Influence of model spatial resolution on simulated aerosol surface concentration

Abstract

Fine particulate matter known as PM2.5 exists in Earth’s atmosphere at varying levels globally. High ambient concentrations of PM2.5 are associated with adverse health impacts, reduced visibility, and have a relatively poorly understood effect on global climate. Global chemical transport models provide an opportunity to simulate PM2.5 with full spatial coverage on a global to regional scale. Satellite observations can be incorporated with simulated PM2.5 to further strengthen PM2.5 estimates. This work explores the differences in simulated PM2.5 using fine (0.25◦ x 0.3125◦) and coarse (2◦ x 2.5◦) model resolution, with the aim of improving PM2.5 estimation and monitoring. Simulating surface concentration of PM2.5 using fine spatial resolution improves agreement with ground-based measurements compared with a coarse resolution simulation, with explained variance increasing by as much as 0.16 seasonally. The fine resolution simulation better resolves spatial gradients in surface PM2.5 which are poorly captured at coarse resolution, such as regions of biomass burning. In urban areas, where population is most dense and accurate health impact assessments are crucial, the fine resolution simulation reveals enhanced surface PM2.5 at the sub-grid scale around city centres. Combining simulated PM2.5/AOD with satellite-derived observations yields further improvements in estimated surface PM2.5. The fine resolution satellite-model PM2.5 estimates show the strongest agreement with ground-based measurements, with correlation coe cients >0.53 and near 1:1 relationship across all seasons. Differences between estimates of PM2.5 and its constituent species at varying model resolutions result from differences in emission density, i.e. the dilution of high density emission sources over coarse grid boxes. Recommendations for future simulations are made based on fine resolution sensitivity tests with varying chemical mechanisms and emission inputs.