Considerable impacts of meteorological field choices on tropospheric ozone simulations by a global chemistry transport model

The meteorological field serves as a vital input for chemical transport models (CTMs) to simulate tropospheric ozone pollution. For global CTMs, it is often directly provided by reanalysis meteorology datasets. Different choices of meteorological fields are likely to yield varied ozone levels and contributions of ozone-related processes, e.g., transport, chemical production/loss and dry deposition, to the variations of ozone pollution. However, relevant comparisons are seldom reported. Here we investigate the impact of meteorological field choices on the results of tropospheric ozone simulations performed by the TM4-ECPL global model. Two generations of meteorological reanalysis products from the European Centre for Medium-Range Weather Forecasts (ECMWF), ERA-Interim and ERA5, are selected as input meteorological fields to drive the model for this comparative study. Results show that when driven by ERA5 meteorology, the model generates considerably higher mean mixing ratios of tropospheric ozone for the period of 2013-2017 compared to ERA-Interim-based results. This outcome is particularly pronounced in the high-latitude areas of both hemispheres and near the surface (by 5-10 ppbV). The model results by using both meteorological fields are validated against air quality monitoring data from over 10,000 sites and ozonesonde measurements globally. When using ERA5, the overestimation of near-ground O_x (ozone + NO₂) levels by the model becomes more notable than using ERA-Interim (increases from 6% to 25%). However, the underestimation of ozone levels in the middle and high troposphere is reduced. Both simulations can well reproduce the annual trends of near-surface ozone pollution, indicating a more important role of ozone precursor emissions in driving ozone changes. Furthermore, through sensitivity simulations and budget analysis, we delve into the reasons behind the considerably higher ozone levels in the ERA5-driven simulation.