The North Atlantic Oscillation

Much of the analysis in this chapter and the next will focus on events in the Middle East and Europe. How the climate of these parts of the world is affected by global circulation patterns is central to the discussion. In this context, it turns out that the features of the North

Atlantic, which have been so much part of the book up to now, continue to dominate our thinking. In particular, behaviour of the winter circulation that is known as the North Atlantic Oscillation (Marshall et al., 2001) is central to this discussion, notably when interpreting events in the Middle East.

Why should this be so? The reason is that the climate of the Middle East is characterised by cool, wet winters and hot, dry summers. Most of the Middle East lacks access to significant surface and groundwater resources because local evaporation far exceeds precipitation. The one important exception is Turkey, which has abundant precipitation resulting from the orographic capture of winter rainfall from eastward-propagating mid-latitude storms generated in the Atlantic Ocean and the eastern Mediterranean Sea. This is vital to the life-giving qualities of the greatest of all the major river systems of the Middle East: the Tigris-Euphrates.

In modern times the flow of the Tigris-Euphrates has two primary flooding periods (Cullen et al., 2002). The first is rainfall-driven run-off from December through March, regulated on interannual to decadal timescales by the NAO as reflected in local precipitation and temperature. The second period, from April to June, reflects spring snowmelt and contributes over half of annual run-off. At the most basic level, this is related to the amount of snow that falls during the winter and the weather conditions during the thaw.

The importance of the NAO is that it is intimately linked to the climate of Greenland and northern Europe. Throughout the preceding chapters reliance on the detailed records of changes in the weather in Greenland have been used to make inferences about changes in other parts of the northern hemisphere. Because we were dealing with relatively large changes in the climate, it was often relatively easy to find examples of proxy records that confirmed that events recorded in either Greenland or in North Atlantic sediments were identifiable across much of Eurasia. As the changes become smaller, this correspondence is more difficult to establish. So we now need to look at more subtle features of Holocene weather patterns in the northern hemisphere. In this respect, the NAO is a particular useful measure of the circulation, especially in winter.

Since the eighteenth century it has been known that when winters are unusually warm in western Greenland, they are severe in northern Europe and vice versa (Van Loon & Rogers, 1978). As the missionary Hans Egede Saabye observed in a diary he kept during the period 1770-78: 'In Greenland all winters are severe, yet they are not alike. The Danes have noted that when the winter in Denmark was severe, as we perceive it, the winter in Greenland in its manner was mild, and conversely.' This seesaw behaviour was quantified by Sir Gilbert Walker in the 1920s in terms of pressure differences between Iceland and southern Europe and defined as the NAO (Walker, 1928).

The NAO shifts between a deep depression near Iceland and high pressure around the Azores, which produces strong westerly winds, and the reverse pattern with much weaker circulation. The strong westerly pattern pushes mild air across Europe and into Russia, while pulling cold air southwards over western Greenland. This flow also pulls cold air down into the eastern Mediterranean and Middle East producing colder, wetter winters than normal. The strong westerly flow also tends to bring mild winters to most of North America. One significant climatic effect is the reduction of snow cover, not only during the winter, but also well into the spring. The reverse meandering pattern often features a blocking anticyclone over Iceland or Scandinavia, which pulls arctic air down into Europe, with mild air being funnelled up towards Greenland. This produces much more extensive continental snow cover, which reinforces the cold weather in Scandinavia and eastern Europe, and this often extends well into spring while the abnormal snow remains in place.

Changes in sea-ice cover in both the Labrador and Greenland Seas as well as over the Arctic appear to be well correlated with the NAO, and the relationship between the sea-level pressure and ice anomaly fields suggests that atmospheric circulation patterns force the sea ice variations (Deser, Walsh & Timlin, 1999). Feedbacks or other influences of winter ice anomalies on the atmosphere have been more difficult to detect. But the frequency of depressions appears to have increased and atmospheric pressure decreased where ice margins have retreated, although these changes differ from those directly associated with the NAO. It may even be possible that a period of one phase of the oscillation produces the right combination of patterns of sea surface temperatures and deep-water production eventually to switch it into the opposite phase.

