Although there is consensus in the scientific community that the planet is warming, and on the overall budget of carbon, large uncertainties still exist concerning the location of natural carbon sources and sinks, and the mechanisms that control their variability and response to climate change in both the land and ocean. The CMI Science Group has already made significant progress in estimating sink magnitudes and explaining their causes. As we continue to improve our observational and simulation capacities, we will increase confidence in estimates of future CO2 change and the mitigation effort needed to avoid damaging climate consequences.

We emphasize that the CMI science component is a synergistic activity building on a broad range of research on the Princeton University campus that includes that carried out by NOAA’s Geophysical Fluid Dynamics Laboratory (GFDL); the Cooperative Institute of Climate Science (CICS), which is a collaboration between Princeton University and GFDL; and the large basic research programs maintained by all the participants in the CMI science program with support from government agencies such as NOAA, NASA, NSF, and DOE. With help from our partners, we will continue to expand and deepen research in our main thematic areas.


Carbon observing system

A series of papers published during 2007 represent a culmination of a long term effort by Sarmiento’s group to determine the long term average of the air-sea CO2 flux and how knowledge of this influences our understanding of terrestrial CO2 fluxes. Now the group is focusing primarily on analysis of new observations being obtained by ongoing repeat surveys, with a major emphasis on the detection of the anthropogenic invasion of carbon in the face of very large interannual variability of the natural carbon cycle.


Ocean inventories

With the goal of reducing uncertainty in estimates of the time-evolving uptake of anthropogenic CO2 from Repeat Hydrography measurements, the Sarmiento group will extend previous work to properly interpret the Repeat Hydrography observations and to quantify the uptake of anthropogenic CO2 by the ocean. Modeling work will consist of the use of forced ocean circulation experiments with GFDL’s MOM4, which includes GFDL’s Ocean Biogeochemistry model (TOPAZ) online. An evaluation of model-simulated DIC fields against repeat measurements in the ocean will facilitate a dynamical interpretation of the observations.

Additionally, output from the coupled earth system model described in the following section will be used to study the dynamical processes that control natural variability of DIC. An important limitation of forced ocean experiments lies in the surface freshwater boundary condition (namely that sea surface salinity is typically restored to observations), and this is expected to damp variability in ventilation processes. Thus the coupled model simulations will serve as an important complement to the forced ocean simulations for identifying and understanding processes controlling DIC changes in the ocean interior.


Air-sea fluxes

Sarmiento’s group is also interested in producing an interannually varying air-sea CO2 flux data product using satellite measurements and the GFDL Earth System Model (ESM2.1). The recently completed joint atmospheric inversion studies of Jacobson et al. (2007a) and Jacobson et al. (2007b) find that the land regions of the tropics and the Southern Hemisphere are likely to be a large source of CO2 to the atmosphere. In contrast to other studies, this result suggests that tropical land ecosystems may not be experiencing enhanced growth due to elevated atmospheric CO2 concentrations. However, the joint inversion result derives mainly from oceanic inversion fluxes, which have no interannual variability and use observations of dissolved inorganic carbon (DIC) and nutrients measured mainly during the early- to mid-1990s. Thus there is a clear need for additional data constraints to evaluate the time-varying component of the problem.

We intend to develop a new gridded interanually-varying air-sea CO2 flux dataset, relying largely on remotely-sensed satellite data products, the GFDL ESM2.1, and a semi-empirical nonlinear reference scheme (neural networks). The semi-empirical scheme will first be used to determine relationships between variability in simulated air-sea fluxes and model variables that correspond to fields available from remote sensing. The relationships will then be used to produce the air-sea CO2 flux dataset from the satellite data.


Air-land fluxes

As a complement to recent research in the Sarmiento group to detect interannual variations in CO2 flux exchanges over North America, work will be conducted by the Pacala and Sarmiento groups towards developing a mechanistic understanding of the detected changes. As a long-term goal, attribution of changes in carbon fluxes will involve distinguishing between a variety of influences, including: (a) land use changes, (b) fertilization associated with increased atmospheric CO2 concentrations, (c) the action of fires, (d) changes in the mean state of the atmosphere associated with global climate change, and (e) natural variability in the carbon cycle. As an important first step towards attribution, we will work with models and observations to understand the natural variability component, which we define to be the background variability that would be expected to occur in the absence of human intervention. For the purposes of detection, this represents the background noise against which the anthropogenic perturbation is to be detected.

