Review of winter 2010 in Europe

The winter of 2010 in Europe was much colder than expected. In many places seasonal temperatures dropped to the values not seen in decades. This cold winter was associated with an exceptionally strong negative phase of the North Atlantic Oscillation. The roles of solar activity, quasi-biennial oscillation of the stratospheric winds and other factors are discussed. Interestingly, temperatures in the European Arctic Seas were well above normal; however, they cannot be explained simply by the negative phase of the Arctic Oscillation, as in the NSIDC report. The unusual behavior of the climate system in the winter of 2010 raises a question about a possibility of a climate regime shift.

Discussion

Unlike the forecast for North America, the forecast for Europe was incorrect. The mean winter (DJF) temperatures were expected to be close to the 1971-2000 average over much of Europe, but they turned out to be much colder (Fig. 1). Preliminary data shows that winter (DJF) temperature in Stockholm was the coldest since 1996, in Berlin since 1987, and in London since 1985. Provisional figures from the Met Office show that the mean UK temperature was the lowest since 1979.  Warmer than normal temperatures were observed only in southeastern Europe.

a) Forecast.	b) Observed.
Fig. 1. Forecast (a) and observed (b) temperature anomalies in the winter (DJF) of 2010.

The primary reason for the cold winter in Europe is unusually strong high-latitude blocking of the westerly flow in the atmosphere and, as a result, an extremely negative phase of the Arctic Oscillation (AO) and the North Atlantic Oscillation (NAO). Figure 2 shows the time series of the winter (DJF) NAO index for the period 1824-2010. The time series was constructed using the data from the CPC since 1951, and prior to that from CRU. The CRU index was adjusted so that its mean and standard deviation correspond to those of the CPC index for the overlapping period 1951-2008. As seen in Fig. 2, in the winter of 2010 the value of the NAO index was the lowest on record.

Fig. 2. Winter (DJF) NAO index, 1824-2010.

What caused the unusual high-latitude blocking and a sharp decline in the AO/NAO indices? One factor that may have played a crucial role is solar activity, which was very low for a long period of time, delaying the beginning of the 24th solar cycle. This prolonged state of low solar activity better matches the 11-yr solar cycles in the late 19th – early 20th centuries, when the climate of the Northern Hemisphere was cold, than solar cycles in the more recent period.

When solar activity is low, there is a tendency for high-latitude blocking of the westerly winds. More specifically, for the period 1948-2010, when sunspot numbers were below 25, in 14 out of 18 of such winters (78%) the Greenland blocking (GB) index (defined as in Zhifang and Wallace, 1993) was positive. The GB index, in turn, is strongly correlated with the NAO index (r = -0.81 for the same period). In the winter of 2010, the GB index was record high, with the value three times greater than the standard deviation.

Another factor that may have contributed to the cold winter in Europe is the Quasi-Biennial Oscillation (QBO) of stratospheric winds. According to Labitzke and van Loon (1993), at solar minimum the circumpolar vortex is weak during the east phase of the QBO, which was observed this winter. It is interesting to note that the phase transition in the QBO was slower than usual, which was consistent with model calculations of Mccormack (2003) showing that the duration of the west phase is about one month longer at solar minimum than at solar maximum.

The combined solar-QBO signal is not as robust in the surface climate as in the stratosphere and is sensitive to the upper boundary conditions of the troposphere (Baldwin et al., 2001); however, it still appears to be traceable (Coughlin and Tung, 2001). Thus,  there is an increased probability of heavier ice winters in the Baltic Sea during low solar activity and the east QBO (Loewe and Koslowski, 1998). Similarly, the composite maps from Bochnicek and Hejda (2006) show negative temperature anomalies at 700 hPa level over Eurasia during winters of low geomagnetic activity and the east QBO phase.

Fig. 3. Correlation between winter (DJF) values of the WPO index and SLP, with the latter lagging the index by one year.

The state of solar activity, QBO phase, and the distribution of summer-fall SST anomalies in the North Atlantic, which favored negative NAO, were all taken into account in the winter 2010 forecast. However, other factors were thought to have a potential to counterbalance the effect of the above factors. Among the predictors pointing to a possibility of a positive NAO phase this winter were anomalously low snow cover in Eurasia in the preceding months (Bojariu and Gimeno, 2003) and some processes in the Pacific basin. An example of the latter is the relationship between the West Pacific Oscillation (WPO) and the NAO. It was found that the correlation between the winter (DJF) values of these two indices was statistically significant when the WPO was leading the NAO by one year. For the period 1961-2009, r = 0.61 (or 0.59 for the detrended values). The NAO dipole is clearly seen in Fig. 3, showing the correlation map between the WP index and SLP values over the Northern Hemisphere with the time lag of one year. In the winter of 2009 the WPO index was 1.24, the third largest value since the beginning of record in 1951. Therefore, based on the above relationship, the NAO index was expected to be positive during the winter of 2010.

