Cycles

Fig. 3. Winter (DJFM) temperature in the upper 200 m of the Kola Section in the Bering Sea, 1922-2008. Values for 2003-2008 were reconstructed using SST data from NCEP/NCAR Reanalysis. .

Through the analysis of a vast collection of observational data, Polyakov et al. (2005) have demonstrated that multidecadal fluctuations on time scales of 50–80 yr are prevalent in the upper 3000 m of the North Atlantic Ocean and adjacent Arctic seas. As an example, Fig. 3 shows Atlantic water inflow to the Barents Sea, as reflected in temperature of the upper 200 m of the Kola Section. Yndestad (1999) identified a cycle with the period of 55.8 years among the dominant cycles in this time series (the other two have periods of 6 and 18 years), although it seems closer to 70-80 years in Fig. 3. Later  Yndestad (2006) added the fourth cycle with the period of 74 years to this list. The 74-yr cycle  was also found in the time-series of Barents Sea ice extent (Yndestad, 2006).

Fig. 4. The NAO index defined as the difference between normalized (by standard deviation) winter (DJF) temperatures in Oslo, Norway, and Jakobshavn, West Greenland. The index is smoothed by 5-yr running means and the values are referred to the center of this interval. The smoothed values are from 1883-2006.

Multidecadal variations are also apparent in the NAO index. A spectral analysis of the NAO index time series from 1886 to 1994 computed by Yi et al., 1999 (their Fig. 9b) shows significant power at periods of approximately 50 years. Using a wavelet analysis, Yndestad (2006) identified a 74-yr cycle in the winter NAO index time series.

 A clear long-term cycle is seen in the winter NAO index since the early 1950s (Fig. 1). The index reached a minimum in the 1960s and a maximum in the late 1980s – early 1990s. Since then it declined, almost completing the full cycle. One can notice the difference in the timing of maxima and minima between the NAO index and Kola section temperature. This phase shift will be discussed in the next section.

The NAO index can be defined not only in terms of atmospheric pressure, but also in terms of temperature. Van Loon and Rogers (1978), for example, used the difference in normalized temperature anomalies in Oslo, Norway, and Jakobshavn, West Greenland. Fig. 4 shows the winter NAO index, defined this way. The correlation between this index and the conventional SLP-based NAO index from CRU is 0.75 for the entire period of observations (1883-2006). As seen in Fig. 4, within the envelope of long-term, secular changes, the NAO index exhibits a shorter, 18-20-yr cycle. Yndestad (2006) links this cycle with the 18.6-yr lunar nodal cycle.

Fig. 5. The sum of normalized winter (DJF) temperatures in Oslo and Jakobshavn, smoothed by 5-yr running means.

The sum of temperature anomalies in Oslo and Jakobshavn (Fig. 5) identifies the periods of warming and cooling in the entire northern North Atlantic. Two major periods of warming (in the 1920s-1930s and in the past 2 decades) are clearly reflected in this time series. Between these two periods, the index experienced fluctuations with a period of approximately 15 years. As shown by Moron et al. (1998), the life cycle of the 13-15 year oscillation in North Atlantic sea-surface temperature (SST) exhibits a striking seesaw, with the two maxima of opposite signs occurring between Cape Hatteras and the Bermudas and due south of the Denmark Straits, near the North Atlantic Drift, respectively. Another source of climate variability on this time scale may lie in the Arctic. In a conceptual model introduced by Dukhovskoy et al., 2004, the Arctic climate system oscillates with periods from 10 to 15 years.

Fig. 6. Mean winter (DJF) values of the Atlantic Miltidecadal Oscillation, 1857-2008. Orange line is the 11-yr running means.

When SST is averaged over the entire North Atlantic and the upward linear trend is removed, the residual time series exhibits a cycle with an approximate period of 70 years known as the Atlantic Multidecadal Oscillation (AMO). As seen in Fig. 6, a new warm phase of the cycle started in the late 1990s. If the cycle continues, the North Atlantic will remain relatively warm for another 20-30 years.