5A.2 An evaluation of the efficacy of using observed lightning to improve convective lightning forecasts

Lynn, Barry H., Guy Kelman, Weather It Is, LTD, Gary Ellrod, WxC, LLC

Lightning forecasts were produced on a convective allowing forecast grid with 4 km grid spacing. Observed total lightning was qualitatively compared to forecast total lightning, while observed cloud-to-ground lightning was compared to forecast lightning at different threshold values using statistical measures. The lightning can be considered as a surrogate for convective intensity. The initial statistical comparison was made on a grid with 36 km grid spacing (about 1200 km squared; the size of New York City, NY). Depending on initial conditions, the forecast period over which lightning assimilation improved the forecast lightning varied from a few hours to as long as 15 hours in four case study forecasts and one reanalysis study. Eight (historical)18-hr forecasts were made for 19-21 March 2012. The results on the grid with 36 km grid spacing show that lightning assimilation leads to improved forecasts of cloud-to-ground lightning during the first 3-hr period. Calculating the Equivalent Threat Score for different radii and thresholds, however, showed that lightning assimilation led to more accurate forecasts over smaller areas and over longer time scales. It is apparent that lightning assimilation has the potential to improve statistical scores of strong convective storms throughout the length of a typical convective forecast.