Since the beginning of CMI grant, the Sarmiento and Pacala groups have performed a variety of studies to improve the understanding of carbon dioxide surface flux location, intensity and evolution in time and space: development of sampling strategies, improvement of model components, and development of inversion methods. This work provides a unique knowledge base to allow the development of a carbon observing system that will incorporate observation data from various origins along with different kinds of models (atmosphere, ocean and land) to monitor both short and long time scale changes in the carbon cycle and to provide predictions for the future. The efforts presented in the following aim to pursue the development of such a system, bringing together atmospheric, oceanic and land studies.


Atmospheric inversions

As part of the North American Carbon Program (NACP) and in collaboration with NOAA/ESRL (the Earth System Research Laboratory, ex-CMDL), Sarmiento’s group has contributed towards the development of the strategy for a high-density observation network over the United States. This network will combine regional measurements of atmospheric mixing ratios of carbon dioxide (CO2), methane (CH4 ), carbon monoxide (CO), oxygen (O2), and other trace gases by tall tower and aircraft. Using simulations, Crevoisier et al. [2006] have highlighted the need for adding some stations in the northwest and the northeast of the United States to catch CO2 variations in these regions. To better assess the characteristics of the network, we plan to refine our study using models with higher spatial resolution. To provide coverage of all North America, contacts have been made with the Meteorological Service of Canada to examine the possibility of implementing five stations in Canada in the framework of the Canadian Carbon Program (CCP). Sarmiento’s group is looking forward to using these dense NACP and CCP observations, in synergy with measurements made by the ESRL global network, in classic atmospheric inversion studies, as well as in more direct approaches, two areas where the group has developed particular skills.

In particular, atmospheric inversions require the use of atmospheric transport models, whose accuracy in reproducing synoptic events has been tested by Gloor et al. [2006]. However, one issue remains concerning the modeling of atmospheric exchange between the boundary layer and the upper atmosphere (convection). Aircraft data provide an interesting way of characterizing these exchanges and thereby improving atmospheric model realism. This task will be undertaken through a close collaboration with GFDL.

The simultaneous measurement of different species also gives the opportunity of taking advantage of the correlation existing between them. Indeed, CO2, CH4, CO, and O2 have sources and sinks in common (biomass burning, fossil fuel emissions) and also specific sources. Therefore, even if global estimation of CO2 sources and sinks have so far relied on inverting atmospheric CO2 concentration alone, the simultaneous use of different gases might provide strong constraints on their sources and on the transport affecting their atmospheric distributions. Sarmiento’s group thus plans on inverting various trace gases (CH4, CO, O2) along with CO2.


Air-sea fluxes

Sarmiento’s group has published a series of 4 papers over the past 3 years that use a variety of methods to estimate the magnitude of the ocean carbon sink over the past couple of decades. Combined together with related work by other colleagues, of which constraints provided by oxygen measurements such as those carried out by Bender’s group is an important component, these studies have reduced the uncertainty in the ocean carbon sink from an estimate of 2.0 ± 0.8 in earlier work, to of order 2.0 ± 0.3 at present. The focus of our research on air-sea fluxes will now turn to understanding what controls the variability of this sink, and the processes that will control the long term response of the ocean carbon sink to climate change.

The synergistic use of observations made over land and in the ocean’s interior has proven to provide a strong constraint on CO2 surface flux estimation [Jacobson et al. 2006a, b]. However, a stronger constraint on air-land fluxes could be obtained through a further reduction of air-sea flux uncertainties. A large source of uncertainty in constraining estimates of air-sea CO2 fluxes stems from the fact that the seasonal and interannual CO2 flux variability between the ocean and the atmosphere is poorly known.

To improve our understanding of the variability, Sarmiento’s group will initiate ocean modeling studies to identify variability in natural air-sea CO2 fluxes and the rate of uptake of anthropogenic CO2. Their particular focus will be on the northern hemisphere, which is where most of the land carbon sink appears to be. A particular area of interest will be the North Pacific Ocean, as the variability here has a significant impact on carbon source and sink estimates over North America. In a recent study, the group has shown that the Kuroshio Extension region of the North Pacific exhibits a very strong seasonality, with uptake being largest during winter, as well as a decadal trend towards increasing wintertime uptake in the model. These results were unexpected and their implications for CO2 fluxes over North America will be tested.

A particular emphasis of Sarmiento’s group has been on the Southern Ocean, which accounts for up to 40% of the oceanic anthropogenic carbon sink, with estimates differing greatly from model to model. In 2002, we significantly narrowed the uncertainty in the model estimates by using new observations of radiocarbon and chlorofluorocarbons to eliminate models that were clearly unrealistic. However, we still do not clearly understand the processes that determine the magnitude of this sink in the Southern Ocean and how it might respond to climate change.

A major factor determining the long-term uptake of anthropogenic CO2 uptake by the ocean is the rate at which deep water that is relatively uncontaminated with anthropogenic CO2 upwells to the surface. The table below gives a compilation of recent estimates of the rate at which water is being added to the “deep” ocean (which occurs primarily in the North Atlantic), and the pathway of the resulting return flow. As can be seen, model simulations generally have most of the upward return flow occurring within the Southern Ocean, whereas observational analyses suggest it should be almost entirely outside the Southern Ocean. Sarmiento’s group will make use of a major new data set of helium-3 emitted from mid-ocean ridges in the deep ocean in order to determine which of these return pathways is most consistent with the observations.

