The Younger Dryas

Figure 10.7 illustrates the YD in terms of high calcium concentrations recorded in the GISP2 ice core. The YD was also associated with high dust concentrations and high magnesium levels. These features imply an intensified circulation over continental regions and increased aridity. In addition, ammonium, nitrate, sulfate, sodium, chloride and potassium concentrations rose (Mayewski et al., 1993). These large increases cannot be explained by increased windiness alone. Destruction of biomass due to colder conditions during the YD could have caused the observed increase in ammonium.

It appears that NADW formation decreased, associated with cooling in the North Atlantic and equatorward retreat of the ocean Polar Front. SSTs decreased rapidly in the GIN seas by about 6 °C, returning to near LGM values (Rochon etal., 1998). However, a sea ice-free corridor persisted along the Norwegian coast. Some of the reconstructed oceanic changes in the GIN seas during the YD are summarized in Figure 10.8. Dates

Figure 10.7 Calcium concentrations (ppb) covering the period 10-20 ka based on GISP2 ice core data. The sample resolution is approximately 2 years through the Holocene, a mean of 3.48 years within the YD (Younger Dryas) and BA (B0lling-Aller0d), and 3-15 years during the OD (Older Dryas) (from Mayewski et al., 1993, by permission of AAAS).

Artemisia (%) Temperature (°C) Sea-ice cover Proposed correlations

(montns/yrj lala-glacial chronozonoE Mangerud etsl (1974)

Artemisia (%) Temperature (°C) Sea-ice cover Proposed correlations

(montns/yrj lala-glacial chronozonoE Mangerud etsl (1974)

Figure 10.8 Inferred changes in Artemisia, sea surface temperature and sea ice cover in the North Sea during deglaciation (from Rochon et al., 1998, by permission of Elsevier).

are 14C years before present, which appear younger than calendar dates (see Section 10.2.1). The reductions in August SST and the increase in the number of months with sea ice cover are especially notable.

Some high-latitude areas experienced a more extreme YD event than others. The YD was strongly expressed in the North Atlantic and in the Nordic Seas. Western Norway experienced a very severe change, seen by a considerable halt in disintegration of the ice sheet and the development of extensive moraines. In the western Nordic seas, there was a slight increase in IRD input (Hebbeln et al., 1998). This compares with little or no change in Svalbard (Birks et al., 1994). A cold dry climate in northwest Russia is inferred from pollen records which indicate that steppe/tundra vegetation was present (Subetto et al., 2002). The YD was not well expressed in the Southern Hemisphere and most of these records point to warm conditions (Siegert, 2001).

Figure 10.7 Calcium concentrations (ppb) covering the period 10-20 ka based on GISP2 ice core data. The sample resolution is approximately 2 years through the Holocene, a mean of 3.48 years within the YD (Younger Dryas) and BA (B0lling-Aller0d), and 3-15 years during the OD (Older Dryas) (from Mayewski et al., 1993, by permission of AAAS).

As mentioned, the transition from the cold climate of the YD to the Holocene occurred very quickly (Dansgaard et al., 1989; Severinghaus et al., 1998). Warming of 7 °C occurred in less than 50 years over Greenland. The Greenland climate also became more humid, most likely due to the reduction in sea ice cover, allowing more evaporation and increased precipitation falling as snow (Alley et al., 1993). Dust concentrations in the GRIP ice core decreased by a factor of three within a 20-year period indicating that storminess rapidly decreased.

While, like other D-O events, the cause of the YD is still uncertain, evidence again implicates changes in NADW production. One theory involves a large influx of meltwater through the St. Lawrence Seaway and into the North Atlantic. This resulted from the re-routing of drainage from Lake Agassiz, a proglacial lake which formed at the foot of the retreating Laurentide ice sheet. The lake had been draining into the Mississippi basin. However, just before the start of the YD, it appears that there was a catastrophic re-routing into the St. Lawrence Seaway, associated with a 40-m drop in the lake level. Drainage was then routed back to the Mississippi (Broecker et al., 1989) (Figure 10.9). As reviewed by Bradley (1999), a number of problems became evident with this view, including indications that the YD occurred at a time when discharge to the world ocean was less than either the preceding or following 1000 years, and that the meltwater flux from the St. Lawrence was actually reduced during the YD episode. In a subsequent study, Broecker (1990) argued that the period of rapid sea level rise prior to the YD lowered salinity in the surface layer so much that the North Atlantic was already predisposed for a shut down or major cessation in NADW production. Fanning and Weaver (1997) offer modeling support for this view.

