Previous: Data Next: Acknowledgements Up: Ext. Abst.

3. Results

Long term ozone change over Marambio was studied with linear regression analysis from monthly means of the classified and unclassified data. Standard error of the slope of the fit estimates the uncertainty of the fit. The results are presented in Table 1. The predominant feature of the ozone behaviour over Marambio is decrease. 4/5 of the changes when all cases are included (Table 1 a) are negative and all changes larger than two times their standard error are negative. The decrease is most pronounced in spring during the chemical

During months of January, February, March and April when there is no strong vortex the mean total column ozone at Marambio has been 277 Dobson units (DU), which is 6 % less than the long term mean at AI.

During the August-December period the total column ozone outside the vortex in Marambio has been 305 DU, which is 9 % less than in AI during 1957-67. The Marambio total ozone inside the vortex has been on average 41 % less than in AI. Significant linear ozone trends inside the vortex are found in September at 435 K and 475 K surfaces (-10%/year) and in total ozone (-2 %/year), in October and November in total ozone (-3 %/year) and in November at 435 K surface (-6%/year).

The overall feature of long term ozone over Marambio during 1987-1999 is decline. This is especially so during spring and for the situations when Marambio situates inside the vortex. There is decline to be seen also for the situations when the vortex edge is above Marambio, but it is less pronounced and actuates little earlier in the season (especially during August). The decline is characterised by a strong decline in the beginning of the period and less or absent decline in the latter half of the period. This is very much comparable to the springtime long term lower stratospheric ozone at Syowa, Neumayer and South Pole presented by Uchino et al. [1999]. Absent decline in the latter half of the period imply to the possibility that the chemical ozone depletion might already have reached its maximum in lower stratosphere. The first results from the early spring of 2000 (not shown) nevertheless do not speak in favour of this.

There is no negative trend seen for the situations when Marambio is outside the vortex at lower stratosphere, which does not speak in favour of substantial mixing through the vortex edge or substantial processing of air just outside the vortex. If substantial mixing had been happened, a long-term decline in ozone would have been expected outside the vortex as well. Strong conclusion is nevertheless hampered by scarcity of the outside vortex data.

The earliest significant linear trends in ozone are seen in August at the edge region (Table 1c) and on the other hand already in May (Table 1a) when all cases are included. According to the sonde temperature measurements in Marambio PSC I treshold temperature can be reached occasionally already in May, more often in June and constantly in July. Thus the suggested early start (May, June, July) for the chemical ozone depletion especially in the vortex edge region [Vömel et al., 1995; Roscoe et al., 1997; Wauben et al., 1997; Lee et al., 2000] is supported by the Marambio data.

Previous: Introduction Next: Results Up: Ext. Abst.