Understanding the NAO is central to interpreting modern climatic events, because of the influence it exerts on average temperatures in the northern hemisphere. Of all seasons, winters show the greatest variance, and so annual temperatures tend to be heavily influenced by whether the winter was very mild or very cold. When the NAO is in its strong westerly phase, its benign impact over much of northern Eurasia and North America outweighs the cooling around Greenland, and this shows up in the annual figures. A significant part of the global warming since the mid-1980s has thus been associated with the very mild winters in the northern hemisphere (Hurrell, 1995). Indeed, since 1935 the NAO on its own can explain nearly a third of the variance in winter temperatures for the latitudes 20° to 90° N.

These modern observations can help us exploit the evidence of climate fluctuations in Greenland during the Holocene. What the NAO shows is that there are close links between events all around the northern hemisphere even when the climate is in a more quiescent state. So it is possible to extend our thinking beyond the transparently global fluctuations of the last ice age when it is safe to assume that as the climate changed rapidly in Greenland, Eurasia and North America were having their fair share of climatic upheaval. In the Holocene, when the connections are subtler, the lessons of the NAO are more important as a strong westerly circulation over the North Atlantic will lead to a consistent pattern of above and below normal precipitation and temperatures around the northern hemisphere. Conversely, weaker circulation will tend to produce the reverse pattern. The value of this correlation is that it enables us to use sometimes -fragmentary data from around the northern hemisphere to build up a more coherent picture of regional climate change.

As was noted in Chapter 2, there have been at least four global periods of rapid climate change during the Holocene. The first of these was the period around 9 to 8 kya, with a notable cold spell around 8.2 kya in the North Atlantic. Then there was widespread rapid change between 6 and 5 kya, plus periods between 3.5 and 2.5 kya and since 0.6 kya. These more turbulent periods show up in the form of stronger mid-latitude circulation in the northern hemisphere, expansion of mountain glaciers around the world and greater ice formation in the northern North Atlantic, as seen in the amount of rafted debris found in ocean sediments.

The modern observations of the NAO show that when the circulation is in a strong westerly phase, winters are warmer and wetter over Iceland, the British Isles and Scandinavia. Farther south there is a reduction in rainfall in a band from the Azores to the Black Sea, and colder winters in the Eastern Mediterranean. In summer, northern Europe is cooler and wetter, which explains glacier expansion in the Alps and Scandinavia at such times. Conversely, when the circulation is more meridional, the pattern of temperature and rainfall anomalies reverses. In particular, there is increased winter rainfall over the Iberian Peninsula, Italy, the Balkans and Anatolia. This means that the periods of more intense circulation in the North Atlantic during the Holocene would have made life more difficult for Neolithic communities north of the Alps. On the other hand, those living around the Mediterranean and in Anatolia and the Middle East may have benefited from the increased winter rainfall.

This use of the NAO to interpret events during the Holocene has to be qualified by noting that, even when the climate had warmed up dramatically, it took a long time for the system to settle into what we would regard as recognisably modern climatic patterns. This may be of particular relevance to the 200-year cold spell that struck around 8.2 kya. This event was more nearly a throwback to the last ice age for two reasons. First, the remnants of the great ice sheets, most noticeably over Canada, still exerted a considerable influence on hemispheric weather patterns. Second, it is generally accepted that this event was a consequence of the massive release of meltwaters from Lake Agassiz via the Hudson Bay (Barber et al., 1999). This probably disrupted the thermohaline circulation of the North Atlantic for the duration of the event. So the conditions at this time were sufficiently different from those in recent centuries to require considerable care in using the modern NAO analogue to interpret circulation patterns at 8.2 kya.

