Atmos. Chem. Phys., 18, 8183–8202, 2018 https://doi.org/10.5194/acp-18-8183-2018 © Author(s) 2018. This work is distributed under the Creative Commons Attribution 4.0 License. Ozone response to emission reductions in the southeastern United States Charles L. Blanchard1 and George M. Hidy2 1Envair, Albany, CA 94706, USA 2Envair/Aerochem, Placitas, NM 87043, USA Correspondence: Charles L. Blanchard ([email protected]) Received: 7 June 2017 – Discussion started: 8 August 2017 Revised: 24 February 2018 – Accepted: 19 May 2018 – Published: 13 June 2018 Abstract. Ozone (O3) formation in the southeastern US is ambient NOy species concentrations offers key insights into studied in relation to nitrogen oxide (NOx) emissions using processes of general relevance to air quality management and long-term (1990s–2015) surface measurements of the South- provides important information supporting strategies for re- eastern Aerosol Research and Characterization (SEARCH) ducing O3 mixing ratios. network, U.S. Environmental Protection Agency (EPA) O3 measurements, and EPA Clean Air Status and Trends Net- work (CASTNET) nitrate deposition data. Annual fourth- highest daily peak 8 h O3 mixing ratios at EPA monitoring 1 Introduction sites in Georgia, Alabama, and Mississippi exhibit statis- tically significant (p < 0.0001) linear correlations with an- Ozone (O3) is a well known and important product of pho- nual NOx emissions in those states between 1996 and 2015. tochemical processes in the troposphere involving nitric ox- The annual fourth-highest daily peak 8 h O3 mixing ra- ide (NO), nitrogen dioxide (NO2), and volatile organic com- tios declined toward values of ∼ 45–50 ppbv and monthly pounds (VOCs). Ozone is of broad interest for its adverse −1 O3 maxima decreased at rates averaging ∼ 1–1.5 ppbv yr . effects on humans and ecosystems, as reflected by regulation Mean annual total oxidized nitrogen (NOy) mixing ratios at through the US Clean Air Act (e.g., U.S. EPA, 2014, 2015a). SEARCH sites declined in proportion to NOx emission re- Regulatory actions address extreme O3 mixing ratios: the ductions. CASTNET data show declining wet and dry nitrate US National Ambient Air Quality Standard (NAAQS), cur- deposition since the late 1990s, with total (wet plus dry) ni- rently 70 ppbv, is applicable to the annual fourth-highest trate deposition fluxes decreasing linearly in proportion to re- daily 8 h maxima averaged over 3-year periods (U.S. EPA, ductions of NOx emissions by ∼ 60 % in Alabama and Geor- 2015b, c). By the early 1990s, US emission control efforts gia. Annual nitrate deposition rates at Georgia and Alabama began to focus on nitrogen oxides (NOx D NO C NO2) in ad- CASTNET sites correspond to 30 % of Georgia emission dition to VOCs (NRC, 1991). O3 management has generally rates and 36 % of Alabama emission rates, respectively. The relied on precursor reduction requirements estimated from fraction of NOx emissions lost to deposition has not changed. models that integrate descriptions of nonlinear chemical and SEARCH and CASTNET sites exhibit downward trends in atmospheric processes (e.g., Seigneur and Dennis, 2011), mean annual nitric acid (HNO3) concentrations. Observed and guidance has also derived from so-called “observation- relationships of O3 to NOz (NOy–NOx) support past model based” models linking O3 and its precursors based on chem- predictions of increases in cycling of NO and increasing re- ical reactions that are believed to drive ambient mixing ratios sponsiveness of O3 to NOx. The study data provide a long- (e.g., NARSTO, 2000; Schere and Hidy, 2000). term record that can be used to examine the accuracy of pro- Most of the work developing an observational basis for cess relationships embedded in modeling efforts. Quantify- O3-precursor chemistry derives from field campaigns, some- ing observed O3 trends and relating them to reductions in times focusing on urban conditions. Short-term data are available from aircraft flights, for example, or summer field Published by Copernicus Publications on behalf of the European Geosciences Union. 8184 C. L. Blanchard and G. M. Hidy: Ozone response to emission reductions in the southeastern US measurements made at a variety of locations. Such studies The photochemical regime in the southeast represents hu- usually are limited to a month or two of intense sampling. mid subtropical conditions with urban emissions yielding el- One example in the southern US is the 1990 ROSE exper- evated O3 levels superimposed on a general regional back- iment at Kinterbish, a rural, forested state park in western ground (Chameides and Cowling, 1995). The EPA O3 and Alabama (Frost et al., 1998). This summer study of rural O3 deposition data provide a regional basis for characterizing at low anthropogenic VOC and low NOx mixing ratios pro- trends since the early 1980s (U.S. EPA, 2016a, b). In addi- vided important insights into rural O3 formation (Trainer et tion, the Southeastern Aerosol Research and Characteriza- al., 2000). Other examples of short-term campaigns across tion (SEARCH) project (Hansen et al., 2003; Hidy et al., the US and elsewhere are reviewed in Solomon et al. (2000). 