Climate change is primarily a story of excess carbon—extra carbon dioxide, methane, soot and other organic compounds distributed in Earth’s environment as a result of human activity. And yet, in spite of carbon’s central role, the patterns and mechanisms by which it flows between Earth’s atmosphere, oceans and terrestrial biosphere—the details of the global carbon cycle—are among the greatest sources of uncertainty in climate models. At heart, this ignorance is a symptom of missing data. ‰ÛÏWe have a real problem of limited observations, which limits our most fundamental understanding of the role of some of the most important regions in the carbon cycle. We know as little about the super-important Southern Ocean as we do about the super-important tropical carbon sinks,‰Û says Josep Canadell, executive director of the Canberra-based Global Carbon Project.

Canadell is among the coordinating authors of a program to tackle that ignorance. Their proposal, organized through the intergovernmental Group on Earth Observations (GEO), is a sweeping plan proposing to multiply measurements of carbon flows and facilitate their conversion into analysis and model-building. The goal of the GEO Carbon Strategy, says Canadell, is to determine carbon’s fluxes through the environment with enough accuracy to pin-point the regions that are carbon sinks and sources, and the processes responsible for absorbing and releasing carbon.

JosÌ© Achache, GEO’s Secretariat Director, foresees the ‰ÛÏintegration of individual carbon observations into a global flux model which shows where the sources and sinks of carbon are distributed and how they behave as a factor of time and season.‰Û

Global negotiations to forge a new treaty on greenhouse gas emissions, such as the talks in Cancun in December 2010, show the urgency to fill the holes in carbon observation, and deliver a robust model of the carbon cycle. ‰ÛÏIn order to properly discuss how to reduce emissions, there has to be agreement on where emissions are coming from,‰Û says Achache. ‰ÛÏThat will make discussions easier.‰Û

Global carbon observation, for example, should increase confidence in estimates of emissions from changing land uses such as forest clearing—or even enable direct observation and quantification of such emissions. That could clinch agreement on proposals to allow rich countries to meet emissions reductions obligations by financing reforestation in developing countries. ‰ÛÏTo put in place mitigation measures such as reforestation, you need to have proof and accuracy on how much carbon is being sequestered,‰Û notes Achache.

Agreement on the science of land-use impacts could also help move the negotiations toward discussion of tougher political issues, such as how fast their societies’ energy practices must change to head off destructive climate impacts.

‰ÛÏPeople won’t be able to, as we say in French, ‰Û÷hide behind their little fingers’,‰Û says Achache. A capacity to verify emissions from regions or even individual power plants, meanwhile, could increase the confidence of global leaders that their negotiating partners in a post-Kyoto treaty will actually deliver on their commitments.

Interviews with GEO Carbon Strategy architects such as Canadell and Achache as well as outside scientists involved in carbon observation and modelling suggest that there is considerable momentum already toward implementing GEO’s vision of an international carbon observing system—but also considerable need to secure the resources required to address under-observed regions and ecosystems.

Towering Observations

GEO’s Carbon Strategy builds upon efforts to coordinate carbon observations that go back nearly a decade. The best known is a network called Fluxnet, forged to integrate data from 30-meter-tall eddy flux towers.

Flux towers combine high-frequency tracking of CO₂ concentrations and vertical wind speed in a technique called eddy covariance that provides a continuous calculation of the net carbon dioxide flow in or out of the ecosystem below. Correlating the changes in the carbon flux with the shifting ecological and climatic conditions driving them provides mechanistic insights into the carbon cycle.

‰ÛÏThey’re not just measuring carbon-in and carbon-out. They’re understanding processes that control that carbon-in, carbon-out,‰Û says Richard Norby, an experimental botanist at Oak Ridge National Laboratory in Tennessee.

Eddy covariance towers expanded over the past decade from 100 to almost 600 largely through the initiative of individual researchers or government agencies, but coordination has grown apace, beginning with the formation of regional networks such as AmeriFlux and CarboEurope, which aligned to form the global Fluxnet body. Fluxnet has hammered out rules for data standardization and interchange to create a coherent dataset of flux readings from over 250 sites worldwide.

The payoff is huge: Fluxnet’s data integration has encouraged modellers to bite into the flux data. ‰ÛÏIt’s making the right people talk to each other. It makes things move,‰Û says Canadian flux covariance pioneer Harry McCaughey, a climatologist at Queens University and participant in the Canadian Carbon Program. The result, according to Canadell, is that Fluxnet has ‰ÛÏdemonstrated the benefits of standardization, allowing a major synthesis effort benefiting both the global investment and the individuals participating.‰Û

Funding Vagaries

While voluntary networks-of-networks such as Fluxnet show what’s possible as far as uniting hitherto disparate carbon measurements, they are not a strong model for ensuring sufficient density and global reach of carbon measurements. Despite recent efforts, Fluxnet remains primarily a North American and European effort. Climates that are cold and dry, and those that are wet and warm, are hardly represented at all. There are only a half-dozen sites in South America, two in Africa, and none in Antarctica.

