Influence of 2000-2050 climate change on particulate matter in the United States: Results from a new statistical model

Abstract

We use a statistical model to investigate the effect of 2000-2050 climate change on fine particulate matter (PM2.5) air quality across the contiguous United States. By applying observed relationships of PM2.5 and meteorology to the IPCC Coupled Model Intercomparision Project Phase 5 (CMIP5) archives, we bypass some of the uncertainties inherent in chemistry-climate models. Our approach uses both the relationships between PM2.5 and local meteorology as well as the synoptic circulation patterns, defined as the singular value decomposition (SVD) pattern of the spatial correlations between PM2.5 and meteorological variables in the surrounding region. Using an ensemble of 19 global climate models (GCMs) under the RCP4.5 scenario, we project an increase of 0.4-1.4μgm-3 in annual mean PM2.5 in the eastern US and a decrease of 0.3-1.2μgm-3 in the Intermountain West by the 2050s, assuming present-day anthropogenic sources of PM2.5. Mean summertime PM2.5 increases as much as 2-3μgm-3 in the eastern United States due to faster oxidation rates and greater mass of organic aerosols from biogenic emissions. Mean wintertime PM2.5 decreases by 0.3-3μgm-3 over most regions in the United States, likely due to the volatilization of ammonium nitrate. Our approach provides an efficient method to calculate the potential climate penalty on air quality across a range of models and scenarios. We find that current atmospheric chemistry models may underestimate or even fail to capture the strongly positive sensitivity of monthly mean PM2.5 to temperature in the eastern United States in summer, and they may underestimate future changes in PM2.5 in a warmer climate. In GEOS-Chem, the underestimate in monthly mean PM2.5-temperature relationship in the east in summer is likely caused by overly strong negative sensitivity of monthly mean low cloud fraction to temperature in the assimilated meteorology (∼-0.04-1) compared to the weak sensitivity implied by satellite observations (±0.01K-1). The strong negative dependence of low cloud cover on temperature in turn causes the modeled rates of sulfate aqueous oxidation to diminish too rapidly as temperatures rise, leading to the underestimate of sulfate-temperature slopes, especially in the south. Our work underscores the importance of evaluating the sensitivity of PM2.5 to its key controlling meteorological variables in climate-chemistry models on multiple timescales before they are applied to project future air quality.