Enhancing Estimation of Fine Particulate Matter Chemical Composition across North America by Including Geophysical A...

Exposure to fine particulate matter (PM_2.5) in ambient air is recognized as the leading environmental risk factor for mortality. A more comprehensive characterization of its chemical composition is needed for its management and health effects research. We improve estimates of total PM_2.5 mass concentration and its chemical composition across North America by developing, optimizing, and applying convolutional neural networks (CNN) with information from satellite-, simulation-, and monitor-based sources to estimate the local bias in monthly geophysical a priori PM_2.5 and component concentrations over 2000-2023. Significant long-term agreement is found with traditional 10-fold spatial cross-validation for total PM_2.5 (R ² = 0.82), sulfate (R ² = 0.98), nitrate (R ² = 0.93), ammonium (R ² = 0.94), organic matter (R ² = 0.83), black carbon (R ² = 0.78), dust (R ² = 0.71), and seasalt (R ² = 0.37). We introduce Buffered Leave Isolated Sites and Clusters Out (BLISCO) spatial cross-validation to evaluate the model extrapolation ability over remote regions, and find that traditional spatial cross-validation may overestimate performance and underrepresent uncertainty due to the spatial autocorrelation of ground monitors. The use of geophysical information from a chemical transport model (GEOS-Chem) significantly increases CNN performance in BLISCO cross-validation, for example, increasing R ² for NO₃ ^- (0.51 to 0.81) and NH₄ ⁺ (0.27 to 0.67). We represent spatial uncertainty for PM_2.5 and its components based on the statistical results of BLISCO cross-validation by integrating information from both the spatial distribution of ground observations and the variability in predictors space representation, and find that distance from monitor is a key predictor of uncertainty.