Source attribution of PM2.5 and O3 concentrations and health outcomes from 2010 and 2011 in Washington D.C. using sensitivity analyses in the GEOS-Chem adjoint model

In order to develop informed policies for pollution mitigation, knowledge of the origins of pollutants is necessary. In this study, we make use of the GEOS-Chem adjoint model, a tool for performing sensitivity analyses that determines the regions, emissions and sectors that make the largest contributions to PM_2.5 and O₃ in a specific location. In this case we choose Washington D.C. as a pilot city, planning to perform similar analyses for Paris, Santiago, Accra and Jakarta. In addition to source attribution of PM_2.5 and O₃ concentrations, we investigate the effect of pollution on health using exposure metrics consistent with those used in epidemiological studies. To more accurately capture PM_2.5 exposure, we scale modeled concentrations at 0.5x0.666 degree resolution with satellite data at 0.1x0.1 degrees. Through these additional steps, we update our "cost function" - a modeled quantity of interest - to represent population weighted pollution and associated death in the D.C. metropolitan area. Running the adjoint model over the US nested domain, we determine the sensitivity of PM_2.5 exposure and related-deaths in Washington D.C. to emissions from the following species: NH₃, NO_x, Black Carbon, Organic Carbon and SO₂ and sectors: Aviation, Agriculture, Energy, Industry, Residential, Shipping, and Transportation. Additionally, we determine the sensitivity of O₃ exposure and related-deaths to NO_x and VOCs. We perform a spatial analysis to determine what regions contribute the greatest to pollution in D.C. and compare our results with O₃ and PM_2.5 source apportionment studies at higher spatial resolution in the D.C. area. We have found that approximately two thirds of PM_2.5 concentration and related-death (520 deaths) in D.C. originates from emissions external to the region, with a large component originating from agricultural NH₃ and industrial SO₂. These results can be used to inform decision making for D.C. and the surrounding region.