‰ÛÏThe development of Fluxnet has been driven by individual or groups of scientists with their own and national interests applying for research funding. That is why 70% of the towers are in Europe and in the US, representing a small fraction of the world’s ecosystems and an even smaller fraction of the global fluxes,‰Û says Canadell.

While new mechanisms are needed to drive the global expansion of flux measurements, there is also a need to secure funding for many existing sites, which remain funded through science grants and are thus vulnerable to the vagaries of soft funding. ‰ÛÏWe could indeed go backwards, as one can predict some research funding fatigue for this type of long-term measurement,‰Û says Canadell.

That backsliding scenario appears to be playing out in Canada, where eight eddy covariance towers are running out of funds. The Canadian towers grew out of BOREAS, a 1990s-era NASA-funded program to study the boreal forest—the circumpolar forest ringing the Arctic where thawing permafrost and potentially run-away methane emissions constitute a major climate wildcard. After BOREAS wrapped up, Canadian scientists such as McCaughey banded together to keep the flux towers alive, securing five years of funding from one agency and then three years from another. One year of emergency funding ended in December 2010.

‰ÛÏAs a participant in this trying to keep two sites running that are part of this global network, I’m struggling to see my way past the end of this fiscal year,‰Û says McCaughey. ‰ÛÏIt could all just stop in four or five months.‰Û

McCaughey says the damage would be felt well beyond Canadian science, because Canada—especially its immense boreal forests—is a major component of the global biosphere. ‰ÛÏCanada has such a humungous chunk of the terrestrial landscape, and, when you get right down to it, we don’t know a great deal about it,‰Û says McCaughey.

Ending the flux tower readings will leave global science in the dark as to how Canada’s ecosystems are responding as CO₂ concentrations rise and climates change, says fellow Canadian Fluxnet participant Andrew Black, an expert in biometeorology and soil science at the University of British Columbia. Black says killing flux measurement is both premature and short-sighted, since barely a decade of readings is hardly sufficient to understand the carbon cycle observed to date, let alone predict what will happen over the century. ‰ÛÏThat’s not enough to figure it all out. It’ll never be enough to predict 50 or 100 years from now how these ecosystems are going to change,‰Û says Black.

Operational Observations

Moving carbon observation from the soft budgets of the research realm onto the steadier funding of operational measurement programs appears to be happening on a large scale in Europe, where the European Union is ramping up its Integrated Carbon Observation System. However, even scientists at premier operational observation agencies say this is not a guarantee of adequate carbon observations.

The development of Fluxnet has been driven by individual or groups of scientists with their own and national interests applying for research funding. That is why 70% of the towers are in Europe and in the US, representing a small fraction of the world’s ecosystems and an even smaller fraction of the global fluxes.‰Û ‰ÛÒ Josep Canadell, executive director, Global Carbon Project

Consider how the world looks from the office of GEO Carbon Strategy coauthor James Butler, director of global monitoring for the Earth System Research Laboratory that’s part of the U.S. National Oceanic and Atmospheric Administration. The Boulder, CO-based lab has assembled a premier carbon observation capability, combining 110 sites that bottle and ship flasks of air to Boulder every week with other sites that make continuous readings of carbon observation from air drawn from half-a-kilometer-tall towers. It has led the world in turning such data into interactive web-based analysis tools such as CarbonTracker, which integrates carbon observations and models of biosphere activity and atmospheric circulation to produce predictions of global carbon ‰ÛÏweather‰Û patterns. Those in turn are checked with readings from carbon-sensing aircraft.

Butler is ‰ÛÏvery confident‰Û of the numbers backing CarbonTracker—thanks to aggressive quality control measures. But he says there simply isn’t enough data being collected to provide accurate global models. Even predictions on a continental scale are marked by high levels of uncertainty.

CarbonTracker images for monthly CO₂ fluxes show color-coded globes flushing from blue to red and back as the regional biospheres fluctuate between sucking carbon out of the atmosphere and respiring it back. The patterns follow the seasonal variations of leafy vegetation that break out to capture CO₂, but then mature and drop off; as well as the footprints of regional droughts and other perturbations that squelch photosynthesis. But another animated image of Earth on the page is in resplendent yellows, illustrating the uncertainty of CarbonTracker’s observation-based predictions (commonly on the order of +/- 50 percent).

‰ÛÏIt’s pretty far off from the kind of information we need to verify emission reductions efforts,‰Û says Butler. ‰ÛÏWe need a lot more observations and model improvement to provide predictions on a regional scale, such as models the size of France or Kansas.‰Û That’s true even for North America, where the bulk of NOAA’s sites are clustered: ‰ÛÏWe’re trying to use North America as an example of how the world would go about doing this. Yet we need 5-10 times as many sites just there,‰Û he says.

Butler was counting on a 1.8-fold increase in the carbon observation budget this year that would have provided $8 million to help with the needed expansion—one that may be doomed by the changes in the U.S. Congress in the mid-term 2010 election. ‰ÛÏYou get a change of government and all kinds of things can happen,‰Û he says.

