Impact of Model Spatial Resolution on Global Geophysical Satellite-Derived Fine Particulate Matter

Abstract

Global geophysical satellite-derived ambient fine particulate matter (PM_2.5) inference relies upon a geophysical relationship (η) from a chemical transport model to relate satellite retrievals of aerosol optical depth (AOD) to surface PM_2.5. The resolution dependence of simulated η warrants further investigation. In this study, we calculate geophysical PM_2.5 with simulated η from the GEOS-Chem model in its high-performance configuration (GCHP) at cubed-sphere resolutions of C360 (∼25 km) and C48 (∼200 km) and satellite AOD at 0.01° (∼1 km). Annual geophysical PM_2.5 concentrations inferred from satellite AOD and GCHP simulations at ∼25 km and ∼200 km resolutions exhibit remarkable similarity (R² = 0.96, slope = 1.03). This similarity in part reflects opposite resolution responses across components with population-weighted normalized mean difference (PW-NMD) increasing by 5% to 11% for primary species while decreasing by −30% to −5% for secondary species at finer resolution. Despite global similarity, our results also identify larger resolution sensitivities of η over isolated pollution sources and mountainous regions, where spatial contrast of aerosol concentration and composition is better represented at fine resolution. Our results highlight the resolution dependence of representing near-surface concentrations and the vertical distribution of chemically different species with implications for inferring ground-level PM_2.5 from columnar AOD.