Previous: Aerosol measurements by LIDAR Next: Conclusion Up: Ext. Abst.

III. Analysis and discussion

III.1. Analyses from daily ECMWF assimilation

From examination of meteorological fields given on a daily basis by the ECMWF as derived for the 400 K potential temperature level (in the lower stratosphere), Potential Vorticity (PV) contours have been computed for the whole studied period (April 21 - June 30). FIGURE 2 shows the longitude´latitude cross-sections of the PV field obtained on April 21^st and 29^th.

One can note that, on the 29^th, a large low-PV zone occurs and stands from 25° to 45°South and from 15° to 75° East : 2 £ PV £ 6 PVU. Its centre is situated around 30°S and between 28° and 44°E with the lowest PV values : PV £ 3 PVU. For comparison, the 21^st PV map, which is representative of the studied period, indicates that expected PV values in the anomaly zone are about 8 PVU.

III.2. High resolution PV advection contours

In order to check about the origin of the increase, a high resolution advection model has been used for the period from 21 to 30 April. That model can produce a continuous evolution. The non-conservative terms of the large scale field are taken into account using a relaxation toward the ECMWF analysis. It has been used for real time 5 days forecasts of the PV fields and has been found useful for the sub-tropical and polar filaments detection and alert in the frame of the METRO project MEridional TRansport of Ozone in the lower stratosphere), a research project included in the European campaign THESEO (THird European Stratospheric Experiment on Ozone).

FIGURE 2 : Potential Vorticity analyses derived from ECMWF fields, for the 400 K potential temperature level. By comparison with the 21st PV map (a), a large low-PV zone is well identified on the 29th (b). It stands from 25° to 45°south and from 15° to 75° east.

Four PV maps derived for the 400 K level from ECMWF assimilation fields and taken between April 21^st and April 30^th, are presented on FIGURE 3. They show that a "tongue" of low-latitude, low-PV air has extended from tropics over the south American continent :

• The tropical air mass was initially localised nearly over Brazil, on the 23^rd, and has moved south-eastward the days after.
• On the 25^th the front side of the event is localised in the middle of the Atlantic Ocean around 30°S and has the same south-east motion direction
• On the 28^th the front side is situated nearby the west coast of southern Africa, with a northward development.
• That low-PV filament has stretched out and wrapped up nearly all South Africa, on the 29^th and 30^th.

(a)	(b)
(c)	(d)
FIGURE 3 : PV contours as obtained with a high resolution PV advection model for April (a) 23rd, (b) 25th, (c) 28th and (d) 30th.

Those PV contours are in favour of a tropical/extra-tropical mass exchange through filament events in a motion similar to the one produced by planetary-scale breaking waves in the stratosphere (Leovy et al., 1985 ; McIntyre and Palmer, 1983). In fact, it is the large-scale Rossby waves in the Westerlies that pull the tracer from tropics into mid-latitudes in the form of large-scale filaments through the surf zone (McIntyre and Palmer, 1984).

From analysis of the ECMWF zonal wind component (not shown) we have noted that the wind direction had changed in the upper stratosphere during the studied period: the zonal wind was getting westerly, what permitted to the Rossby planetary modes to propagate upward into the stratosphere. Besides, the LIDAR temperature profiles recorded over Durban reveal much variability from day to day (not shown). what highlights disturbances due to the planetary waves that propagate through the westerlies upward into the stratosphere in middle and high latitudes.

Moreover, when we examine the water vapour distributions in the troposphere, as revealed by satellite Infrared images (not shown), we note the presence of a convection event localised over Brazil on the 23^rd.

In connection with the PV contours shown above (FIGURE 3), the observed convection event could be associated to a tropospheric source of aerosols, which may have been :

• first pulled upward through the tropopause into the stratosphere,
• and then horizontally advected from tropics to mid-latitudes through the southern subtropical barrier.

Examination of water vapour distributions in the troposphere (not shown) obtained on the 29^th, shows that the tropospheric jet stream, which could have also induced a tropospheric-stratospheric exchange, was localised nearly in the middle of the Atlantic ocean, far from the south-western coast of Africa. What supports the suggestion that the observed aerosol loading event, that happened in the upper troposphere and lower stratosphere over Durban site, results meanly from a tropics/extra-tropics meridional air mass transfert.

Previous: Aerosol measurements by LIDAR Next: Conclusion Up: Ext. Abst.