d'Almeida et al. [1991] have reviewed many available aerosol
measurements and from these have defined 22 aerosol components. These
components are used to build 12 aerosol types: clean Arctic; clean
Antarctic; polar-polluted; desert-background; desert-wind carrying
dust; clean continental (rural); clean-forest; average continental;
urban (industrial); clean-maritime; maritime-mineral; and
maritime-polluted. They further provide 5
5
spatial resolution of the distribution of these aerosols and monthly
computed radiative characteristics which include parameterized effects
of relative humidity. Their work is also useful as a summary of many
of the reported observations (prior to 1991) of aerosol
characteristics from around the globe.
The approach used by d'Almeida et al. [1991], that of cataloguing observations of ambient aerosol concentrations, is useful for modeling studies which require as input estimates of average aerosol quantities (optical depth, for example). An alternative approach is to characterize emissions, ideally of both primary aerosol and of secondary aerosol precursors. This approach is needed for modeling studies which investigate the response of ambient aerosol concentrations to variations in transport and removal (e.g., seasonal variations or variations caused by climate changes) and for climate feedback studies. Significant progress has been made in developing emissions estimates for important precursor gases, such as natural sources of sulfur ( Bates et al. [1992b]) and fossil fuel combustion sources of nitrogen and sulfur oxides ( Hameed and Dignon [1992]), which are then used to estimate secondary aerosol production.
Inventories of primary emissions of species such as dust and smoke are also needed. Penner et al. [1993b] constructed a global inventory of black carbon emissions using two methods: correlating ratios of black carbon to sulfur dioxide, and employing published emission factors and fuel production and use statistics. These estimates agreed to within a factor of two and yielded reasonable global concentrations when used in the Lawrence Livermore National Laboratory global chemistry / climate model.
The development of synoptic- to global-scale transport, transformation and deposition models for the cycling of aerosols and precursor gases, and comparison of the results with observations, will permit testing of current understanding of the interactions between aerosol sources, sinks, and transformation processes. Models which couple radiative transfer calculations are needed to explore direct climate effects of aerosols. Examination of the indirect effects of aerosols will require significant advancements in understanding and modeling of aerosol/CCN/cloud relationships Penner et al. [1993a].
Existing U.S. models include the Lawrence Livermore National Laboratory (LLNL) global model ( Galloway et al. [1992] and Taylor and Penner [1994]) which has been used to examine the global sulfur cycle and effects of sulfate aerosols. The Pacific Northwest Laboratory (PNL) model ( Luecken et al. [1991]) also treats sulfur species; the Brookhaven Global Chemistry / Aerosol Model ( Benkowitz et al. [1994]), derived from the PNL model, is driven by operational meterological data, allowing simulation of specific conditions for which observations are available, in contrast to other models which are driven by general circulation model output of representative meteorological conditions.