The 60year Cycle And Its Role In Ice Extent Changes In Various Regions

While the history of Arctic sea ice extent variability in the twentieth century is characterized by a negative linear trend, there were prolonged periods of time when sea ice extent showed more or less stable increases or decreases that resulted in positive or negative trends, respectively. Some studies (Zakharov, 1996; Mironov,

Figure 2.5. Forms of elementary wave variability used in the wavelet analysis. The vertical axes are dimensionless amplitudes of variability and the horizontal axes are dimensionless shifts in time.

2004; Zubakin, 1987; Vinje, 2000) calculated alternating-sign linear trends for such intervals. These data allow us to coarsely assess the duration of the corresponding cycle of ice extent change over 50 to 60 years.

The presence of such variability both in the western and eastern regions allows us to distinguish three typical epochs: a decrease in ice extent in the first part of the twentieth century and its subsequent increase in the late 1960s-early 1970s, which was again replaced by a decrease in ice extent during the last three decades. The boundaries of the indicated time intervals and the intensity of the changes varied slightly from region to region and from winter to summer.

In Karklin et al. (2001), data on interannual changes in the total ice extent of the Siberian Arctic Seas in the twentieth century were approximated by a polynomial to the sixth power. The curve thus obtained well reflects the long-period variability of the ice extent in the region under consideration. The half-century wave period, identified by these authors for the first time, lasts 55-60 years, which is quite close to the rough estimates given above. Similar variability was detected in air temperature changes in a zone from 72°N to 87°N and in recurrence of the main atmospheric circulation forms identified by Vangengeim (1935) and Girs (1960). Karklin et al. (2001) noted that the long-period ice extent variability in the Arctic Seas, unlike shorter-period variability, does not indicate an opposition in phase between the western and eastern regions, which may testify to their common nature.

Figures 2.7-2.9 show a similar approximation of the changes in total ice extent in the Greenland and Barents Seas in April and in the total ice extent of three western

L, 1000 km3

2200-

2200-

1900

1920

1940

1960

1980

2000

Figure 2.7. Variability of total ice extent in the Greenland and Barents Seas for the period 1900-2003 (April): 1) linear trend, and 2) polynomial trend.

1900

1920

1940

1960

1980

2000

Year

Figure 2.7. Variability of total ice extent in the Greenland and Barents Seas for the period 1900-2003 (April): 1) linear trend, and 2) polynomial trend.

Figure 2.8. Variability of the total ice extent in the Greenland, Barents, and Kara Seas for the period 1900-2003 (August): 1) linear trend, and 2) polynomial trend.
Figure 2.9. Variability of total ice extent in the Laptev, East Siberian, and Chukchi Seas for the period 1900-2003 (August): 1) linear trend, and 2) polynomial trend.

and three eastern seas in August. All three figures have common features: a negative linear trend from the beginning to the end of the twentieth century and long-period (55-60-year) variability. However, the amplitudes of the variability differ significantly: the phases of variability between the western and eastern seas are slightly displaced.

These figures show examples of data plots that can be used to estimate the "50-year" variability of ice extent in various seas and regions. The average amplitudes of variability under consideration, determined at the moments of the largest polynomial curve deviation from a linear trend, allow calculation of ice extent variance, as created by the wave in question, depicted by a curve, using the equation

where A50 equals the average amplitude of the 50-60-year wave.

This equation is based on changes in the harmonic curve, which can differ slightly from the real fluctuation. However, its advantage is that it allows an estimate of variance for the full period, whereas the real fluctuation can include the arbitrary parts of this period, which will influence the value of the variance. The results of calculations using Equation 2.2 for various seas, regions, and seasons are presented in Table 2.5. It shows the small effect (3-6%) of a "60-year" cycle in total ice extent variability of the seas in the North European Basin (Greenland, Barents) in winter and of the seas located to the east of Severnaya Zemlya in summer. This fluctuation is largest (up to 20%) in the Greenland Sea in summer.

The presence of such a clear cycle in ice extent changes in the Nordic Seas region influences estimates of the linear trend describing these changes. As Figure 2.7 shows,

Table 2.5. Average amplitudes of "60-year" components of ice extent and corresponding variance, along with its contribution to the total variability of the ice extent of the seas

Sea, month

^50,

2

thousand km2

106 km2

%

Greenland, April

35.3

623

3.2

Barents, April

55.9

1562

6.5

Greenland + Barents, April

77.9

3034

5.2

Greenland, August

41.2

849

12.3

Barents, August

81.4

3313

19.0

Greenland + Barents, August

111.7

6238

17.6

Greenland + Barents + Kara, August

175.0

15312

17.6

Kara, August

97.5

4753

20.1

Laptev, August

30.0

450

4.6

East Siberian, August

38.0

722

6.2

Chukchi, August

18.0

162

7.6

the phase of the "60-year" variability characterizes the conditions under which positive ice extent anomalies were noted at the beginning of the century and in the 1970s, and negative anomalies were noted in the 1940s and at the end of the century (two waves). Gudkovich and Kovalev (2002) show that these conditions lead to the appearance of a false negative linear trend. The calculations indicate that 10% of the variance described by the corresponding trend (Table 2.1) expresses the false trend influence.

Was this article helpful?

0 0

Post a comment