Vertical meteorological processes, such as subsidence and stratospheric intrusions, can transport significant concentrations of ozone down from the lower stratosphere and upper troposphere into the mid to lower troposphere. These processes are often neglected or poorly simulated by regional and urban air quality models because of inadequate specification of the initial and lateral boundary conditions for ozone and other trace gases. Previous studies have also found a strong correlation between ozone and potential vorticity (PV) in the mid to upper troposphere such that the ozone concentration equals PV times a constant. PV has higher values at levels above the tropopause; therefore, large PV values within the troposphere can be used as an indicator of stratospheric air at tropospheric levels. In this study, we examine the downward transport of ozone over a 14-day period in the summer of 1991 over eastern North America using PV to supplement ozone observations. During this period, ozonesonde measurements were made once or twice a day at Cape Race, Newfoundland.
The National Center for Environmental Prediction’s (NCEP) global reanalysis fields (2.5o grid spacing) are used to produce three-dimensional PV fields. High values of PV are obtained in the mid troposphere between 5 and 8 km MSL over Newfoundland that correspond to peaks in the ozone concentrations at these levels as a result of several tropopause folds during the period. In contrast to other studies, we have found that a height-dependent constant of proportionality between ozone and PV best fits the observed ozonesonde profiles. As expected, there is a high correlation between ozone and PV in the lower stratosphere. While there still is a high correlation in the troposphere, it gradually decreases closer to the ground. Because observations have shown that intrusions can develop into narrow and elongated streamers, the NCEP reanalysis fields may not adequately represent the spatial and temporal characteristics of every stratospheric intrusion event. A dynamical mesoscale model (RAMS) is therefore used to produce meteorological analyses at a higher resolution (120 and 60 km grid spacing). The variations in the PV fields from the mesoscale model are in better agreement with the observed ozone profiles. Ozone concentrations based on the mesoscale PV values are then used for the initial and lateral boundary conditions of a regional-scale photochemical model. Large variations in the ozone concentrations are predicted in the mid troposphere over eastern North America, suggesting that the “background” concentrations of ozone in the mid-troposphere needs to be accounted for by regional-scale air quality models. The predicted ozone fields and diagnostic ozone fields (based on PV values) are also compared with satellite measurements from the Total Ozone Mapping Spectrometer (TOMS). The sensitivity of the ozone predictions to the specification of the boundary conditions will be discussed.
Symposium on Interdisciplinary Issues in Atmospheric Chemistry