A global-through-urban scale nested and coupled air pollution and weather forecast model was developed and applied to the SARMAP Field Campaign. The model, GATOR-GCMM (Gas, Aerosol, Transport, Radiation, General Circulation, and Mesoscale Meteorological) model, treats nesting of all important air quality (including size- and composition-resolved aerosol) and meteorological parameters simultaneously, from the global scale through the urban scale (<5-km grid spacing). It treats any number of nested layers and any number of meteorological and air quality grids in each layer between the global and urban scales. The model does not use nudging, or data assimilation; hence, it is entirely prognostic, except for the emissions inventory. It is configured so that, regardless of the number of grids used during a single continuous simulation, the total central memory requirement of the model never exceeds 1.5 times and 2.1 times the central memory requirement of the largest grid for gas simulations and gas/aerosol simulations, respectively. The model includes feedbacks of air pollutants to meteorology. New aerosol numerical algorithms were developed for the model. New algorithms accounting for subgrid-scale soil classification and vegetation cover heterogeneity in ground-temperature calculations were also developed. The ground-temperature module treats temperatures over sea ice, urban areas, and snow-covered soil, vegetation, sea-ice, and urban areas. Urban areas are broken down into road surfaces, rooftops, vegetation, and bare soil. The model was applied with five global-through-urban scale nested grids to the August 3-6, 1990 SARMAP field campaign. Parameters compared with observations included air temperatures, air pressures, relative humidities, wind speeds/directions, and the following 20 gases and carbon bond groups: O3, NO, NO2, CO, SO2, nonmethane organic carbon, CH4, C2H6, C3H8, paraffins, ethene, olefins, formaldehyde, higher aldehydes, acetone, other ketones, toluene, xylene, benzaldehdye, and isoprene. In the absence of nesting, the gross error in near-surface temperatures, normalized over all 8617 day and night observations in the model domain, was 1.06% of observed kelvin temperatures. The gross error in near-surface relative humidity (RH), normalized over all 5597 day and night observations in the domain was less than 24.8% of observed RH values. The gross error in near-surface ozone mixing ratio, normalized over all 2866 measurements above 50 ppbv during the 4-day period, was 23.5%. Nesting improved model statistics for all near-surface parameters, but not significantly, probably due to the fact that the air pollution episode occurred under a synoptic high pressure system. Finer-resolution grid spacing resulted in higher average ozone mixing ratios than coarser-resolution spacing, except in the largest regional grid; hence, the coarser the grid, the greater the underprediction of ozone in most cases.

Material from IMA Talk

Back to Workshop Schedule

Back to Atmospheric Modeling

Back to top of page