In order to improve our understanding of natural controls on atmospheric CO2, we are investigating why atmospheric CO2 varies with recent climate cycles, being lower during ice ages. Work conducted previously by CMI members and others suggests that the Southern Ocean, the continuous band of ocean surrounding the Antarctic continent, holds the answer.

Understanding the effect of atmospheric iron (i.e. dust) inputs to the Southern Ocean has arisen as a common goal among the research groups of Sarmiento, Bender and Sigman. The Bender group has deployed new methods aboard ships of opportunity to measure net carbon production by Southern Ocean biology. This process is related to CO2 export from the surface ocean, which reduces the atmospheric CO2 inventory. The results to date of their work indicate that the input of dust from arid regions of southern hemisphere continents to the Southern Ocean is a major factor enhancing net production. The results tend to support John Martin’s hypothesis that atmospheric CO2 was lower during the ice ages because lands were more arid, the dust flux was greater, and Southern Ocean net production was higher.

The Sarmiento group is carrying out model experiments to clarify the effects that iron fertilization of different regions of the Southern Ocean would have on the carbon cycle and on lower latitude ocean productivity. The Sigman group’s novel nitrogen isotope measurements on deep sea sediments provide evidence of iron fertilization during the last ice age in two regions of the Southern Ocean. The first is the Subantarctic Zone, which is in the path of the westerly winds carrying dust from South America, South Africa and Australia. The second is the Antarctic Zone, where sea ice melting can release iron in a summertime pulse ideal for stimulating algal growth.

In the coming years, Bender’s group plans to extend this work by increasing the geographic extent of their Southern Ocean carbon flux observations, by working with groups making ancillary measurements that inform us about causes of CO2 variability, and by making continuous measurements of diagnostic properties along cruise tracks (rather than periodic observations). The first objective of this work is to achieve a process-level understanding of the factors controlling net production in the Southern Ocean. The second goal is to gain an understanding of how to scale local measurements, using remotely sensed properties, to determine seasonal and areal variability in net production throughout the basin. These results will in turn allow a much more comprehensive understanding of the role played by enhanced dust fertilization in driving atmospheric CO2 concentrations lower during the ice ages. They will also be relevant to Sarmiento’s ongoing studies of the capacity of purposeful iron additions to increase CO2 sequestration in the Southern Ocean.

In addition to the evidence for iron fertilization of the ice age Southern Ocean noted above, Sigman group’s has found that the Antarctic was more strongly stratified during ice ages. Stratification lowers atmospheric CO2 by suppressing the upwelling of deep waters, with their high CO2 concentrations, at the surface. These findings, taken together, may do much to explain the lower CO2 of glacial climates. However, fundamental questions persist about this hypothesis. In particular, what processes lead to polar ocean stratification, and why did stratification break down abruptly at the shift toward the current interglacial period? In short, can the observed polar ocean changes be incorporated into a complete, self-consistent model for glacial cycles?

Sigman’s group will address these questions along two parallel paths. First, they will investigate the phasing of the Southern Ocean changes with other components of climate change upon glaciation and deglaciation. To accomplish this, new high resolution sediment records from the polar ocean have been identified that are more easily correlated to glaciological records, and Sigman’s group has begun to develop protocols to facilitate our work in these new archives. Second, Sigman plans to continue his collaborations with NOAA Geophysical Fluid Dynamics Laboratory scientists and the Sarmiento and Philander groups to develop a model framework that allows for physically explicit simulation of glacial cycles.

A second area of interest within the CMI science group is 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 colleague Gerald Haug (GFZ Potsdam) 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. This work has demonstrated shifts in the mean position of the intertropical convergence zone, which has affected rainfall in these regions and influenced ancient human civilizations such as that of the Maya.

The effort to generate new and better data is ongoing at many institutions on an international basis. However, there is now a critical need for synthesis, as there appears to be a remarkable degree of internal consistency among the data, such as global coordination of the observed regional shifts in the intertropical convergence zone. This synthesis will open the way for new modeling activities at Princeton and GFDL to understand the origins and significance of Holocene climate change.