards. The low-level warming approach often is more effective at initiating deep convection than is increasing the humidity. Additionally, the assimilation methods have contrasting effects on average temperature and humidity in subsequent forecasts. When not using a Newtonian nudging coefficient, both methods excite strong acoustic waves that have the potential to cause numerical instabilities, thereby indicating the benefits of gradual nudging. The paper also explores altering the time window of observations and the metric denoting simulated deep convection.