Convective Transport of Trace Gases and Lightning NOx Production in DC3 Storms Simulated with Cloud-Resolving, Regional, and Global Models

Deep convection is responsible for numerous effects on tropospheric chemistry, the most prominent being the transport of trace gases and aerosols from the boundary layer to the upper troposphere. Cloud-resolving models in many cases can provide an accurate explicit representation of convective transport. However, this process is parameterized in regional and global models. A second major effect of deep convection is the production of NOx by lightning (LNOx). Most models of all scales rely on parameterizations of LNOx production. In this talk we will focus on the ability of cloud-resolving, regional and global models to represent these two processes in storms for which observations of a variety of chemical and physical properties are available. The Deep Convective Clouds and Chemistry (DC3) field program took place over the Central US during May and June 2012. Airborne observations from the NASA DC-8, NSF/NCAR G-V and DLR Falcon aircraft characterized the low-level inflow and upper level outflow of numerous storms in three study regions: northeastern Colorado, west Texas to central Oklahoma, and northern Alabama. In each of these three regions dual-Doppler radar coverage and Lightning Mapping Array data were available. Cloud-resolved simulations of two or more DC3 storms using WRF-Chem at 3-km or finer horizontal resolution are evaluated in terms of hydrometeor distributions and vertical velocities using available radar data, and trace gas mixing ratios in the inflow and outflow regions are compared with the aircraft data. Vertical fluxes of CO and of hydrocarbons with lifetimes substantially longer than the convective time scale are computed from the model output. WRF-Chem was also run at 15-km horizontal resolution with parameterized convection (Grell 3-D scheme). The NASA Global Modeling Initiative chemical transport model is run at 2 x 2.5 degree resolution, driven by meteorology from the GEOS-5 MERRA assimilation. Convection is parameterized using the Relaxed Arakawa-Schubert scheme. Vertical fluxes of the same species in the two models with parameterized convection are compared with those computed in the cloud-resolved simulations. LNOx production is computed in all three models as a function of flash rates estimated from parameterization schemes. However, the storm parameters used in these schemes vary among the models. The model simulations all assume the same LNOx production per flash. The vertical distribution of the LNOx production is treated differently in the cloud-resolved WRF-Chem than in the other two simulations. Comparisons of the horizontal fluxes of LNOx out of the anvils of these storms are made among models and with data from aircraft anvil transects. The results of the model comparisons of vertical and horizontal fluxes of trace gases due to deep convection will be summarized, and implications for impacts of deep convection on upper tropospheric composition as predicted by regional and global models will be assessed.