In order to characterize the amplitude and structure of CO2 flux variations over North America over interannual to decadal timescales, we intend to use the LM3V model (developed jointly by Princeton and GFDL) coupled to GFDL’s three-dimensional dynamical atmosphere model AM2. In doing so, we intend to test the hypothesis that interannual variability in CO2 fluxes over North America is largely driven by internal variability of the atmosphere, whereas decadal variability in CO2 fluxes is a largely deterministic non-local response to sea surface temperature variations.


Climate and biogeochemistry

An important direction of future research in the Sarmiento group will be to explore the ways in which carbon in the ocean couples to the other components of the climate system in earth system models, and how this can help us understand paleoclimate and predict the future response of the earth system to global warming.


Improved models of nutrient cycling

Recent research at Princeton and GFDL has clarified that the critical parameters that control the air-sea balance of natural CO2 are the absolute concentrations of preformed and remineralized nutrients in the ocean. This new understanding has led us to a renewed emphasis on improving our simulations of the processes that determine the distribution of these two types of nutrients in the ocean interior, and how climate change modifies those processes. Our new more careful examination of these properties has revealed a number of problems that reflect limitations in both ocean circulation and ocean biogeochemistry models. We are tackling these problems on three fronts:

We have a major new effort underway to develop improved models of ocean circulation by a strategy of new tracer simulations that will enable us to test the models, and by a wide range of sensitivity studies aimed at improving the models, particularly in the critical region of the Southern Ocean.

In addition to our ongoing efforts to develop surface ecology models, we have a new effort underway to more carefully examine ocean interior nutrient, oxygen, and carbon distributions and determine how these can be improved. This research will benefit greatly from the deployment of new methods for high speed convergence of the models to solutions that will enable us to carry out a much wider range of parameter sensitivity studies than has ever been possible before.



To increase our understanding of past climatic changes, the Sigman and Sarmiento groups will continue to improve modeling tools and collect new data from ocean sediments.


Model development

Development of a new coarse-resolution coupled climate model of the earth system including the terrestrial as well as oceanic carbon cycles will be continued. This is an essential component of our strategy that will make it possible for us to carry out long (multi-millennial) simulations of the climate, a much more complete set of tests of the many theories for the ice ages that have been developed at Princeton and GFDL, and a wide array of studies of the sensitivity of the carbon cycle to future climate change.


Understanding glacial/interglacial cycles

Available data suggest that a hydrographic front within the modern Antarctic Zone, which today has modest but measurable impacts on the carbon cycle, was far more important during the last ice age. To evaluate this hypothesis, Sigman’s group will work in partnership with international collaborators to recover and analyze new sedimentary materials from the different regions of the Antarctic, as well to study different regimes of the modern Antarctic to ground-truth paleoceanographic tools. This effort should also yield records with better age control than previously available, allowing comparison of the detailed timing of Antarctic ocean changes with other records of climate change and with the end-glacial rise in atmospheric CO2.

Over the past year, graduate student Abby Ren in Sigman’s group has developed a new paleoceanographic tool, based on the organic matter internal to the calcium carbonate microfossils of planktonic foraminifera (surface-dwelling animals), to reconstruct nutrient conditions across the ocean outside the most polar regions. While the utility of Sigman’s previous tools was limited to the more productive regions of the ocean, this new tool opens up essentially the rest of the ocean to parallel investigation. Although the development work has many aspects and will continue for some time, Sigman’s group is now entering the application phase, with the work beginning on several sediment records from across the global ocean. Among the central goals is testing the expected low latitude consequences of the polar stratification hypothesis.


Geoengineering and a carbon sequestration “truth squad”

This will continue to be a low-level effort of examining geoengineering and carbon sequestration proposals using our earth system models. Past research has included examinations of deep sea CO2 sequestration and iron fertilization. A major ongoing project that will be completed during the following year is a re-examination of the iron fertilization scenario using new ecosystem models that were finally completed over recent years.