Lehman and Keigwin (1992) give a different perspective. Their study addressed the YD as part of a focus on sudden changes in the North Atlantic circulation during the deglaciation phase. They show that prior to the YD, episodes of freshening, associated with meltwater inputs from the decaying Fennoscandian Ice Sheet, immediately preceded faunal evidence for coolings associated with weakenings of the THC. The THC then recovered. The cause and effect is hence reminiscent of the salt-oscillator model. The more recent of these melting events (11.5-11.2 ka 14C, ~ 13.4-13.0 ka calendar) may have triggered the much longer and more severe YD. They suggest a link between freshening beginning about 50 years after the onset of the YD and initial discharge of stored meltwater from the so-called Baltic Ice Lake. This may have prolonged and intensified suppression of NADW production during the YD. They find another interval of freshening after about 10.4 ka 14C (~12.6 ka calendar) associated with final discharge of the Baltic Ice Lake. The major point is that warming and associated enhanced meltwater input promoted weakening of the THC and cooling, followed by recovery. The intensity of the YD event may have been in part fostered by additional freshwater inputs provided by drainage of the Baltic Ice Lake.

However, as outlined by Lehman and Keigwin (1992), a complete focus on the Fennoscandian Ice Sheet seems insufficient - the Laurentide Ice Sheet must also be involved. It appears that, while the melting rates of both the Fennoscandian and Laurentide ice sheets may have been influenced by changes in the THC, their individual

Laurentide Ice Sheet

Figure 10.9 The Laurentide Ice Sheet and the routing of overflow from the Lake Aggasiz basin (dashed line) to the Gulf of Mexico just before the Younger Dryas (a) and routing of overflow from Lake Aggasiz through the Great Lakes to the St. Lawrence and northern North Atlantic during the Younger Dryas (b) (from Broecker et al., 1989, by permission of Nature).

Figure 10.9 The Laurentide Ice Sheet and the routing of overflow from the Lake Aggasiz basin (dashed line) to the Gulf of Mexico just before the Younger Dryas (a) and routing of overflow from Lake Aggasiz through the Great Lakes to the St. Lawrence and northern North Atlantic during the Younger Dryas (b) (from Broecker et al., 1989, by permission of Nature).

meltwater contributions may have had very different effects on the THC. Meltwater from the Laurentide Ice Sheet (e.g., from the drainage of Lake Agassiz) must travel farther to reach areas of convection in the GIN seas, so a greater or prolonged melting of the Laurentide Ice Sheet may have been needed for an equivalent damping effect on the circulation. It may be that the Laurentide Ice Sheet meltwater built up in the Atlantic, leading to a non-linear response of NADW production to ice sheet melting. This may account for the increasingly severe reductions in the THC during the deglaciation period that culminated in the millennial-scale YD.

Yet another idea involves sea ice. During the LGM, when sea level was much lower, the Arctic Ocean was very isolated, and Fram Strait was narrower, restricting sea ice transport into the North Atlantic. Very thick sea ice would have built up in the Arctic Ocean, remaining largely immobile until rising sea level, breakup of the Barents Sea Ice Sheet and the return of warmer Atlantic waters to the Arctic Ocean allowed for massive discharge of sea ice into the North Atlantic. A final trigger may have come when sea level rose sufficiently that the Bering land bridge was flooded, allowing throughflow from the Pacific to the Atlantic, flushing out the ice (Bradley, 1999). Dating of terrestrial peats on the Chukchi shelf place this crucial event at about 11 ka 14C (~13 ka calendar), just before the YD (Elias etal., 1996). The processes are hence reminiscent of those that led to the Great Salinity Anomaly.

0 0

Post a comment