In the case of primary aerosols, the largest uncertainties often lie in the
extrapolation of experimentally determined source strengths to other regions
and seasons. This is especially true for dust, for which many of the observations
are for a Saharan source. The spatial and temporal distribution of biomass fires
also remains uncertain. The non-linear dependence of sea salt aerosol formation
on wind speed creates difficulties in parametrizations in large-scale models
and the vertical profile of sea salt aerosols needs to be better defined.
Secondary aerosol species have uncertainties both in the sources of the precursor gases and in the atmospheric processes that convert some of those gases to aerosols. For sulphate, the uncertainties in the conversion from SO2 (factor of 2) are larger than the uncertainties in anthropogenic sources (20 to 30%). For hydrocarbons there are large uncertainties both in the emissions of key precursor gases as a function of space and time as well as the fractional yield of aerosols as those gases are oxidised. Taken at face value, the combined uncertainties can be a factor of three for sulphate and more for organics. On the other hand, the success some models have had in predicting aerosol concentrations at observation sites (see Section 5.4) as well as wet deposition suggests that at least for sulphate the models have more skill than suggested by a worst-case propagation of errors. Nevertheless, we cannot be sure that these models achieve reasonable success for the right reasons.
Besides the problem of predicting the mass of aerosol species produced, there is the more complex issue of adequately describing their physical properties relevant to climate forcing. Here we would like to highlight that the situation is much better for models of present day aerosols, which can rely on empirical data for optical properties, than for predictions of future aerosol effects. Another issue for optical properties is that the quantity and sometimes the quality of observational data on single scattering albedo do not match those available for optical depth. Perhaps the most important uncertainty in aerosol properties is the production of cloud condensation nuclei (Section 5.3.3).
Other reports in this collection