Drivers of Interannual to Multidecadal Variability in U.S. Air Quality and Implications for Air Quality Forecasts on Seasonal to Decadal Timescales

Abstract

This dissertation investigates the meteorological drivers of interannual to multidecadal variability in U.S. air quality. Toward that end, we quantify the meteorological processes influencing ozone and fine particulate matter (PM2.5) at different spatial scales, including local weather (~ 100 km), synoptic circulation (~1,000 km) and large-scale climate patterns (~10, 000 km). As expected, local temperature drives ozone episodes in summer across most of the United States; we also identify a marked decrease in the ozone-temperature relationship at high temperatures, a phenomenon known as ozone suppression, at ~20% of U.S. observation sites. We also find a strong relationship between monthly mean PM2.5 and local temperatures across the East in summer. On synoptic scales, we find that the polar jet wind and the Bermuda High control much of the ozone variability in the eastern United States. The variability in June-July-August (JJA) ozone in the East is also closely linked to large-scale ocean-atmosphere interactions involving patterns of sea level pressure (SLP) and Atlantic and Pacific sea surface temperatures (SSTs). Using these large-scale climate patterns, we build a linear regression model that can predict ~45% of the variability in JJA maximum daily 8-hour average (MDA8) ozone one season in advance. Our research also reveals that the Atlantic Multidecadal Oscillation (AMO) drives much of the variability of JJA air quality in the eastern United States on long timescales. The shift of cold AMO to warm AMO can lead to warmer and drier summers in the eastern United States, which in turn may influence regional air quality. We find that in one-half AMO cycle (~35 years), JJA MDA8 ozone concentrations change by 1-3 ppbv in the Northeast and 2-5 ppbv in the Great Plains; JJA PM2.5 concentrations change by 0.6-1.0 μg m-3 in the Northeast and Southeast. This result suggests that a complete picture of air quality management in coming decades would benefit from consideration of the AMO influence. By applying present-day statistical relationships of air pollutants and meteorology to future climate projections, we can estimate future changes in U.S. air quality. Here we use an ensemble of projections following the Representative Concentration Pathway (RCP) 4.5 scenario archived for the Coupled Model Intercomparison Project (CMIP5). Our approach assumes constant present-day emissions of ozone precursors and PM2.5 sources. For annual mean PM2.5, we project an increase of ~1 μg m-3 in the eastern United States and a decrease of 0.3-1.2 μg m-3 in the Intermountain West. In summer, PM2.5 increases as much as 2-3 μg m-3 across a large swath of the East. For ozone, we find an average increase of 2.4 episode days a-1 across the United States; the change is as large as 3-9 days a-1 in many sites in the Northeast, Midwest, and California. Our results imply that the “climate penalty” on air quality could undercut ongoing regulatory efforts to control pollution.