Modeling atmospheric sulfur over the Northern Hemisphere during the Aerosol Characterization Experiment 2 experimental period

A high‐resolution (1° × 1°, 27 vertical levels) Eulerian chemical transport and transformation model for sulfate, SO2, and related species driven by analyzed forecast meteorological data has been run for the Northern Hemisphere for June–July 1997 and extensively evaluated with observational data, mainly from air quality and precipitation chemistry networks. For ∼5000 evaluations, 50% of the modeled sulfate 24‐hour mixing ratios were within a factor of 1.85 of the observations; 50% of ∼328 concurrent subgrid observations were within a factor of 1.33. Much greater subgrid variation for 24‐hour SO2 mixing ratios (50% of ∼3552 observations were within a factor of 2.32) reflects high variability of this primary species; for ∼12600 evaluations, 50% of modeled mixing ratios were within a factor of 2.54 of the observations. These results indicate that a substantial fraction of the modeled and observed differences is due to subgrid variation and/or measurement error. Sulfate mixing ratios are identified by source type (biogenic, volcanic, and anthropogenic) and production mechanism (primary and by gas‐phase and aqueous‐phase oxidation). Examination of key diagnostics showed substantial variation for the different types of sulfur, e.g., SO2 aqueous‐phase oxidation rates of 29–102% d−1 and sulfate residence times of 4–9 days. Volcanic emissions contributed 10% of the sulfate burden and 6% of emissions, because the elevated release allows large fractional conversion of SO2 and long residence time. Biogenic SO2 was generally at lower concentrations than H2O2, resulting in efficient aqueous‐phase oxidation; this source type contributed 13% of emissions but only 5% of sulfate burden. Anthropogenic sources were the dominant contributors to sulfur emissions (80%) and sulfate burden (84%).