Defying Expectations

To try to understand the ocean’s carbon limits, Feely and the rest of the oceanography community measure dissolved carbon and the ocean’s pH. They probe its temperature, alkalinity, salinity, and record the presence of tracers like chlorofluorocarbons or helium to find out when the water was last exposed to the atmosphere. By the end of a typical cruise, they will have collected 50,000 or more measurements. And then they go out the next year to repeat it all again, and they have done this for more than three decades.

  Photograph of a researcher sealing ampules containing gas sampels retrieved from the southern ocean.

“When we started in the 70s and 80s, we had this concept: we’ll measure [carbon dioxide concentrations in the ocean] 10 years later, and we’ll just see the anthropogenic input,” says Feely. “We had very simplistic ideas that the anthropogenic changes would be the only changes we would see,” he adds a little ruefully.

Feely and his colleagues saw changes, but they weren’t at all the changes they expected. Carbon concentrations in the ocean did rise as atmospheric carbon dioxide skyrocketed, but in 2006, Feely and several colleagues announced that the equatorial Pacific seemed to be venting more carbon dioxide to the atmosphere between 1997 and 2004 than it had in previous years. And in 2007, Ute Schuster and Andrew Watson, oceanographers from the University of East Anglia, reported that amount of carbon that the North Atlantic Ocean soaked up decreased by a factor of two between 1994 and 2005. The ocean, or parts of it, seemed to be taking up less, not more, carbon.

Had the ocean already stratified, slowing the rate at which it soaked up carbon? Schuster and Watson believed they saw stratification at work in the North Atlantic, but the drop in the amount of carbon being taken up was too large for global warming to be acting alone. During the decade Schuster and Watson made their observations, a large-scale weather pattern, called the North Atlantic Oscillation, shifted.

Like El Niño in the Pacific, the North Atlantic Oscillation (NAO) changes weather on a large scale. In the early 1990s, says Watson, the North Atlantic Oscillation brought stronger and more frequent winds to the northern regions of the North Atlantic during the winter. The winds stirred the ocean, tucking carbon-dioxide-laden surface water down and pulling unsaturated water to the surface, in effect increasing the rate at which the North Atlantic took up carbon. By 2000, the North Atlantic Oscillation shifted, calming winds and allowing warmer waters to expand north. These two changes increased stratification in the North Atlantic and slowed the carbon uptake between 1994 and 2005, says Watson.


A scientist uses a torch to seal glass ampoules containing dissolved gases extracted from seawater. Tens of thousands of samples were collected during the Climate Variability and Predictability cruises. Scientists were surprised when decades of sampling revealed a complex relationship between man-made carbon dioxide emissions, cyclical changes in climate, and the oceanic carbon cycle. (Photograph ©2008 Brett longworth.)

Graph of winter North Atlantic Oscillation, 1980 through 2007.

In the Pacific, Feely tracked the increased venting at the equator to a shift in another natural pattern. The Pacific Decadal Oscillation is a decades-long climate pattern that alternately warms and cools the ocean. “When the Pacific Decadal Oscillation shifts into its cold phase, you get stronger winds and stronger upwelling,” says Feely.


In the early 1990s, the North Atlantic Oscillation was strongly positive (green bars) during winter. In the mid-90s, the climate pattern became more variable, but overall, was weakly negative. Lower winds and higher temperatures in the North Atlantic slowed carbon uptake. (Graph by Robert Simmon, based on NCAR data.)

Graph of Pacific Decadal Oscillation index, 1905 through 2005.

The ocean does not take up carbon uniformly. It breathes, inhaling and exhaling carbon dioxide. In addition to the wind-driven currents that gently stir the center of ocean basins (the waters that are most limited by stratification), the ocean’s natural, large-scale circulation drags deep water to the surface here and there. Having collected carbon over hundreds of years, this deep upwelling water vents carbon dioxide to the atmosphere like smoke escaping through a chimney. The stronger upwelling brought by the cold phase of the Pacific Decadal Oscillation apparently enhanced the size of the chimney and let more carbon escape to the atmosphere.


In the equatorial Pacific, carbon dioxide venting increased between 1997 and 2004. This coincided with a shift in the Pacific Decadal Oscillation (PDO) from a warm phase (positive) to a cool phase (negative), during which winds and upwelling of deep water are stronger. The graph shows the values of the PDO index as 1-year averages (line) and five-year averages (shaded). (Graph by Robert Simmon, based on JIASO data.)

  Map of deep ocean circulation, including carbon sinks.

After 30 years of measurements, the ocean carbon community is realizing that tracking human-induced changes in the ocean is not as easy as they thought it would be. It wasn’t a mere matter of measuring changes in carbon concentrations in the ocean over time because the natural carbon cycle in the ocean turned out to be a lot more variable than they imagined. “We discovered that natural processes play such an important role that the signals they generate can be as large as or larger than the anthropogenic signal,” says Feely. “Now we are trying to address how these decadal changes affect the uptake of carbon. Once we account for these processes, we can remove them from the data set and calculate the anthropogenic carbon dioxide as the residual.” But to track the increasingly complicated carbon balance sheet, the ocean community needed models, mathematical simulations of the natural world.


The global oceans are connected by deep currents (blue lines) and surface currents (red). Carbon from the atmosphere enters the ocean depths in areas of deep water formation in the North Atlantic and offshore of the Antarctic Peninsula. Where deep currents rise towards the surface, they can release “fossil” carbon dioxide stored centuries ago. (Map by Robert Simmon, adapted from the IPCC 2001 and Rahmstorf 2002.)