2014) provides measurements that can be used to investi- More recent field studies include New England in 2002 gate changes in O3 production resulting from changes in an- (e.g., Griffin et al., 2004; Kleinman et al., 2007), Texas in thropogenic emissions in the southeastern US. The SEARCH 2006 (e.g., Neuman et al., 2009), the mid-Atlantic region in network of eight sites began with the Southeastern Oxidant 2011 (He et al., 2013), California in 2010 (Ryerson et al., Study (SOS) (Chameides and Cowling, 1995; Meagher et al., 2013), Colorado in 2012 and 2014 (e.g., McDuffie et al., 1998) rural locations, which were near (1) Centreville, AL, 2016), and the southeastern US in 2013 (e.g., Neuman et al., ∼ 85 km southwest of Birmingham; (2) at Yorkville, GA, 2016; Warneke et al., 2016). These campaigns and accom- ∼ 60 km northwest of Atlanta, GA; and (3) at Oak Grove, panying analyses of O3 production and accumulation typi- MS, ∼ 40 km southeast of Hattiesburg, MS, and 75 km north cally address summer, which historically has the strongest of Gulfport, MS, on private land within the confines of the photochemical activity. However, strong photochemical O3 De Soto National Forest (Hansen et al., 2003). Measurements production can occur under special circumstances in winter of some gas-phase species began at these rural sites in 1992, (e.g., Schnell et al., 2009). thus providing a rural data record of over 20 years. Beginning Accounting for an O3 background is important. O3 back- in 1999, SEARCH added five sites in metropolitan Atlanta, ground is associated with biogenic influence, large-scale GA; Birmingham, AL; Pensacola, FL; and Gulfport, MS. transport, or the potential influence of the upper atmosphere Our goal for this study is to extend earlier analyses of the (e.g., stratospheric intrusions, especially during spring; Lin photochemical response of O3 to precursors through 2014, et al., 2012; Langford et al., 2015). The nature and mag- emphasizing relationships between O3 and NOy. We first nitude of background O3 remain an active area of research summarize relevant O3 photochemistry to provide a con- in the US and Europe (Naja et al., 2003; Solberg et al., text for the observational analysis. We then describe trends 2005; Ordóñez et al., 2007; Cristofanelli and Bonasoni, in emissions and ambient pollutant concentrations, and dis- 2009; Arif and Abdullah, 2011; Zhang et al., 2011; Wilson cuss O3, NOz, and HNO3 observations at the SEARCH sites. et al., 2012). Hidy and Blanchard (2015) discuss definitions The trends in O3 mixing ratio, NOx precursor emissions, and of continental and regional background O3. For this study, ambient nitrogen oxide mixing ratios offer important insight we adopt a definition of “background” that includes both the into future changes in O3 and NOy. Blanchard et al. (2014) non-anthropogenic component and the southeastern regional previously explained the majority (66–80 %) of the day-to- component (Sect. 4.4.1). day variations in daily peak 8 h average O3 at SEARCH Field studies have provided observational evidence of non- sites during March–October of 2002–2011 using meteoro- linearity in O3–NOz relationships (e.g., Trainer et al., 1993, logical variables coupled with ambient measurements of O3 1995; Kleinman et al., 1994; Hirsch et al., 1996; Frost et al., precursors (NO, NO2; limited measurements of VOCs) and 1998; Kasibhatla et al., 1998; Nunnermacker et al., 1998; NOx photochemical reaction products (NOz) and a statisti- St. John et al., 1998; Sillman et al., 1998; Zaveri et al., cal model. The previous analyses are extended here for data 2003; Griffin et al., 2004; Travis et al., 2016). Long-term, through 2014 to help understand ongoing and potential future post-1990s data are widely available for O3 and NO2 but O3 changes in relation to changes in ambient NOz and HNO3 detailed observations of total oxidized nitrogen (NOy) and mixing ratios in the southeastern US. Results are discussed VOC, and especially their component species, are typically in relation to modeling predictions by Reynolds et al. (2004) lacking (e.g., Hidy and Blanchard, 2015). One of the longest and others. records of urban and suburban data, comprising a series of short-term campaigns as well as continuous measure- ments, is from southern California. This region exemplifies a 2 Ozone–nitrogen oxide chemistry photochemically active urban regime. An analysis of multi- decadal (since the 1960s) data by Pollack et al. (2013) re- 2.1 Key atmospheric reactions linking O3 with NOx veals how changes in atmospheric chemical reactions have contributed to the observed reductions of O3 in southern Cal- Net tropospheric O3 accumulation occurs when sunlight acts ifornia since 1973. Long-term (more than one decade) mea- on VOC and NOx emissions and the O3 production rate ex- surements characterizing O3 and NOy relationships in both ceeds O3 loss (Trainer et al., 2000).