Investigation of the Relationship between LNOx Production Efficiency and Flash Rate through Cloud-Resolved Model Simulations of Observed Storms

Satellite observations of NO₂ columns over thunderstorms have proven to be a valuable method for sampling the NO_x production by lightning in a large number of storms over broad areas of the Earth. However, there are several significant uncertainties associated with satellite-based estimates of lightning NO_x (LNO_x) production efficiency (PE) per flash. In addition, analyses of OMI satellite-based PE estimates have suggested that PE is inversely related to flash rate. We investigate this relationship through analyses of two storms observed over Oklahoma during the Deep Convective Cloud and Chemistry (DC3) field campaign. Aircraft observations of NO_x in storm anvils are helpful in providing a basis to compare with satellite estimates. However, typically aircraft sampling is conducted for only short periods of time and only in a small fraction of the cloud volume. Therefore, we have simulated the observed storms (May 29, 2012 and May 19, 2012) and associated chemistry to provide a more comprehensive view of LNO_x in the storms using the WRF-Chem model. The May 29 system was a high flash rate event, whereas the May 19 storm had a much lower flash rate. Total lightning flashes are parameterized in the model to best match the flash rate magnitude and time series evolution observed by the Oklahoma Lightning Mapping Array (LMA). For each system a series of simulations were conducted with varying mean LNO_x PE to determine the best match with the aircraft NO_x observations. We contrast the resulting mean PE values for the two storms in relation to the observed flash rates. In addition, we use the model simulation output to investigate uncertainties in the OMI LNOx retrieval scheme, such as the magnitude of LNO_x injection above the tropopause and the background NO_x due to convective transport from the PBL. The combination of aircraft, model, and satellite data furthers our skill in estimating LNO_x production.