Previous: Mean Fields  Next: ConclusionsUp: Ext. Abst.

4. Case Study

4.a Synoptic Evolution

During the second week of the month, strong ozone perturbations were detected over southern SA, consisting of a transient ozone wave evolving with a southwest-northeast direction across the continent. From July 7^th to10^th (figs.4a-d) ozone values as low as 240 DU could be followed along four days from southern Patagonia to southwestern Atlantic. At the same time, high ozone values (up to 400 DU -fig.4b-) were observed to evolve over Buenos Aires and to the east.

The analogous evolution of the atmospheric fields shows a clear correspondence between ozone relative minimum (maximum) and: ridges (troughs), lowered (enhanced) tropopause pressure values (figs. 4e-h) and relative maximum (minimum) of potential vorticity on isentropic surfaces (figs. 4i-l). By use of a simple conceptual model, Vigliarolo et al. (2000) exhibited the dynamical relationship between a synoptic wave and the resulting ozone distribution, provided ozone-mixing ratio is conserved on these time scales. Moreover they showed by composite analysis for the preferred winter synoptic-scale mode of variability, that waves evolving along subpolar jet latitudes are responsible for the corresponding transient ozone-pattern. This is due to the barotropic-equivalent structure of these waves and the fact that ozone partial-pressure attains a maximum very close to the level where these waves also maximise.

In order to address the relative dynamical contributions of the atmospheric waves to ozone changes over southern SA region, two approaches are followed. The first, relates ozone daily changes to horizontal and vertical transports of ozone mixing ratio by transient motions (subsection 4b), while in the second ozone distribution is fitted to a multivariate linear relationship with tropopause pressure and potential vorticity at several isentropic levels (subsection 4c).

Figure 4: Temporal evolution from July 7^th to 10^th of: (a)-(d) ozone, (e)-(h) tropopause pressure, (i)-(l) potential vorticity (shaded) and streamlines (thick contours) on the 330ºK surface. Contour interval is (a)-(d) 20 DU; (e)-(h) 20 hPa; and (i)-(j) 2*10^-8 m2 s^-1 kg^-1.

4.b Mapping Technique

A mapping technique was used to provide estimates of three dimensional ozone-mixing ratio horizontal gradients based on Meteor 3/Toms total ozone and SBUV/2- NOAA 11 ozone mixing ratio vertical distribution. This technique allows meridional ozone mixing ratio estimations supposing a linear relationship between the former and the meridional ozone gradient. Then, the same derived linear coefficients are used to estimate zonal ozone mixing ratio gradients from total ozone zonal gradients. It is worth to note that for July 7^th to 10^th, the SBUV/2 profiles used correspond to the scan of the satellite that passes over the continent, as the subsequent scans are spread about 25º east and west. Hence, the validity of the above assumptions is much accurate in a domain close to the satellite path. In addition, only data from 100 to 30-hPa were used, as above that upper boundary the dispersion relationship between the "mapped" variables becomes non-significant.

With a similar procedure, we also estimate three-dimensional ozone mixing ratio-distribution in terms of total ozone distribution and SBUV/2 ozone mixing-ratio profiles.

Figure 5: (a) Local ozone change from July 8^th to 7^th and (b) total (zonal, meridional and vertical) advection contributions to ozone changes integrated over the 100-30 hPa layer. Contour interval 2*10^-4 DU s^-1, all zero contours have been omitted.

The most simple transport ozone equation where the ozone-mixing ratio is conserved following the motion is:

where  stands for ozone mixing ratio, u and v are the horizontal wind components and  is the vertical velocity. Then multiplying by the appropriate quantities and performing a vertical integration of (1), ozone local changes between two close pressure levels may be obtained as the combined response of both horizontal and vertical transports of ozone mixing ratio integrated in the pressure layer.

In that sense, local ozone change between July 8^th and 7^th is shown in fig.5a while the contributions from zonal, meridional and vertical (total) advection terms integrated in the 100-30 hPa layer is shown in fig.5b. Both fields are in good agreement although the total advection field overestimates ozone local changes by approximately a factor of 1.7. This may be due to: a) the poor estimates of the vertical velocity field in the stratosphere, which may be very low compared to the real ones. b) The limitation of the mapping technique itself for levels above 30-hPa, although the analysis of SBUV/2 profiles over the region indicate that the contribution to local ozone changes came from the 125-62.5 hPa layer, with secondary contributions coming from both 62.5-37 hPa and 250-125 hPa layers. c) The transports taken only from 100 to 30-hPa, which may be misrepresenting the real situation.

