Analysis of regional budgets of sulfur species modeled for the COSAM exercise

The COSAM intercomparison exercise (comparison of large‐scale sulfur models) was organized to compare and evaluate the performance of global sulfur cycle models. Eleven models participated, and from these models the simulated surface concentrations, vertical profiles and budget terms were submitted. This study focuses on simulated budget terms for the sources and sinks of SO₂ and sulfate in three polluted regions in the Northern Hemisphere, i.e., eastern North America, Europe, and Southeast Asia. Qualitatively, features of the sulfur cycle are modeled quite consistently between models, such as the relative importance of dry deposition as a removal mechanism for SO₂, the importance of aqueous phase oxidation over gas phase oxidation for SO₂, and the importance of wet over dry deposition for removal of sulfate aerosol. Quantitatively, however, models may show large differences, especially for cloud‐related processes, i.e., aqueous phase oxidation of SO₂ and sulfate wet deposition. In some cases a specific behavior can be related to the treatment of oxidants for aqueous phase SO₂ oxidation, or the vertical resolution applied in models. Generally, however, the differences between models appear to be related to simulated cloud (micro‐)physics and distributions, whereas differences in vertical transport efficiencies related to convection play an additional rôle. The estimated sulfur column burdens, lifetimes and export budgets vary between models by about a factor of 2 or 3. It can be expected that uncertainties in related effects which are derived from global sulfur model calculations, such as direct and indirect climate forcing estimates by sulfate aerosol, are at least of similar magnitude.     

A comparison of scavenging and deposition processes in global models: results from the WCRP Cambridge Workshop of 1995

We report on results from a World Climate Research Program workshop on representations of scavenging and deposition processes in global transport models of the atmosphere. 15 models were evaluated by comparing simulations of radon, lead, sulfur dioxide, and sulfate against each other, and against observations of these constituents. This paper provides a survey on the simulation differences between models. It identifies circumstances where models are consistent with observations or with each other, and where they differ from observations or with each other. The comparison shows that most models are able to simulate seasonal species concentrations near the surface over continental sites to within a factor of 2 over many regions of the globe. Models tend to agree more closely over source (continental) regions than for remote (polar and oceanic) regions. Model simulations differ most strongly in the upper troposphere for species undergoing wet scavenging processes. There are not a sufficient number of observations to characterize the climatology (long‐term average) of species undergoing wet scavenging in the upper troposphere. This highlights the need for either a different strategy for model evaluation (e.g., comparisons on an event by event basis) or many more observations of a few carefully chosen constituents.     

Vertical distributions of sulfur species simulated by large scale atmospheric models in COSAM: Comparison with observations

A comparison of large‐scale models simulating atmospheric sulfate aerosols (COSAM) was conducted to increase our understanding of global distributions of sulfate aerosols and precursors. Earlier model comparisons focused on wet deposition measurements and sulfate aerosol concentrations in source regions at the surface. They found that different models simulated the observed sulfate surface concentrations mostly within a factor of two, but that the simulated column burdens and vertical profiles were very different amongst different models. In the COSAM exercise, one aspect is the comparison of sulfate aerosol and precursor gases above the surface. Vertical profiles of SO₂, SO²⁻₄, oxidants and cloud properties were measured by aircraft during the North Atlantic Regional Experiment (NARE) experiment in August/September 1993 off the coast of Nova Scotia and during the Second Eulerian Model Evaluation Field Study (EMEFSII), in central Ontario in March/April 1990. While no single model stands out as being best or worst, the general tendency is that those models simulating the full oxidant chemistry tend to agree best with observations although differences in transport and treatment of clouds are important as well.