Increasing the predictive power of systems such as CarbonTracker will be a slog, predicts Butler, if current trends continue. ‰ÛÏIt’s going to take a decade, maybe two, because you can’t count on any country to hold the line very long,‰Û says Butler.

Satellite Hopes

Satellites serve to expand measurement power in many realms of Earth observation, and could do so for carbon cycle measurements as well, but that promise is long in coming. Technical factors explain part of the delay. For example, measuring carbon fluxes means peering down from orbit to pick up changes in levels of fast-mixing gases occurring right at the Earth’s surface.

‰ÛÏThe active signals are at the bottom. It’s got to go all the way through the atmosphere,‰Û says Butler. But bad luck has also played a part: in February 2009 a failed launch destroyed NASA’s Orbiting Carbon Observatory (OCO) satellite, one of the first satellites to track CO₂ levels, leaving carbon-cycle scientists worldwide crestfallen. ‰ÛÏWe were relying on OCO,‰Û says Jean-NoÌÇl ThÌ©paut, who runs the data division for the European Center for Medium-Range Weather Forecasts in Reading, UK.

OCO’s ill-fated launch put the spotlight on the Japan’s Greenhouse Gas Observing Satellite (GOSAT), launched in January 2009. GOSAT and OCO had similar goals of measuring levels of CO₂ and CH₂ in columns of atmosphere to augment surface flux estimates from ground-based measures, especially over regions such as the Tropics and Siberia that are poorly covered by the surface network. But GOSAT’s predicted resolution was lower (4ppm versus 1ppm for OCO).

Researchers such as ThÌ©paut say that calibrating GOSAT’s instruments is proving difficult as a result of atmospheric constituents such as aerosols and cirrus clouds that can distort the signal from CO₂ in the reflected sunlight that GOSAT detects. ‰ÛÏThis type of measurement is very sensitive to the scattering effects of the atmosphere,‰Û says ThÌ©paut. Given the weakness of the CO₂ signal, he expects a ‰ÛÏlong learning curve‰Û for GOSAT.

That calibration challenge increases the urgency for launching a second satellite aimed at greenhouse gases ‰ÛÒ which NASA is rushing to do with the launch of a second OCO satellite scheduled for early 2013. ‰ÛÏHaving two satellites making measures with slightly different procedures and instruments should be a terrific advantage,‰Û says Achache, the GEO secretariat director.

Japan, for its part, is already making plans to launch a second GOSAT in early 2014 when the first GOSAT mission ends.

Staying Grounded

GEO’s Carbon Strategy architects say that, despite the slow progress on the satellite front, the pipeline is now in place and focus should shift back to the need for further ground-based or in situ monitoring and sea-surface measures. Achache calls the lack of dedicated funding for ground-based monitoring such as eddy flux towers ‰ÛÏa real problem‰Û that GEO will be advocating for in 2011.

Canadell agrees, and emphasizes the need to ensure a wide range of in situ measurements beyond the atmospheric CO₂ and flux measurements handled by Fluxnet and NOAA’s networks. He points to the decadal soil carbon and forest inventories called for by the GEO strategy. ‰ÛÏThese are a bit less fashionable, but we think equally important,‰Û says Canadell. ‰ÛÏEven more as they help us also to validate other data products such as satellite-based measurements.‰Û

And there must be emphasis to push all of these measurements out to every ecosystem and region on the planet, says Canadell. GEO’s Carbon Strategy calls, for example, for equipping all research vessels and support ships going to Antarctica with high accuracy automated CO₂ measurement systems. Canadell warns that if state-level funding decisions continue as they have, the result will be the same patchwork of observations with an ‰ÛÏover-emphasis in observations in the rich temperate world and under-investment in the remote regions which have much larger importance for the carbon cycle and carbon-climate feedbacks, such as the tropics and boreal regions.‰Û

Butler points to one more focus for researchers: Learning how to communicate their findings to the citizens whose behavior and opinions ultimately matter most in determining anthropogenic influence on the carbon cycle, and global climate. ‰ÛÏWe can’t seem to capture [the attention of] Joe 6-pack. He hasn’t gotten it yet,‰Û says Butler.

His point is that it may take decades to deliver models with high precision—decades that could put civilizations in a precarious position if carbon emissions continue unabated. ‰ÛÏIt’s very clear that the sooner we reduce CO₂ emissions to zero the better. CO₂ has not been this high on Earth for millions of years. We’re really running a risk here,‰Û says Butler. He points out that—while precision modeling is still unavailable—everything scientists know now about oceans and the terrestrial biosphere suggests that they will not continue to absorb half of the CO₂ humanity adds to the climate.

McCaughey shares the sentiment: ‰ÛÏWe are putting the carbon balance of the globe fundamentally out of whack, and the atmosphere and the oceans can’t handle it anymore. All of these systems have capacity limits at some point.‰Û