Development and application of a new air pollution modeling system .3. Aerosol-phase simulations

Result from a new air pollution model were tested against data from the Southern California Air Quality Study (SCAQS) period of 26-29 August 1987. Gross errors for sulfate, sodium, light absorption, temperatures, surface solar radiation, sulfur dioxide gas, formaldehyde gas, and ozone were lowest among parameters compared (1-40%). Gross errors for elemental carbon, organic carbon, total particulate mass, ammonium, ammonia gas, nitric acid gas, and light scattering, were larger (40-61%). Gross errors for particulate nitrate were largest (65-70%). Reducing the baseline land-based particulate emissions inventory to one-third its original value did not affect gross errors significantly; however, it did turn overpredictions into underpredictions for many species. Doubling emissions increased gross errors for nearly all parameters. Setting lateral boundary inflow concentrations of particles to zero caused slight (<1%) erosion of results for most species, large erosion (10%) for sodium and chloride, but slight improvement (<1%) for a few species. Setting both lateral inflow and initial concentrations for gases and particles to zero caused severe degradation of results for many species but relatively mild degradation or improvement for a few. Spinning up the meterological model 24 h in advance caused most gross errors to increase. Finally, the presence of aerosols reduced peak daytime surface solar radiation by approximately 6.4% (55 W m(-2)), increased nighttime temperatures by about 0.77 K, decreased daytime temperatures by about 0.08 K, and increased overall temperatures (day plus night) by 0.43 K compared to a no-aerosol case. The presence of aerosols also caused ozone mixing ratios to decrease by 2%. Copyright (C) 1996 Elsevier Science Ltd