Daniel Sigman’s and Jorge Sarmiento’s groups continue to develop new models and tools for studying ancient changes in carbon fluxes and climate. Data from sediment cores and results of computer simulations have continued to strengthen the evidence that the Southern Ocean, the continuous band of ocean surrounding the Antarctic continent, plays an important role in glacial-interglacial CO2 changes


Model development

Eric Galbraith led the work in the Sarmiento group in the continued development of the coarse resolution coupled model CM2.1C (a coarser resolution version of GFDL’s climate model CM2.1). This work is motivated by interest in studying coupling between carbon and the physical components (ocean/atmosphere/land) of the climate system under climate change. The goal is to develop a model which runs an order of magnitude faster than the state-of-the-art generation of models developed at GFDL such as CM2.1, which was considered to be one of the premier coupled models presented in the most recent IPCC report. An order of magnitude increase in speed would facilitate the long simulation times needed for paleoclimate studies.

Work over the last year has focused on improving the skill of CM2.1C in its simulation of the preindustrial climate state. As a means of evaluating model skill, the root-mean-square error of sea surface temperature with respect to observed sea surface temperature is a commonly used diagnostic of the mean bias (area-weighted) in the model representation of surface temperatures. The RMS temperature error of CM2.1C is compared with a number of models that were considered as part of the IPCC report in Figure 10. CM2.1C is now an IPCC-class model with regard to its temperature and salinity errors, as well as other characteristics such as stable ocean overturning, the global surface winds, as well as a respectable representation of El Niño variability.

Figure 10. A comparison is shown between the Root Mean Square (RMS) sea surface temperature error for IPCC simulations. The models shown in the chart include the full-resolution CM2.1 model (the last model shown) from GFDL and the coarse-resolution version of CM2.1 (CM2.1C). The first five bars in the chart indicate different stages of improvement in the representation of physical processes for CM2.1C, with the second (CM2.1C 1860 a500) representing the control run that is now an IPCC-class model. The models from other institutes include those of the Goddard Institute for Space Studies (GISS), the UK Met Office (UKMO), and the National Center for Atmospheric Research.

In a related modeling effort, Richard Slater has continued the development of the modules used for ocean tracers in the GFDL ocean models. As a test of model skill, a comparison was made between simulated and observed CFC-12 changes along P6 (32°S) in the Pacific Ocean. Both the observations and the model reveal a clear subsurface maximum in the CFC-12 change, establishing that the circulation in the coupled model is able to reproduce to first-order important structures and features in passive tracers in a region known to be significant for the global ocean uptake of anthropogenic carbon, as well as for paleoclimate changes.


Studying glacial-interglacial changes and the Southern Ocean

Daniel Sigman and his collaborators continue to pursue the evidence for reduced vertical exchange (i.e. “stratification”) in the halocline-bearing polar ocean regions under colder climates of the past 3 million years. A major motivation for this focus is that the reconstructed polar ocean changes have the capacity to affect atmospheric CO2 in the observed sense of its glacial/interglacial oscillation, by, during ice ages, reducing the natural leak of biologically Daniel Sigman and his collaborators continue to pursue the evidence for reduced sequestered CO2 out of the 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.

Work by graduate student Brigitte Brunelle on ice age conditions in the Bering Sea of the Subarctic North Pacific indicates reduced nutrient supply from below during the last ice age, strengthening the case for a bipolar (Antarctic and North Pacific) increase in stratification during ice ages. In related work by Eric Galbraith, it has been demonstrated that these polar ocean changes were indeed paralleled by reduced ventilation and increased storage of CO2 in the deep ocean of the last ice age, and Galbraith has produced new constraints on the timing with which this deep CO2 reservoir was dissipated at the end of the ice age, eventually accumulating in the atmosphere.

Based upon the measurement results, Sigman’s group has recently begun to consider physical mechanisms for the apparent climate/polar stratification link mentioned above. Former postdoc Agatha de Boer adapted the current GFDL ocean model (MOM4) for paleoclimate-scale simulations and used this platform to investigate the sensitivities of deep ocean ventilation to fundamental climate parameters, including the mean temperatures of the ocean and atmosphere and the strength of the southern hemisphere westerly winds. As described above, an underlying motivation was that both the Antarctic and the subarctic North Pacific were less efficient at ventilating the ocean interior under cold climate regimes of the geologic past, potentially explaining glacial/interglacial variations in atmospheric CO2. de Boer’s work has demonstrated that both ocean cooling in itself and cooling-induced changes in the westerly winds work to reduce Antarctic overturning, such that both have the potential to explain past ventilation changes (Figure 11). At the same time, negative feedbacks arise in the model, placing bounds on the ventilation rate of the ocean interior and its susceptibility to a cooling-driven decrease. These include the role of North Atlantic overturning in the energy balance of the deep ocean.

Figure 11. Sensitivity of the deep ocean ventilation age to ‘dynamic’ ocean temperature change in a version of the GFDL ocean model (after de Boer et al., 2007). In the experiments shown, the temperature input to the seawater density calculation step in the model is changed globally from the standard case. Ventilation age, the mean time since deep water has contacted the ocean surface, is plotted for the global ocean as well as for each of the major deep water formation regions. The experiments predict that a colder ocean than today would have reduced overturning in the Southern Ocean and North Pacific, leading to a globally “older” deep ocean that sequestered more CO2 from the atmosphere, but increased overturning in the North Atlantic.

In light of the confluences of de Boer’s modeling work with paleoceanographic measurements by Sigman’s group and other investigators, a new hypothesis has been posed to explain the most recent set of major deglaciations. From the glacial state of Antarctic and North Pacific stratification, orbital forcing and glacial melting in the Northern hemisphere are called upon to decrease the relative driving forces for overturning in the North Atlantic versus the other ocean basins. As North Atlantic overturning ceases, Antarctic overturning increases, releasing CO2 to the atmosphere and reducing the surface albedo associated with southern hemisphere sea ice. These processes drive global warming, which leads to large scale deglaciation. This hypothesis has many testable aspects and also provides a straightforward conceptual framework for ongoing investigation.


Impacts of climate change on ancient civilizations

Sigman’s group is also working to improve our understanding of the climate that directly preceded the era of large scale human alteration of the global environment. Sigman has collaborated with his close colleague Gerald Haug, Professor in the Geological Institute at ETH Zurich, on high resolution studies of marine and lake sediment cores to reconstruct climate changes in the northern hemisphere tropics over the Holocene, the last 10 thousand years during which the Earth has been in a warm (“interglacial”) climate state. Sediment records from a lake in southeast China and the Cariaco Basin off the coast of Venezuela indicate that at least some past migrations of the Intertropical Convergence Zone (ITCZ) appear to have stretched across the Pacific. Moreover, there is suggestive evidence that shifts in the ITCZ represented caused environmental changes that contributed to sociopolitical events in dynastic China and the Classical Maya of Central America.


Climate change and ocean productivity

Also in collaboration with the Haug group, Sigman’s group is working to understand the response of the ocean’s nitrogen cycle to climate change. As part of this effort, high-resolution sedimentary nitrogen (N) isotope records across two major deglaciations were generated from the Cariaco Basin, a low-oxygen basin with a remarkable well-preserved sedimentary sequence. The reconstruction appears to connect competing views on the controls of N2 fixation, suggesting that N2 fixation is tightly coupled to ocean N loss but is also sensitive to iron supply and so is stimulated by dust inputs. These results argue for some degree of N cycle response to changing dust input to the ocean, but little opportunity for the large N2 fixation-driven changes in ocean fertility that have been hypothesized as a part of plans to add iron to the tropical and subtropical ocean.