The shifts in the climate around 4.2 kya are also difficult to interpret. In this instance, the palaeoclimatic data are equivocal. The Greenland ice cores and North Atlantic sediment measurements suggest that around this time there was a short and dramatic decline in the extent of sea ice. This appears to have been linked to a period of weaker circulation in mid-latitudes, except across North America, and in Europe glaciers either retreated or remained stable. There was also widespread drought in equatorial Africa and the Middle East. Although this event was not as extensive as others during the Holocene, it was sufficient to have a significant impact on early civilisations.

5.2 EUROPE, THE MIDDLE EAST AND NORTH AFRICA The changes in the climate across northern Europe after the LGM reflected closely the records presented in Chapter 2 for Greenland and the North Atlantic. So the succession of the Bolling, Older Dryas and Allerad, followed by the clear setback of the Younger Dryas, can all clearly be seen in the proxy records of climate change across the continent (see Fig. 2.8). Analyses of beetle assemblages (Coope et al., 1998) from Ireland in the west to Finland and Poland in the east have confirmed the earlier pollen analyses and provided estimates of July temperature levels. As the dying throes of the ice age swung the temperature and precipitation patterns north and south, the flora and fauna responded first to the relaxation of the glacial conditions and then had to retreat in the face of renewed cold.

During the Younger Dryas there was a temporary disappearance of the woodland cover that had previously extended over much of Europe and a replacement by dry steppe and steppe-tundra. Across northwestern Europe, the conditions may have been less severe, with forest-steppe (a mixture of patches of trees and grassland) being widespread. Indeed there may have been forest tundra, intermingled with some steppe elements, across most of Poland and Germany, but close to the Fennoscardian ice sheet there was shrub tundra.

In southeastern Europe and the Levant, with the warming after the LGM, precipitation rose to a peak around 13.5 kya. Thereafter, the aridity during the Younger Dryas may have been much more severe than farther north. Pollen analysis suggests that in many areas of Greece and across Turkey, it was even more arid than during the most extreme part of the LGM. Annual precipitation may have been less than 150 mm across much of lowland Greece, and this aridity extended into northwest Syria, Turkey and the western Zagros mountains of Iran, but Northern Israel was perhaps less arid than other parts of the region. Temperatures were also markedly lower.

When the Younger Dryas ended, the change for northern Europe was dramatic, even allowing for the occasional hiccup like the Preboreal Oscillation. In the British Isles the combined influence of the ice sheet over Scotland and the ice cover over the North Atlantic to the west had led to particularly low temperatures. Here, annual temperatures rose by about 15 °C, with a midsummer rise of about 5 °C, and in midwinter by over 20 °C. Comparably large changes occurred across most of northern Europe, although the places closest to the ice sheets over northern Britain and Scandinavia remained under the baleful influence of their icy neighbours.

The succession of changes in the Holocene climate in Europe was initially identified in terms of pollen data (see Section 2.6), which apply to most of northern Europe. This sequence recognises five general climate periods during the Holocene: the Preboreal and Boreal (11.5 to 9 kya, a rapid transition period followed by a warm and dry period that reflected the more continental nature of the early Holocene in Europe) with summers warmer than now but colder winters than at present;the Atlantic (9 to 6 kya, a warm and wet period);the Sub-Boreal (6 to 2.5 kya, a warm and dry period) and the Sub-Atlantic (2.5 kya to the present, a cool and wet period).

A recent and much more comprehensive analysis of pollen data examined well over 2000 records from North Africa and Europe west of the Urals (Davis et al., 2003). This has produced estimates of changes in temperature for six regions of Europe (covering the east and west halves of the continent for northerly, central and southern latitude zones) since 12 kya. This work confirms the broad features of the earlier analysis described above for the northern and central part of the continent. The interesting clarification of these results is that they show that most of the changes in temperature, notably the warming in the mid-Holocene, occurred in the summer. By comparison winter temperatures have changed little since 8 kya, with western Europe, if anything, showing a continuing warming trend up until recent times.

