Parameterizations for lightning NOx in atmospheric models

Representation of the lightning source of NOx in atmospheric chemical transport models is a major source of uncertainty in computations of the upper tropospheric NOx and NOy budgets and upper tropospheric ozone production. Uncertainties exist in the methods of specifiying lightning flash rates, the relative proportions of intracloud (IC) and cloud-to-ground (CG) flashes, the NO production rates per CG and IC flash, and the specification of the effective vertical distribution of lightning NOx production. Better understanding of lightning NOx production on the cloud scale is key for developing better parameterizations for regional and global chemical transport models. Therefore, we have been approaching the problem using a cloud-resolving model as well as mesoscale and global models.

We describe our parameterization of lightning NOx production in the Goddard Cumulus Ensemble Model and the results obtained in relation to airborne measurements taken in and around storm anvils during the 1996 STERAO-A field program in Colorado. The results of this investigation suggest that in the storms considered NOx production per IC flash substantially exceeded what would be expected using commonly accepted values of production per IC flash relative to that per CG flash.

Cloud top height has commonly been used as a predictor of total flash rate in mesoscale and global chemical transport models. We have tested such algorithms using observed flash rate data and find that despite reasonable prediction of cloud tops by the model, the cloud top height was not a good predictor of flash rate. A wide range of observed flash rates were present for a given cloud top altitude. Upward cloud mass fluxes from the convective parameterizations yielded better predictions of flash rates. Results of such a new flash rate algorithm have been compared with NOx and NOy data from the upper troposphere observed during the 1997 SONEX field campaign over North America and the North Atlantic.