Another factor that suggested a possibility of a mild winter in 2010 was the existence of quasi-biennial fluctuations in surface temperature in central Europe. The analysis of the 130-yr time series of January temperature in Berlin (1881-2010) showed that if temperature dropped from one January to the next by 1.06 standard deviation or more, the temperature in the following January is very likely to increase. Indeed, out of 25 such cases in the time series, there was just one case when temperature continue to drop for the second year in a row. And that case was January 2010. This and other examples show how unusual was the winter of 2010.

a) November 2009 - forecast.	b) November 2009 - observed.
c) December 2009 - forecast.	d) December 2009 - observed.
e) January 2010 - forecast.	f) January 2010 - observed.
g) February 2010 - forecast.	h). February 2010 - observed.
Fig. 4. Forecast(left column) and observed (right column) monthly temperature anomalies from November 2009 through February 2010.

A comparison of forecast and observed monthly temperature anomaly fields for the months from November 2009 through February 2010 is shown in Fig. 4. These monthly forecasts were significantly influenced by the work of Moron and Plaut (2003), who examined the frequencies of weather regimes in the Atlantic–European sector during El Niño and La Niña events. They found that in November-December zonal weather regimes (positive NAO) were more prevalent during El Niño, while blocking of the westerly flow (negative NAO) was frequently observed during La Niña. In January-February, the situation is almost inverted, that is, blocking tends to occur more frequently during El Niño than during La Niña. This relationship worked very well for our 2008 forecast during the La Niña event, but mostly failed in the winter of 2010.

November 2009 was anomalously warm in Europe, which was in line with the forecast (Figs. 4 a,b). The first decade of December was also anomalously warm, but then a strong upper atmospheric ridge (so-called “omega block”) was formed over the eastern North Atlantic, accompanied by a deep trough over Europe (Fig. 5). This trough sort of opened the door for cold Siberian air, which penetrated as far as western Europe. The storm track was shifted far south of its normal position. Sharp temperature contrasts between cold Siberian air and warm Atlantic air led to intensification of those storms that brought heavy precipitation to southern Europe and often wreaked havoc along their route (see a Wikipedia article about Winter of 2009–2010 in Europe). Frequent blocking patterns continued with some variations in January and February.

It is interesting to note that while Europe was freezing, the Arctic was anomalously warm. The National Snow and Ice Data Center (NSIDC) in their online report  explained it by the negative phase of the AO, that “tends to bring warmer than normal temperatures to the Arctic.” That was quite surprising, because the negative AO can explain warmer than normal temperatures and reduced ice cover in places such as the Baffin Bay-Labrador Sea area, but not in the European Arctic Seas, where sea-ice extent this winter was well below normal (map). As an example of anomalous warmth in this region, monthly temperature departures from the 1978-2009 average in Svalbard Luft (78.2N, 15.5E), were +6.1C in December 2009 (third warmest after 1984 and 2005) and +6.5C in January 2010 (third warmest after 2006 and 1990).

Fig. 5. Omega block in mid-December 2009.

Moreover, until recently the accelerated Arctic warming that started in the 1990s was largely explained by the positive phase of the AO, which steered ocean storms farther north, especially after its winter index jumped to the record high value in 1989 (e.g., Moritz et al., 2002; Rigor and Wallace, 2004; Wu and Straus, 2004). Also, it has been speculated that the greenhouse forcing may favor the positive mode of the AO that also favors sea ice losses (Shindell et al., 1999; Moritz et al., 2002). Prior to the 1990s there was no evidence of warming in the Arctic Ocean in the second half of the 20th century. In contrast, Kahl et al. (1993) found significant surface cooling over the Arctic Ocean during the period 1950-1990. In more recent years, the AO index was fluctuating closer to its normal values, but the Arctic remained anomalously warm. This discrepancy between the index and high-latitude temperatures was called “the Arctic paradox” (Overland and Wang, 2005).

Fig. 6. Regime shifts in winter temperature in Berlin.

Anomalously warm temperatures in the European Arctic Seas this winter are the result of warm air advection along the western periphery of the strong upper atmospheric ridge in the Northeast Atlantic. The latitudinal position of the blocking ridges, however, significantly changes from month to month and from year to year. As noted by Lamb (1972), even during the Little Ice Age, some regions occasionally experienced anomalously warm winters. The overall atmospheric circulation during the LIA was characterized by increased frequency of meridional types of atmospheric circulation and reduced zonal flow consistent with the negative phase of the AO (Moritz et al., 2002). Therefore, in the long run, if the negative AO regime continues, the Arctic will see not warming (as one may infer from the NSIDC report), but significant cooling.

Now, an important question is whether this cold winter in Europe marks the beginning of a new cold climate regime, or it is simply a short-lived anomaly, as, for example, the winter of 2003. The regime shift analysis applied to the time series of mean winter (DJF) temperatures in Berlin shows that the period 1988-2008 represents a distinct regime of much milder winters than anytime from the beginning of the record (Fig. 6). Unfortunately, the prominent regime shift in the late 1980s in Europe have not received as much attention in the climate literature as the shift in the late 1970s in the Pacific. Although it is too early to draw a definitive conclusion, it is possible that the warm climate regime in Europe has come to an end.

References