An Evaluation of the Efficacy of Using Observed Lightning to Improve Convective Lightning Forecasts

Abstract

Abstract Dynamic lightning forecasts [total and cloud to ground (CG)] were produced on a convection-allowing forecast grid with 4-km grid spacing with lightning assimilation (ASML) and without lightning assimilation [control (CNTL)]. A dynamic scheme produces time- and space-dependent potential electrical energy, which then converts this energy into lightning (e.g., number per hour per grid element). The assimilation scheme uses observed, gridded total lightning to determine how much water vapor is added at constant temperature in the mixed-phase region, leading to a convective response. ASML and CNTL lightning forecasts were compared to observed total and CG lightning. Four case studies—each representing a different type of convective regime—demonstrate that the spatial distribution and intensity of forecast lightning were improved when lightning assimilation was used. Over 3 days in March 2012, eight 18-h lightning forecasts quantified the advantages in forecast accuracy. Equitable threat scores for forecast CG lightning associated with strong [25 (3 h)−1], very strong [50 (3 h)−1], and extreme [100 (3 h)−1] events were significantly more accurate for convective storms that developed in forecasts with lightning assimilation than without. Improvements in forecasts of very strong and extreme events occurred out to 9 h of forecast time, while the forecasts of strong events showed improvement out to 15 h. Spurious convection was removed with filtering in one case study, which led to a notable improvement in the timing and intensity of the squall line. Sensitivity tests examined the utility of this filtering approach, and the importance of reducing mass imbalances liable to occur when too much water vapor mass is introduced into the model.