Figure 6: Contribution to total advection for July 8^th to 7^th from (a) zonal, (b) meridional and (c) vertical motion fields; each integrated over the 100-30 hPa layer. Contour interval 2*10^-4 DU s^-1, all zero contours have been omitted.

For the same day, each component of the total advection term integrated over the 100-30 hPa layer is shown in fig.6. Negative local ozone changes over the central part of Argentina and extending south-east to the Atlantic ocean are due to the combined contribution of both zonal and meridional advection. Meanwhile, the positive ozone changes located to the north of Uruguay and eastern Atlantic come from the zonal advection term partially offset by meridional and little vertical advection. To the eastern part of Tierra del Fuego and over 57ºS, 55ºW too much negative values are produced by the zonal advection term, although ozone changes over the region are positive (fig.5a).

4.c Lineal multivariate regression model

Suggesting that total ozone variability on synoptic time-scales is mostly explained by quasi-columnar motion of air along isentropic surfaces, Salby and Callahan (1993) derived a linear regression model that relates total ozone content to tropopause pressure and potential vorticity. In order to test this hypothesis and to find suitable predictors for ozone distribution, the following relationship was proposed:

where  is total ozone,  represents the tropopause pressure,  stands for the potential vorticity on isentropic surfaces of 315ºK, 330ºK and 450ºK, and , i=0, 4 are the corresponding regression coefficients.

To find the regression coefficients that fit the relationship given by (2), data from the entire month of 1994 July was used according to least square method. Then, (2) was resolved for the grid points of the SH when the corresponding time series of the variables involved have no gaps within the month. As an example, figure 7a shows the time evolution of the relative contributions from terms in (2) averaged over 35º-55ºS, 50º-70ºW, an area that encloses the eastern part of southern SA. Contributions from all terms (tropopause pressure and the sum of potential vorticity contributions) are in very good agreement with almost over the whole month, excepting in the period July 25^th to 30^th when the differences between each curve become the largest (about 28 DU). Apart from some few days at the beginning of the month, potential vorticity remains the main contributor to , while an analysis of potential vorticity terms discriminated for isentropic surfaces (fig.7b) shows that  pv₃₁₅  is the most important term, which in turn is comparable to tropopause pressure contribution.

Figure 7: Time evolution of contributions from the different terms in (2). (a)    is represented by the continuous curve in red; contribution from tropopause pressure is labelled in green while that belonging to the sum of all potential vorticity at 315ºK, 330ºK and 450ºK is denoted by blue. The long dashed orange line represents the sum of all contributions (from tropopause pressure and potential vorticity). (b) Individual contributions from potential vorticity at 315ºK (light blue curve), 330ºK (orange) and 450ºK (rose); the sum of all this contributions are labelled in blue while contribution from tropopause pressure (green long-dashed) was plotted for comparison purposes.

For July 8^th, figure 8 presents: (a) field, and the contributions from: (b) the sum of potential vorticity from the three isentropic levels and (c) the tropopause pressure over southern SA. In particular, over the western Pacific, southern Patagonia and Malvinas Islands, and the band extending from the continent at 43ºS into the Atlantic Ocean to the east, relatively low values of the field are mainly due to decreased tropopause pressure (figs.8c-4f) and relatively low potential vorticity (figs.8b-4i). On the other hand, high values of over north of Uruguay and southeastward over the Atlantic (fig.8a) are due to relatively low potential vorticity (fig.8b-4I) and tropopause pressure enhancement (fig.8c-4f) both linked to a cyclone evolution over the area.

Figure 8: (a)  field, and the contributions from (b) the sum of terms of potential vorticity from 315ºK, 330ºK and 450ºK, and (c) tropopause pressure. Contour interval is 30 DU.  

Previous: Mean Fields  Next: ConclusionsUp: Ext. Abst.