Global Burden of Disease-Major Air Pollution Sources – a GLOBAL Approach
Michael Brauer1, Randall V. Martin2,3, Erin E. McDuffie2,3 (Presenter), Joseph Spadaro4, Richard T. Burnett5
1The University of British Columbia, Vancouver, BC, Canada; 2Institute for Health Metrics and Evaluation, University of Washington, Seattle, WA, USA; 3Washington University in St. Louis, MO, USA; Dalhousie University, Halifax, NS, Canada; 4Spadaro Environmental Research Consultants (SERC), Philadelphia, PA, USA; 5Health Canada, Ottawa, ON, Canada
Background: Identification of contributing sources is an important next step in research to address air pollution as a global health risk factor. Source identification can be useful to initiate air quality management and to evaluate different policy options. Recent work conducted in China and India used updated national emissions inventories and chemical transport model simulations to identify current and future contributions to exposure and disease burden from multiple source sectors. There is now interest in conducting similar analyses for all countries included in the Global Burden of Disease (GBD). This project will estimate source sector contributions to disease burden from ambient PM2.5 in 2017 at the national level for all ~195 countries and territories included in the GBD.
Methods: We use a new 2017 global emissions dataset from the Community Emissions Data System (CEDS), together with inventories for biomass burning (GFED4) and other sources (e.g., windblown and anthropogenic dust) in the GEOS-Chem 3D chemical transport model to examine the sensitivity of PM2.5 mass to 16 source sector categories (including Residential and Commercial Energy, Industry, Power Generation, Agriculture, On-Road and Off-Road Transportation, Waste, Solvents, International Shipping, Open Fires, Agricultural Waste Burning, Anthropogenic and Windblown Dust, Biogenic sources). Additional fuel-specific simulations (including residential coal and biofuel combustion, as well as industrial and power generation coal combustion) provide additional information with increased policy relevance. We bias-correct and downscale simulation results using the 0.1° x 0.1° GBD satellite-based exposure estimates to directly compare with the GBD and further downscale these results to the 0.01° x 0.01° resolution to facilitate analysis at an urban area scale. For disease burden estimates, downscaled population-weighted PM2.5 mass concentrations from each source sector are combined with exposure response functions to estimate the population attributable fraction for each disease (ischemic heart disease, stroke, lung cancer, COPD, acute lower respiratory infections, type 2 diabetes) included in the GBD. We also conduct a sensitivity analysis using the newly developed Global Exposure Mortality Model (GEMM), which estimates non-accidental mortality and cause-specific mortality in functions derived only from studies of ambient air pollution health impacts.
Results: Simulations for 2017 indicate that industry, power generation, and residential emissions are dominant global contributors to population-weighted PM2.5 mass (~15%, 14%, 18% contributions, respectively), but that their relative contributions vary by country and region. Fuel-specific simulations also indicate that total coal and biofuel combustion individually contribute to about 10% of global PM2.5 mass, but that each have contributions of up to 30% in heavily polluted countries such as India and China. When combined with the exposure response, downscaled fractional source contributions for one of the most polluted countries, India, suggest that 60% of the ambient pollution deaths in 2017 were attributable to regional power generation, industry, and residential combustion emissions.
Conclusions: Chemical transport models, when constrained with satellite and ground-based observations, offer valuable information about sources contributing to PM2.5 pollution and their health impacts at the global scale.