Abstract
We use the stretched‐grid version of the three‐dimensional global University of Maryland Chemical Transport Model (UMD‐CTM) to examine the effects of mesoscale meteorological features such as fronts and deep convection on regional‐scale chemistry and transport. The stretched‐grid model simulation, with a grid configuration featuring a mesoscale resolution region centered over the central United States, was conducted for June 1985, and evaluated through comparisons with a set of aircraft observations of trace gases. We also present results from a uniform‐grid UMD‐CTM simulation with a more conventional 2° × 2.5° horizontal resolution for the same time period to examine how well the stretched‐grid global model simulates mesoscale features. The changes in middle and upper tropospheric CO and O3 due to convection from the model simulations are in good agreement with the range of measurements. The stretched‐grid model shows better agreement with measurements than the uniform‐grid model for the enhancement of trace gases in upper troposphere outflow due to deep convection and for the gradient of trace gas mixing ratios across a cold front. Peak convective enhancement of CO in the upper troposphere is larger in the stretched‐grid model simulation than in the uniform‐grid simulation, indicating a better representation of locally focused deep convective transport of polluted boundary layer air in the former. This type of vertical transport feature must be handled accurately if a model is to be used for intercontinental transport calculations. However, we find that deep convection in both model simulations, although better simulated in the stretched‐grid model, is too widespread and too frequent. We find that net ozone production in the polluted boundary layer is ∼15% less in the fine‐grid region (0.5° resolution) of the stretched‐grid model than in the same region of the 2° × 2.5° model due to less artificial dilution of ozone precursors. The net ozone production in convective outflow plumes is also smaller in the stretched‐grid model than in the uniform‐grid model. We estimate the net flux of ozone from North America in the lowest 7 km to be 10 Gmol d−1 for the month of June using the results from the stretched‐grid simulation. This value includes direct horizontal boundary layer flux, ozone that has been vertically transported from the boundary layer to free troposphere, and ozone that had been produced photochemically in the free troposphere.