G z

sheet. Cold conditions with minimal melting and the continual accumulation of snow and superimposed ice during the LGM would have led to the build-up of extremely thick sea ice (tens of meters in thickness, like the ice shelves of northern Ellesmere Island today) which was essentially locked in place within this restricted basin. This composite of snow, superimposed ice, and sea ice would have remained more or less immobile until rising sea level, break-up of the Barents Sea ice sheet, and the return of warmer sub-surface waters to the Arctic Ocean induced some disturbance in (and heat flux to) the floating ice mass. The final trigger for a massive discharge of very thick, low salinity ice into the Atlantic Ocean may have come about when sea level rose to the point where Bering Strait was flooded (at -40 to -50 m) causing the Trans-Polar Current to become established. Recent 14C dates on terrestrial peats from the Chukchi Shelf, north of Bering Strait (which were flooded by the transgressing sea) place the time of this crucial event at around 11,000 14C yr B.P. (Elias et al., 1996b). This suggests a direct link to the onset of Younger Dryas conditions in western Europe as a result of the North Atlantic being flooded, not by icebergs from the Laurentide Ice Sheet, but by Arctic Ocean sea ice, which then reduced, or displaced, NADW production for the next 1000 yr.

Although the exact sequence of events that took place in late Glacial and early Holocene time is not yet entirely understood, it is clear that meltwater entering the North Atlantic is implicated as a critical factor influencing climatic fluctuations at that time (Manabe and Stouffer, 1995; Bard et al., 1996). Very abrupt changes in thermohaline circulation took place, involving reductions in deepwater formation, or shifts in the pattern of deepwater production, with compensatory reductions in the flux of warm surface waters to the North Atlantic. Nevertheless, several enigmas remain. Perhaps because of 14C dating problems, the timing of observed changes does not always fit together nicely, and sometimes the geochemical evidence is contradictory. Distinguishing deepwater produced in different areas is difficult, compounded by the fact that deepwater production regions probably shifted over time (Duplessy et al., 1980, 1988). The timing and magnitude of meltwater discharges from the Arctic and Fennoscandia, and from both the northern and southern drainage systems of the Laurentide ice sheet remain problematic. Indeed, Clark et al. (1996) believe that meltwater from the Antarctic ice sheet must be implicated in the late glacial oceanographic changes of the North Atlantic; they argue that the estimated volumes of meltwater from the Laurentide ice sheet and the timing of discharge events (based on what is known from the geological record on land) do not fit with the various explanations so far proposed for changes in North Atlantic circulation. Further high resolution sediment studies will no doubt help to resolve many of these issues. Meanwhile, the important question to ask today is whether, in the absence of ice major sheets, changes in the North Atlantic salinity structure can be brought about by imbalances in the precipitation-evaporation-runoff relationship. This issue is especially relevant as greenhouse gases increase to unprecedented levels, with a possible consequence being higher precipitation and runoff into the North Atlantic Basin or a drastic reduction in Arctic Ocean ice cover. If the thermohaline circulation were to be disrupted by such changes, there may indeed be "unpleasant surprises in the greenhouse" (Broecker, 1987).


It is clear from ice-core records that carbon dioxide levels in the atmosphere (pC02) were considerably lower (by 90-100 p.p.m.v.) at the Last Glacial Maximum than in the Holocene (see Section 5.4.3). Indeed, long-term carbon dioxide changes parallel 8lsO in ice and must have played a role in glacial-interglacial climatic changes, either directly or indirectly. How did such changes come about? Because the ocean carbon content is 50-60 times that of the atmosphere, even relatively small changes in the rates of ocean carbon dioxide uptake or loss (de-gassing) can have large effects on atmospheric C02 levels (Broecker, 1982). Changes in sea-surface temperature and salinity in glacial times (to a colder, more saline ocean) can account for -10% of the observed pC02 changes (simply by the difference in COz solubility under such conditions) but the bulk of the change must be related to an increase in biological productivity in the ocean during glacial periods. The C02 dissolved at the ocean surface is removed by biological activity in the photic zone, which forms the organic tissue and carbonate shells of marine organisms. As these die and fall through the water column, the carbon fixed near the surface is transferred to deeper layers of the ocean and may be deposited in the sediments. This process can be thought of as an "ocean carbon pump" whereby carbon is continually removed from the ocean surface (Volk and Hoffert, 1985); upwelling of carbon-rich deepwa-ter returns C02 to the atmosphere. Thus, some areas of the ocean are carbon sinks and some are carbon sources for the atmosphere. Factors that alter the biological productivity of surface waters and the rate of upwelling, or which change the distribution of sources and sinks of carbon, can therefore have a significant effect on atmospheric C02 levels (Ennever and McElroy, 1985).

An important index of photosynthetic activity in the surface waters of the ocean is the relative proportion of 12C to 13C. During photosynthesis, 12C is preferentially removed from the water to produce organic material with low 813C. In many regions of the ocean, productivity is limited by a lack of nutrients (particularly phosphate and nitrate). Broecker (1982) suggested that during glacial periods, when sea level was lower, phosphates that had accumulated on the continental shelves would be eroded and dispersed into the ocean, increasing nutrient levels and hence productivity in the near-surface waters. This would lead to 12C removal from near the surface and cause the 813C gradient between the upper and deep ocean to increase. At the same time, higher productivity in the photic zone would increase the rate of C02 removal from the ocean, causing an increase in the C02 flux to the atmosphere, thereby lowering pC02. Hence, there should be evidence of lower atmospheric C02 levels preserved in the skeletal remains of organisms that lived near the surface, compared to those that lived in the deep ocean. Shackleton et al. (1983) examined this question by measuring the difference between 13C in the carbonate tests of planktonic and benthic forams in an equatorial Pacific core (Fig. 6.54a). The resulting record of the 813C gradient between near surface and deep waters (A813C) provides a proxy of atmospheric C02 levels, with stronger gradients signifying increased productivity at the surface and hence lower pC02. This record is remarkably similar to that obtained from the Vostok ice core (Fig. 6.54b), suggesting

50 100 150 200 250

0 0

Post a comment