The water that upwells in the Southern Ocean can either turn to the north and sink below the main thermocline, or turn to the south and sink to the abyss as “bottom” water. In related modeling studies, Sarmiento’s group will be studying this upper ocean flow and its sensitivity to climate. The northward flowing branch supplies nutrients to the upper ocean that account for about two-thirds of biological production outside the Southern Ocean, so its response to climate could have major consequences for biological production. The southward flowing branch is the primary determinant of the air-sea balance of CO2 and changes in this process could have a significant impact on the future trajectory of CO2 in the atmosphere.

Table 3: This table shows the formation rate of deep water that occurs in the North Atlantic (NADW; estimated in parentheses), and the upward return flow partitioned into that which occurs within and outside the Southern Ocean. Note that the models have most of the upward return flow in the Southern Ocean, whereas the observational analyses have most of the upward return flow outside the Southern Ocean.

Ocean carbon models will then be used to forecast the potential impact of global warming on ocean carbon sinks. This study will also allow testing the realism of the models and of their representation of ocean circulation.


Air-land fluxes

Changes in climate have the potential to affect the geographic distribution of ecosystems, and the mix of species that they contain. For instance, in North America, there are over twenty major forest types and over 250 tree species, with marked, but contrasting, correlations with climate. However, little is known about how these ecosystems might respond to climate change, or how these changes might feed back on climate. Partly, this is because current global vegetation dynamics models do not represent richness of species and make simplifying assumptions about species diversity by assigning identical biophysical parameters to broadly-defined plant functional types, each containing tens to hundreds of individual species.

Now, the availability of various data sources for North America, ranging over different spatial and temporal scales, makes it possible to move to a new generation of vegetation models containing species-specific parameters and dynamics. Thus, the Pacala and Sarmiento groups propose to develop the capability to update species-specific biophysical parameters recursively in a data-assimilation scheme, given regional measurements of CO2 atmospheric concentration (NACP aircrafts and tall towers), local measurements of CO2 fluxes (flux towers), information on species composition and individual tree growth and mortality (USDA forest inventories), and possibly satellite data. The derived parameters will then be used as the basis for the aggregation of the species into a new set of functional types for use in a dynamic land model.

In this regard, LM3V, the new land model jointly developed at NOAA/GFDL and Princeton University, is well suited: it simulates vegetation dynamics and exchanges of water, energy and CO2 between land and atmosphere, and it has been successfully coupled to GFDL’s atmospheric and climate models. With the improved set of functional types and parameters, LM3V will be used to improve our understanding of the key processes governing the exchanges of CO2 between land and atmosphere over North America.

A particular focus will concern the modeling of carbon emissions by fires. Understanding how fires influence the structure and carbon dynamics of various ecosystems is needed to accurately simulate the future behavior of these pools of carbon in the context of global warming. The Pacala and Sarmiento groups have started the development of a fire model that will predict the conditions needed for a fire to start, and the surface burned by the fire. The calibration of this model will be performed through the use of on-ground and satellite observations of fire characteristics (burnt area, fire intensity), and the use of measurements of atmospheric concentrations of various trace gases more or less affected by fire. As well accurately simulating carbon fluxes, the fire model will give the opportunity to simulate the future behavior of vegetation in response to the increase of both fire occurrence and intensity that are expected in the context of global warming.

The proposed studies aim to reduce uncertainty in current estimates of North American carbon stocks and fluxes, and to improve our knowledge of their response to future changes in CO2, climate and land management. The previously presented atmospheric and oceanic studies will provide the boundary conditions needed for this data assimilation approach through the use of the atmospheric pulse response functions simulated for the purely atmospheric concentration data-based flux inversions.


Monitoring atmospheric oxygen

Bender’s core CMI research has involved measurements of the atmospheric O2/N2 ratio. This ratio provides a primary constraint for partitioning the sequestration of fossil fuel CO2 that does not remain in the atmosphere between ocean uptake and land biosphere uptake. The idea is that CO2 dissolution in the oceans has no effect on the atmospheric O2 burden, while CO2 uptake by the land biosphere makes biomass and O2 in nearly equal amounts. Hence measurements of the changing O2/N2 ratio of air allow us to determine the part of CO2 sequestration due to the land biosphere, and the part due to the oceans.

The challenge of this work is that the diagnostic O2 changes are very small. 1 gigaton of carbon uptake by biomass raises the O2/N2 ratio of air by 2.5 parts in 106. Since 1991, two laboratories have maintained global observing networks for measuring these changes, Bender’s lab and the lab of Ralph Keeling at Scripps Institution of Oceanography. The product of these laboratories has become a fundamental part of the data set used to characterize and understand global change.

There are 2 ancillary products of this work. First, O2/N2 ratios also reflect fundamental aspects of the ocean carbon cycle, including the interaction of carbon fluxes with circulation, and the biological productivity of the ocean basins. Second, Ar/N2 ratios, measured along with O2/N2 ratios, have a seasonal cycle due to ingassing and outgassing of the oceans as temperature changes, and can be used to study the fidelity of atmospheric mixing models.

We plan to continue our measurement program during the coming 5 year period. We also plan to do various studies to interpret our results. These studies, which are mostly done in the context of complex models of the oceans and atmosphere (and sometimes led by modelers rather than people from Bender’s group), have the following objectives:

  1. Partitioning CO2 sequestration between land biosphere and ocean for the length of the record.
  2. Refining, by inverse studies, the distribution of CO2 sources and sinks.
  3. Examining the links between CO2 sequestration and climate variations.
  4. Determining the fertility of the ocean basins and its interannual variability.
  5. Understanding mechanistic controls on these terms.