Presented at the AQAST 9 th Semiannual Meeting Wednesday June 3 rd, 2015 Presentation by: Dan Goldberg, Ph.D. Candidate Co-authors: Tim Canty, Tim Vinciguerra, Ross Salawitch & Russ Dickerson Evidence for an increasing geographic region of influence on ozone air pollution in the eastern United States 0

Motivation for this study… 1 The state of Maryland is very interested in the role of the interstate transport of ozone. Science objectives: How much ozone is generated locally vs. transported from upwind regions? What are the future trends in surface ozone? Is ozone being simulated for the right reasons?

Ozone Source Apportionment Examples 2 The CAMx software can attribute ozone to different source regions. Surface ozone, once formed, can be transported 100’s km downwind from the original source. MD OHPA

Ozone from the model boundary 3 Ozone from the boundary is uniformly greater than 15 ppbv. Some locations, especially close to the boundaries, are higher. Ozone attributed to areas beyond the model domain, i.e., Texas, Cal, Asia

Summer 2011 & 2018: Ozone Source Attribution The attribution of ozone in all states decreases 10 – 25 % over 7 years. The only portion to increase is the ozone attributed to the model boundary. Baltimore, MD BC=boundary conditions 4

Trends in the Apportionment of Surface Ozone Boundary and meteorology are initialized identically in each simulation. Total surface ozone has decreased and is projected to further decrease. Sources inside the model domain will decrease. If the sources outside the model domain remain the same, ozone attributed to these sources will increase. 5 Mean 8-hr max in Baltimore, MD

Trends in the Apportionment of Surface Ozone An increasing role of the boundary is seen in all metropolitan areas in the eastern United States. 6 Mean July percentage of ozone (%) attributed to the boundary Mean July concentration of ozone (ppbv) attributed to the boundary What is causing this increase??? We think it’s related to HO 2 chemistry!

HO 2 Chemistry in the eastern United States Mean July 2011 daytime (7 AM – 7 PM EDT) HO 2 concentrations The HO 2 + O 3 reaction can be an important sink of O x (O 3 +NO 2 +…) when HO 2 >15 pptv. 7

Change in daytime HO 2 concentrations between July 2002 and 2018 Surface 2 km5 km 1 km Daytime HO 2 concentrations are decreasing in most areas. This is increasing the lifetime of ozone when reaction with HO 2 is important. Ozone lifetime with respect to reaction with HO 2 is increased ~5% over 16 years. 8

Change in daytime HO 2 concentrations between July 2002 and July 2018 (with VOC reductions ONLY) Surface 2 km5 km 1 km Anthropogenic VOC reductions are partially (but not fully) responsible for the decrease in HO 2 concentrations. 9

Curtain plots of Ozone at the Boundary during July 2011 Modeling domain MOZART is marginally higher in the lower layers at all boundaries, except the Southeast. GEOS–Chem has higher ozone in the mid– troposphere, especially at the western boundary (which is the boundary that most often influences ozone in the eastern United States). 10

Prediction of O 3 precursors: Using DISCOVER-AQ data 11 The comparison with data from the P3-B aircraft during DISCOVER-AQ MD shows a significant over prediction of NO y and a significant under prediction of HCHO. We’ve made three changes to update the model, “Beta”: CB6r2 gas-phase chemistry (Old: CB05) MEGAN v2.1 biogenic emissions (Old: BEISv3.14) Reduce emissions from mobile sources by 50% (Anderson et al., 2014)

12 Prediction of O 3 : Using DISCOVER-AQ data Prediction of ozone is similar in each case, but how the ozone produced is much different.

Policy Implications of the model updates 13 The model will be more responsive to NO x emission changes. Percentage of O 3 formed in a VOC-limited environment during the daytime of July 7, 2011 BaselineBeta

Conclusions Boundary conditions play a critical role: –Ozone from the boundary is larger by percentage and by concentration in future years. –HO 2 decreases are responsible for an increase in the ozone lifetime. This is an unintended consequence of the policies to reduce NO x and VOC emissions. –There are large differences between MOZART & GEOS-Chem aloft. Updates to the model to give a better prediction of NO y and HCHO. –The model will respond better to reductions in NO x emissions, which is a better representation of what is happening in reality. 14

EXTRA MATERIAL 15

Fate of the HO 2 16 In urban areas: High concentrations of NO react with HO 2 to create NO 2 In rural areas: Low concentrations of NO HO 2 + O 3 is not a sink for ozone VOCs oxidize or photochemically react to create HO 2 As VOCs and NO x decrease, so does HO 2 Low concentrations of HO 2 High concentrations of HO 2 HO 2 + O 3 can be a sink for ozone Sources of HO 2 :

Policy Implications of the model updates 17 On-road vehicles contribute ~40% less than in the baseline version. Power plants contribute 10 – 20% more than in the baseline version. *Similar increases are seen for all non-vehicular emissions (ships, area sources, etc.) Slides show ratio of the Beta version: Baseline version “Beta”: CB6r2, MEGANv2.1, 50% reduction of NO x from vehicles

South EastWest North July 2011: Ozone attributed to the Model Domain Boundaries ~2/3rds of Boundary ozone in Maryland came from the Western Boundary 18

Fraction of total surface ozone attributed to the boundary conditions, Maryland, and everywhere else in the modeling domain The magnitude of the worst air quality days seems to be determined by the amount of interstate transport. Baseline Local Southerly Transport Southwesterly Transport 19 July 2011: Ozone Source Apportionment Modeling domain