Key as in Table 6.5

The above analysis should lead to the conclusion that, from 1951 to 1990, the Arctic was dominated by decreases in P and, moreover, that these decreases were more significant than the increases. Thus, it may be reckoned that the changes in this region are not accidental and that they constituted one of the most prominent features of P variability during this time. This can also be seen clearly in Figure 6.13, which presents lines of regression and curves of confidence delineating, with a 95% probability, the scope of theoretical changes in lines of regression for three selected stations with a significant decrease in P. As follows from Figure 6.13 and Table 6.7, the possibility of change in the strength and direction of the trends is significantly less at the stations in the Russian Arctic than in the remaining area of the Arctic. Ifthe factors which shaped P from 1951 to 1990 do not change in the near future, it may be predicted with 95% probability that they will keep decreasing in the area under discussion. Beyond the Russian Arctic, the stability of the tendencies is much lower, and thus the credibility of all prognoses decreases.

Over the period 1961-1990, the spatial distribution of trends is very similar to an analogous distribution over the period 1951-1990 (Figure 6.11). In particular this concerns the border dividing areas with positive and negative coefficients of regression. These coefficients are usually higher over the period from 1961 to 1990. This happens in the areas characterised by both a decrease and an increase in P. The courses of their seasonal and annual values and their trends in selected Arctic stations over the analysed period are presented in Figures 6.10a-i. Both the 30-year trends and those referring to whole period of observations have a similar direction at almost all stations. Significant decreasing tendencies in P (Ostrov Vize, Mys Kamenny, Ostrov Kotelny, Barrow) were caused by high values of P in the 1960s (Figures 6.10c-f). An analysis of all the remaining series of negative trends revealed that their temporal distribution was similar to that of the above mentioned stations in the period under discussion. An analysis of Table 6.6 demonstrates that in the eastern part of ATLSRe, and in the western and central parts of SIBR, there occurred a significant attenuation of trends which remained negative, but which lost their statistical significance. Significant changes in P also occurred in the south of Greenland. At Angmagssalik the trend changed its value from 47.3 mm/10 years to -16.45 mm/10 years, and at Godthab from -10.54 mm/10 years to -92.39 mm/10 years. Similar results of the variability of annual totals of P over the Greenland ice sheet during the period 1963-1988 (calculated using mathematical formulae) were presented by Bromwich et al. (1993). They observed an average decrease in P by 15% for the entire ice sheet. According to them, the reason for such a tendency was that the development of cyclonic activity near Greenland was much stronger in the 1960s than in the subsequent years. In autumn and winter, negative trends were characteristic of about 70% of the analysed stations, and of about 58% of them in spring and summer. In the cool half-year, the number of statistically significant trends was also greater (Table 6.6).

Barrow

Was this article helpful?

## Post a comment