Using Short-Term Forecasts of Total Lightning to Improve Numerical Weather Forecasts

Data assimilation has been shown to improve short-term convective forecasts and forecasts of longer-lived tropical storms. Standard approaches are to assimilate wind, temperatures, humidity from high resolution observational nets, or to use spatially contiguous radar to force convection in the right locations. Alternatively, observed lightning has been used to supplement these data sources. The advantage of using lightning observations is that they are usually also available where these other data sources are not, and this will especially be the case when lightning observations from satellite becomes available mid-decade. Up to now, however, a typical approach for assimilating radar and/or lightning is to prime the atmosphere for convection at appropriate locations, and then compare the evolution of model forecast parameters other than lighting to observed parameters (like cloud top height, cloud-derived radar echoes). A more direct way is suggested, where the forecast model also predicts total lightning values, which can then be directly and absolutely compared to observed lightning densities. We present case study results using forecast and observed lightning to improve forecast convection in RAP-WRF forecasts at 4 km grid resolution. It is suggested that this approach can potentially improve convective and tropical storm forecasts.