Tropospheric Ozone Dsitrubtions in the Tropical Western Pacific Based on Observations, CAM-Chem, and Reanalysis Simulations

Tropospheric ozone affects air quality and human health, and tropospheric ozone also absorbs both longwave and shortwave radiation, causing it to be a significant greenhouse gas (Anderson et al., 2016). In order to accurately represent ozone in atmospheric models and reanalysis simulations, controlling mechanisms of ozone and ozone distributions must be understood and represented in models and simulations. This study will use in-situ data from the NSF/NCAR CONTRAST campaign to validate ozone distributions from CAM-Chem, MERRA-2, and ERA-Interim.

Previously, a bimodal distribution of ozone ad an anti-correlation between ozone and RH have been observed over the tropical western Pacific Ocean (e.g., Pan et al., 2015). When evaluating model and reanalysis data, bimodal ozone distributions are seen in CAM-Chem, MERRA-2, and ERA-Interim. However, their primary modes of ozone probability density function (PDF) are much higher than the observed value (18 pbbv), which are 20 ppbv, 31 ppbv, and 36 ppbv, respectively. In addition, their secondary modes are smaller than the observed values (60 ppbv) which are 45 ppbv, 45 ppbv, and 43 ppbv, respectively.Focusing on the anti-correlation between ozone and water vapor, both MERRA-2 and ERA‑Interim data miss many of the dry layers that are respectivelyobserved in the CONTRAST data, at 200 hPa-350 hPa and between 800 hPa-1000 hPa. Global maps interpolated on the 320 K isentropic level show that dry air (RH < 20%) frequency and high ozone concentrations ( > 40 ppbv) spatially correlate with each other in CAM-Chem data but not in MERRA-2and ERA-Interim data. These results show that CAM-Chem, MERRA-2, and ERA-Interim do reproduce the anti-correlation between ozone and RH, but to varying degrees.