Tropospheric O3 modeling study: Contributions of anthropogenic and biogenic sources to O3-CO and O3-CH2O correlations

Thumbnail Image
Cheng, Ye
Wang, Yuhang
Associated Organization(s)
Organizational Unit
Supplementary to
Tropospheric O3 and CO are major pollutants in the troposphere. Strong correlation between O3 and CO was observed during the DISCOVER-AQ aircraft experiment in July 2011 over the Washington-Baltimore area. The observed correlation does not vary significantly with time or altitude in the boundary layer. The observations are simulated well by a regional chemical transport model. We analyze the model results to understand the factors contributing to the observed O3-CO regression slope, which has been used in past studies to estimate the anthropogenic O3 production amount. We trace separately four different CO sources: primary anthropogenic emissions, oxidation of anthropogenic VOCs, oxidation of biogenic isoprene, and transport from the lateral and upper model boundaries. Modeling analysis suggests that the contribution from biogenic isoprene oxidation to the observed O3-CO regression slope is as large as that from primary anthropogenic CO emissions. As a result of decrease of anthropogenic primary CO emissions during the past decades, biogenic CO from oxidation of isoprene is increasingly important. Consequently, observed and simulated O3-CO regression slopes can no longer be used directly with an anthropogenic CO emission inventory to quantify anthropogenic O3 production over the United States. The consistent enhancement of O3 relative to CO observed in the boundary layer, as indicated by the O3-CO regression slope, provides a useful constraint on model photochemistry and emissions. As an extension, we analyze the scenario of O3-CO regression slopes in the entire United States and China regions. The O3-CO regression slope ~ 0.3 is simulated over the eastern outflow regions over the ocean. Over the eastern inland regions of both countries, the O3-CO regression slope is lower than that over the outflow region, reflecting in part continuous O3 production in the outflow region. The simulation result shows that the proportion of contribution from biogenic isoprene to the regressed O3-CO slopes various depending on the corresponding local emission scenario. While biogenic isoprene oxidation makes a comparable contribution as anthropogenic emissions in the eastern US, the latter dominates over eastern China. Over the western inland regions of both countries, the O3-CO regression slope can be higher than the eastern inland regions due to transport from lateral and upper boundaries. The observations of O3-CO regression slope provide the means to understand the relative importance of anthropogenic and biogenic emissions on O3 as well as transport. In addition to O3-CO, strong correlations and consistent linear regression slopes of O3-CH2O and CO-CH2O were also observed during the DISCOVER-AQ aircraft experiment in July 2011 over the Washington-Baltimore area. Same as CO, we also analyze the model results to understand the factors contributing to the observed O3-CH2O regression slope by tracing separately three different CH2O sources: primary secondary anthropogenic sources, biogenic isoprene oxidation, and transport from model boundaries. Results show biogenic isoprene oxidation makes the largest contribution to the regression slope of O3-CH2O across much of the eastern United States, providing a good indicator for O3 enhanced by biogenic VOCs. In contrast, the regression slope of O3-CO is controlled by both anthropogenic and biogenic emissions. Therefore, the CO-CH2O linear relationship can be applied to track the contributions to surface O3 by anthropogenic and biogenic factors. Making use of these linear dependences, we build a fast-response ozone estimator using near surface CH2O and CO concentrations as inputs. We examine the quality of this O3 estimator by increasing or decreasing anthropogenic emissions by up to 50%. The estimated O3 distribution is in reasonably good agreement with the full-model simulations (R2 >0.77 in the range of -30% to +50% of anthropogenic emissions). The analysis provides the basis for using high-quality geostationary satellites with UV, thermal infrared, or near infrared instruments for observing CH2O and CO to improve surface O3 distribution monitoring. The estimation model also provides 6 observation-derived regional metrics to evaluate and improve full-fledged 3-D air quality models. The NASA DISCOVER-AQ airborne campaigns were also carried out around the Houston and Denver metropolitan areas in the summers of 2013 and 2014, respectively. Using the 2011 national emissions inventory (NEI), a regional chemical transport model (REAM) is applied to analyze the aircraft observations. We find that major model discrepancies are driven by large underestimates of alkane emissions in both regions. Modeling analysis suggests increases of alkane emissions by a factor of 15 in the Houston Ship Channel, where ship-transport, ship-unloading, storage, domestic transportation of oil take place, and by a factor of 5 in the regions of oil and gas exploration of Denver. The large increase of alkane emissions has drastically different effects on O3 concentrations depending on the strength of biogenic emissions. A useful metric to diagnose the effects of alkane emissions on photochemistry is the least-squares regression slope of O3 to CH2O, which increases by 30% and 80% in Houston and Denver, respectively, due to the increases of alkane emissions, leading to good agreement between model simulations and aircraft observations. Our finding implies that alkane emissions from oil and gas related sources may be substantially underestimated by the NEI, leading to corresponding underestimates of anthropogenic contributions to O3 particularly over the western United States where biogenic VOC emissions are low. In regions like Denver, reducing alkane emissions is urgently required to control summertime O3 concentrations.
Date Issued
Resource Type
Resource Subtype
Rights Statement
Rights URI