Development of a three-dimensional cloud-scale chemical transport model

We have developed a preliminary version of an off-line 3-D cloud-scale chemical transport model (CTM) which is driven by output from the 3-D Goddard Cumulus Ensemble (GCE) Model. This is the first step toward our ultimate goal of running the chemistry and transport on-line within the cloud model. The CTM can be used to study transport within convective storms, the production of NOx by lightning, scavenging of soluble species, and ozone photochemistry. We have applied the model to a convective storm that was observed over SE Wyoming and NE Colorado during the Stratosphere-Troposphere Experiment - Radiation, Aerosols, and Ozone (STERAO-A) in July, 1996.

Three-dimensional fields of winds, temperatures, water vapor, diffusion coefficients, and five classes of hydrometeors from the GCE model are saved every 10 minutes and input to the cloud-scale CTM. These fields are interpolated to 15-s intervals for use in transport calculations, which are performed using the Van Leer advection scheme. The SMVGEAR-II chemistry solver is used along with a simplified version of the Jacobson chemical mechanism. A cloud-perturbed photolysis scheme has been devised for thin, thick, and very thick clouds. Soluble species are removed from the gas phase based on the Henry's Law constants of the individual species and the liquid water content of the cloud. Lightning NOx production is parameterized using the scheme of DeCaria et al. [2000, JGR] in which observed cloud-to-ground (CG) and intracloud (IC) flash rates are used along with methods of specifying the location of flashes within the cloud.

The 3-D GCE model simulation of the 12 July 1996 STERAO-A storm produced a multiple thunderstorm cell structure very similar to that observed on radar. We have applied the cloud-scale CTM to this storm with the objective of determining the amount of NOx production per flash and the effect of lightning NOx on photochemical ozone production in the cloud and downstream. The chemistry was initialized with observed data from an aircraft profile ahead of the storm. The results show that NOx production per flash was roughly similar for CG and IC flashes. However, due to the factor of seven greater number of IC flashes, the IC flashes dominated the NOx production within the cloud. The lightning NOx had only a small effect on ozone production during the storm, increasing ozone by only 1-3 ppbv. However, after storm dissipation, the lightning NOx resulted in a net ozone production of up to 12 ppbv/day in the upper troposphere over the subsequent 24 hours.