Estimating Model-Based Marginal Societal Health Benefits of Air Pollution Emission Reductions in the United States and Canada

A Hakami,S Zhao,M Soltanzadeh,P Vasilakos,A Alhusban,B Oztaner,N Fann,H Chang,A Krupnick,T Russell

Abstract:We developed spatially detailed source-impact estimates of population health burden measures of air pollution for the United States and Canada by quantifying sources-receptor relationships using the benefit-per-ton (BPT1) metric. We calculated BPTs as the valuations of premature mortality counts due to fine particulate matter (PM2.5; particulate matter ≤2.5 μm in aerodynamic diameter) exposure resulting from emissions of one ton of a given pollutant. Our BPT estimates, while accounting for a large portion of societal impact, do not include morbidity, acute exposure mortality, or chronic exposure mortality due to exposure to other pollutants such as ozone. The adjoint version of a widely used chemical transport model (CTM) allowed us to calculate location-specific BPTs at a high level of granularity for source-impact characterization. Location-specific BPTs provides a means for exploiting the disparities in source impact of emissions at different locations. For instance, estimated BPTs show that 20% of primary PM2.5 and ammonia emissions in the United States account for approximately 50% and 60% of the burden of each species, respectively, for an estimated burden of $370B USD. Similarly, 10% of the most harmful emissions of primary PM2.5 and ammonia emissions in Canada account for approximately 60% and 50% of their burden, respectively. By delineating differences and disparities in source impacts, adjoint-based BPT provides a direct means for prioritizing and targeting emissions that are most damaging. Sensitivity analyses evaluated the impact of our assumptions and study design on the estimated BPTs. The choice of concentration-response function had a substantial impact on the estimated BPTs and is likely to constitute the largest source of uncertainty in those estimates. Our method for constructing annual BPT estimates based on episodic simulations introduces low uncertainty, while uncertainties associated with the spatial resolution of the CTM were evaluated to be of medium importance. Finally, while recognizing that the use of BPTs entails an implied assumption of linearity, we show that BPTs for primary PM2.5 emissions are stable across different emission levels in North America. While BPTs for precursors of secondary inorganic aerosols showed sensitivity to emission levels in the past, we found that those have stabilized with lower emissions and pollutant concentrations in the North American atmosphere. We used BPTs to provide location-specific and sectoral estimates for the cobenefits of reducing carbon dioxide emissions from a range of combustion sources. Cobenefit estimates rely heavily on the emission characteristics of the sector and therefore exhibit more pronounced sectoral fingerprints than do BPTs. We provide cobenefit estimates for various subsectors of on-road transportation, thermal electricity generation, and off-road engines. Off-road engines and various heavy-duty diesel vehicles had the largest cobenefits, which in most urban locations far exceeded estimates of the social cost of carbon. Based on our cobenefit estimations, we also provide per-vehicle burden estimates for different vintages of vehicle subsectors such as transit buses and short-haul trucks in major US cities.