Effect of the addition of an ultrafine (4-km) grid on ozone formation predicted by UAM-V

An interest in the effects of model grid resolution and vertical diffusivity arose during the OTAG modeling when ozone simulation results suggested a sensitivity to grid resolution, particularly in urban areas, as well as boundary layer height. Both factors can affect the rate with which emissions from different source types become mixed. Vertical diffusivity was especially problematic because the meteorological modeling produced values that in some cases seemed too small. In addition, early modeling results produced daytime boundary layer heights that were lower than observed. Thus, changes were made in the modeling that produced somewhat larger diffusivities and deeper mixed layers, and influenced subsequent computed ozone concentrations.

An outgrowth from this experience was a series of experimental UAM-V simulations conducted for the Tennessee Valley region using two grid configurations. The first configuration, (G12), contained a 12 km grid nested inside a 36 km grid, the configuration which had been used during the OTAG ozone modeling efforts. The second configuration, (G04), contained an additional 4 km grid of dimensions 692 X 442 km, which was nested inside the 12 km grid. The vertical resolution of the 4-km grid was 7 layers, the same as the vertical resolution of the 12-km grid. RAMS was used to prepare wind, height/pressure and vertical diffusivity files at 4 km resolution. An alternate vertical diffusivity file at 4-km resolution was prepared by re-gridding the OTAG 12-km vertical diffusivity file. The vertical diffusivities produced in this way were about one order of magnitude greater than the 4-km vertical diffusivities obtained from RAMS. An emissions inventory at 4-km resolution was obtained and processed to produce the appropriate area source emissions files. A land use file for the 4-km grid was obtained by appropriate processing of ROM data. Results from a number of cases were compared to determine the effects of specific changes in configuration.

Comparisons were conducted with respect to rural areas, large urban areas, and areas downwind of large NO_x point sources which were located within the boundaries of the 4-km grid. The effects on ozone formation were analyzed to determine the influence of (1) the presence of the 4-km grid; (2) the use of plume-in-grid (PIG) treatment at 4-km resolution, and (3) the magnitude of vertical diffusivity values on the 4-km grid. When the 4-km grid was added to the grid configuration, mid-afternoon ozone concentrations observed in rural areas were typically reduced by about 20 ppb. The use of PIG did not have a significant impact on ozone production. Use of the smaller diffusivities produced higher ozone concentrations at Memphis and Nashville. This result may be due to the higher isoprene concentrations which were observed in the lowest layers as a consequence of the reduced vertical mixing.