The analysis does, however, give a different picture for southern Europe. The striking feature is how much the western Mediterranean has warmed since around 8 kya. Here the temperature has risen by about 2 °C in both summer and winter. In the eastern Mediterranean the rise has been much less and largely restricted to winter. More interesting is the decline in temperature here in the period from 11 to 8 kya during which the winter temperature dropped by over 4 °C and in summer by nearly 2 °C, in both cases from levels some 2 °C above current values. In the western Mediterranean the drop in temperature was restricted to summertime, and was less dramatic, falling from around 0.5 °C below current values at 10 kya to 2 °C below current values at 8 kya. The explanation for these changes was the onset of markedly wetter conditions throughout the Mediterranean in the early Holocene.

To identify the impact of specific events, such as the cooling at 8.2 kya, we have to turn to other records, as their impact is not easily picked out in the pollen records. In Greenland it led to a cooling of about 2.7 °C. Analysis of the isotope ratio of the shells of small crustaceans (ostracod valves) preserved in the sediment of Ammersee, a small deep lake in southern Germany (see Fig. 2.8), suggests a cooling of about 1.7 °C in the annual air temperature at the time (Von Grafenstein et al., 1998). Studies in Lake Annecy, in nearby eastern France, indicate an increase in precipitation at the time (Magny et al., 2003). This work also concludes on the basis of these data, together with other hydrological records produced in Europe for the same period, that mid-latitudes between around 50 and 43° N experienced wetter conditions during this cooling episode, whereas northern and southern Europe and the Mediterranean had a drier climate. These wetter conditions appear to have extended eastwards as far as Lake Van in eastern Anatolia (Wick, Lemcke & Sturm, 2003).

The decline in rainfall in the Mediterranean appears to have extended to the Arabian Sea. Ocean-sediment records here show a marked decline in the flow of freshwater from the Indus River (Staubwasser et al., 2003), while speleothem records from Oman confirm that the summer monsoon weakened when temperatures were low over the North Atlantic (Fleitmann et al., 2003). In equatorial Africa records of lake levels show a sharp decline suggesting a weakening of the monsoon rains to the north of the Equator (Stager & Mayewski 1997; Gasse, 2000). So it is reasonable to assume that when looking for evidence of climatic-induced change in human affairs, any events that coincide with 8.2 kya deserve close scrutiny. But, as noted in Section 5.1, we need to be careful about using present NAO behaviour to interpret its global impact.

Cores from the eastern Mediterranean and Aegean Sea provide additional information (Rohling, 2002). There are three markedly cooler interludes centred on 8.2, 6.2 and 3.2 kya. Rainfall levels rose markedly at the beginning of the Holocene (10.5 to 9.5 kya) and then drifted back down to levels more typical of modern times by around 5 kya. Possibly more intriguing are the overall changes that appear to have altered the circulation of the Mediterranean around 9.5 kya. Starting around this time oxygen-starvation (anoxia) wiped out the entire deep ecosystem of the eastern Mediterranean. This condition prevailed for some 3.5 kyr. Environmental evidence of such anoxic episodes is found in dark, olive green to pitch black, organic-rich sedimentary layers (sapropels). What this shows is that the amount of freshwater from both rainfall and river run-off increased substantially, reducing the salinity (in the eastern Mediterranean the water is currently very saline and sufficiently dense to sink to great depths). The less dense layer formed a cap and switched off the formation of oxygenated deep water.

Although this episode was a catastrophe for the deep ecosystem, for people living in parts of the region it was a peculiarly benign climate. The extension of monsoon rains from the south covered much of the Sahara with savannah vegetation that suited nomadic herdsmen. In the Nile valley the effect was for the much greater stream flow to scour deep into the thick sediment, which had formed during the period since the end of the LGM, making the valley less habitable.

Was this article helpful?

0 0

Post a comment