The Sensitivity of Southeast U.S. Ozone Abundance and Production to Dry Deposition

The dry deposition (DD) sink for ozone and its precursors receives less attention than emission sources when interpreting observations, evaluating models, and projecting future ozone abundances. We use the GFDL AM3 chemistry-climate model to bound the role of U.S. DD on ozone abundances over the Southeast USA (SE) during the summer of 2013, coincident with measurements from the NOAA Southeast Nexus aircraft campaign (SENEX). Shutting off DD of oxidized nitrogen (NO_y), ozone, or all species increases mean surface ozone in the SE by 3.4, 12 and 17 ppb, respectively. Based on recent work investigating strong observed inter-annual changes in ozone dry deposition velocity at a northern mid-latitude forest, we perform sensitivity tests with a ±50% change in ozone DD. We compare this with a ±25% change in U.S. anthropogenic NO_x (NO_x = NO + NO₂) emissions to show that if ozone DD varies coherently across the U.S.A., this process exerts a similar influence on regional mean surface level ozone abundance as recent reported trends in NO_x emissions. For example, -50% ozone DD or +25% NO_x increase mean surface ozone by 4.6 or 2.5 ppb, respectively. This suggests that observational constraints on year-to-year variability in ozone DD across broad regional scales are needed to determine the influence of ozone DD on trends in measured ozone concentrations. Further, the sensitivity of ozone to DD implies that ozone DD may play a role in previously reported systematic warm season biases over the Southeast U.S.A. across multiple models.

We also use our sensitivity simulations to compare two metrics for inferring the ozone production efficiency (OPE) based on: (1) an assumed linear relationship between abundances of ozone and NO_x oxidation products (NO_z = NO_y - NO_x), often used to estimate OPE directly from observed concentrations, and (2) on model diagnostics of the production of odd-oxygen and nitric acid. The former offers insight into ozone production chemistry from field measurements, while the latter cannot be observed directly but reflects the underlying chemistry. Our analysis demonstrates that each metric responds differently to changes in simulated deposition velocity, implying that variations in dry deposition (of both NO_y and O₃) may lead to incorrect interpretation of trends in the concentration-based OPE. The present dearth of long-term observational sites for dry deposition may thus preclude comprehensive attribution of recent trends in ozone chemistry to changes in its sources or sinks.