Alex presented a seminar “A cloud-scale lightning data assimilation technique implemented within the WRF-ARW model: Proof of concept and real-time evaluation”
To improve the analysis and short-term forecasts of convection, a new technique for assimilating total lightning data into the WRF-ARW model at cloud-resolving scales has been developed and evaluated over a large number of thunderstorm days. Assimilated lightning data forces deep, moist precipitating convection to occur via a nudging function for the total lightning data. This computationally inexpensive, smooth continuous function locally increases the water vapor mixing ratio and virtual buoyancy within the graupel-rich mixed phase region of storms at observed lightning locations. The assimilation of gridded pseudo-GOES-R resolution (9 km) flash rate via EarthNetworks® total lightning data for only a few hours prior to the forecast initialization significantly improved the representation of the convection at the initial analysis time and during the subsequent 1-6 forecast hours within the convection-permitting (≤ 5 km) and-resolving (≤ 2 km) grids.
Subsequent evaluation of this lightning assimilation algorithm against a standard cloud-scale 3DVAR technique (assimilating radar data) for the case of the 29 June 2012 “super derecho” event will also be briefly reviewed. This talk will end with an overview of the statistics derived from the implementation of this lightning assimilation scheme into the real-time 4-km CONUS NSSL-WRF forecast testbed (spanning~70 forecast days). These modeling exercises revealed, in particular, that the greatest benefit from the assimilation of lightning data arose from improved forecasts of nocturnal outflow dominated mesoscale convective systems, which evolution is dictated by the initial location and strength of the cold pool.
The presentation is available on the anonymous ftp site: