New sediment data on the polar oceans during the last ice age

As part of a long term collaboration with the group of Professor Gerald Haug at ETH Zurich, Daniel Sigman and colleagues continue to pursue evidence for reduced vertical exchange (i.e. “stratification”) in polar ocean regions under colder climates of the past 3 million years. A major motivation for this focus is that the reconstructed changes have the capacity to affect atmospheric CO2 during ice ages, reducing the natural leak of biologically sequestered CO2 out of the polar ocean and into the atmosphere. Moreover, this work bears on the expected response of the different polar ocean regions to anthropogenic climate change, with implications for regional weather, fisheries, and the oceanic sink of anthropogenic carbon.

In 2010, work by recent graduate student Brigitte Brunelle provided the first regionally complete view of Subarctic North Pacific conditions over the last two glacial cycles based on diatom-bound nitrogen isotope and productivity proxy data. These results confirm her earlier work focused on the Bering Sea indicating reduced nutrient supply from below during the last ice age, and demonstrate that this behavior applied over the entire western Subarctic North Pacific. Brunelle’s data also highlight a mysterious deglacial change in the North Pacific that appears to coincide with changes in the North Atlantic. Members of the Sigman Group are currently seeking to understand this apparent North Atlantic/North Pacific connection, which may provide mechanistic insight into the global cascade of events that ended the last ice age.

Sigman also published a review article with Mathias Hain, and Gerald Haug on the status of polar ocean hypotheses to explain glacial/interglacial changes in atmospheric CO2 (Sigman et al., 2010). From geochemical and biogeochemical perspectives, the community working on this problem now has a set of fully plausible hypotheses for glacial/interglacial CO2 change, and new data should be able to address them in a more or less direct fashion. The conceptual challenge is that the currently favored hypotheses do not seem consistent with the physical models of climate that are being used to predict the Earth system’s changes under global warming. While this is disturbing, it highlights the centrality of the ice ages in our understanding of climate, including its anthropogenic future.


Mechanisms of glacial CO2 drawdown

The reduction of atmospheric CO2 during the last ice age is believed to be due to the increased storage of CO2 in the deep ocean that was respired from sinking organic particles. The increased storage, in turn, has been attributed to a slowdown in the rate at which deep water with respired CO2 was brought back up to the surface (i.e. by a reduced ventilation rate for the deep ocean). Whether a reduced ventilation rate for the deep water leads to a large reduction of ~90 ppm in atmospheric pCO2 is still under debate.

In 2010, the Sigman & Sarmiento Groups both investigated the impact of ocean circulation on atmospheric CO2, using box and 3-D global ocean models of biogeochemistry models, respectively. Both groups showed that the response of atmospheric pCO2 to reduced ventilation is rather modest, because a substantial portion of the change in atmospheric pCO2 caused by the accumulation of respired carbon is offset by a concomitant accumulation of alkalinity in deep waters and out of contact with the atmosphere.

In the box model, stratification, nutrient drawdown, and sea ice cover (within bounds set by observations in the Antarctic) lowered atmospheric CO2 by only 35 ppm. Total drawdown reached 65 ppm when shoaling of North Atlantic overturning was considered. Adding the effects of increased nutrient consumption in the sub-Antarctic caused as much as an additional 35 ppm CO2 drawdown, achieving more than the 90 ppm glacial-interglacial change.

Among the suite of 3-D circulation models examined in a study by Kwon and colleagues, the largest reduction in atmospheric pCO2 of 44-88 ppm occurred in a model where reduced overturning rates of southern and northern sourced deep waters resulted in a four-fold increase in the Southern Ocean deep water ventilation age. The large uncertainty of a factor of two in the 3-D studies arises due to the uncertainty regarding how surface productivity responds to climate and circulation change. The same 3-D model also produced a large reduction in the calcite saturation state in the Atlantic bottom water, which would likely drive additional drawdown of atmospheric pCO2 due to dissolution of CaCO3 sediments (Figure 26).

Both studies support the view that a globally synchronous change in the ventilation rate of deep waters, changes in productivity in the Southern Ocean, and associated feedback from the CaCO3 system could produce the large fluctuations in atmospheric pCO2 during the glacial-interglacial climate cycles.



Figure 26. Organisms growing at the surface of the ocean export organic carbon and calcium carbonate to the deep ocean where most of the organic carbon is eaten by organisms and returned to the water as respired carbon dioxide, and most of the calcium carbonate dissolves. On average, the ratio of calcium carbonate to organic carbon in the material exported from the surface is less than 10%, but most of the organic carbon decomposition occurs very shallow in the water column, whereas most of the calcium carbonate dissolution occurs quite deep so that the observationally based estimate of the ratio shown in the figure increases to 50% with increasing depth. The relative behavior of the cycling of organic matter and calcium carbonate has a major impact on the air‐sea balance of carbon dioxide as the respiration of organic carbon and the dissolution of calcium carbonate have opposing effects on atmospheric carbon dioxide levels. Kwon et al. (in prep.) derived a simple analytical framework in which atmospheric carbon dioxide concentration is related to the oceanic storage of respired carbon and dissolved carbonate. Using this analytical framework along with a 3‐D global ocean biogeochemistry model, Kwon et al. (in prep.) examined the control of atmospheric CO2 concentration by the ocean ventilation change with the focus of the effect of the redistribution of respired carbon and dissolved carbonate. The figures above present the ratio of dissolved carbonate to respired carbon along the Atlantic (30°W, left) and Pacific (180°E, right). The ratio is calculated from the data of WOA05 [Garcia et al. 2006] and GLODAP [Key et al. 2004]


Ice core studies of greenhouse gas concentrations prior to 800,000 years ago

One of the most important lines of evidence for the link between greenhouse gases and global temperature is the close covariation between global temperature and atmospheric CO2 concentration over the past 800 ka. The CO2 records come from trapped gases in ancient ice sampled by coring the 3-km thick ice sheet in East Antarctica. The timespan of the records is limited by the age of ice in the East Antarctic Ice Sheet.

The Bender Group has undertaken two collaborative projects searching for much older ice, in two regions, Mullins Valley, and the Allan Hills. Both are near the coast of Antarctica, south of New Zealand. At both sites, there is some equivocal evidence that ice is present dating back to 2.5-9 Ma. Their earlier studies at one of these sites revealed an ice age of about 1 Ma, which was not as old as had been expected. It is unclear if this relatively young age was an artifact of contamination of shallow ice samples with modern air.

Work of the past year shows that, at one of these sites (Mullins Valley, Dry Valleys region of Antarctica) trapped gases in the ice do not date to earlier than 1 Ma (1,000,000 years before present). It is not known whether this unexpectedly young age represents the true age of the ice or contamination of trapped gases by the modern atmosphere. The Allan Hills site (Figure 27) gives even younger ages, about 0.5 Ma. At this site, however, there is direct evidence for the presence of contamination by modern air throughout the 15-25 m depths of the ice cores. Contamination is known to occur, but was previously found only in the top 5 m or so of the ice. To deal with this problem, the team returned this past year and drilled 2 deeper cores in the region, on going to 125 m and the other to 230 m depth. Dating of these cores will reveal the antiquity of the ice.

Figure 27. Preparing to drill for old ice in the Allan Hills, Antarctic. The “blue ice” (foreground) is ancient ice brought up to the surface by the flow of the ice sheet.