Riser and other members of Argo's steering team are fighting to ensure that the international collaboration doesn't unravel for lack of maintenance money.

"We have 3,000 floats now — but you have to keep replacing them," Riser said. "This is only the beginning."

But it's a good position, considering skeptics said the project would never work.

It's not that anyone doubted the powerful role oceans play in global climate. Warm currents in the Atlantic keep England and northern Europe from being as frigid as the Canadian plains. When the equatorial Pacific heats up, the resulting El Niños can change weather patterns around the world.

As greenhouse gases from cars and power plants trap more heat in the atmosphere, the oceans — which absorb much of that heat — will largely determine which places will get rainier, which will dry out and how storm patterns will shift.

What worried skeptics was the technical difficulty in pulling off the $80 million plan. Previously, ocean monitoring was a laborious and costly job, requiring a research ship and crew to maintain the finicky instruments.

It was the development of compact, automated sensors capable of withstanding years of battering in the ocean that made Argo possible. The first buoys were deployed in 2000.

 

A probe's cycle

At about 5 feet tall and weighing less than 60 pounds, the probes are considerably smaller than the oceangoing marvel for which they were named: The mythical sailing ship that carried Jason and his Argonauts on their quest for the golden fleece.

Each probe is programmed to sink to a depth of more than a mile, then drift in total darkness for nine days — deeper than submarines prowl. Every 10th day, the probes ascend, collecting temperature and salinity measurements as they go. At the surface, they transmit their data, then descend to start the cycle anew.

"This is an entirely new way of observing the oceans," said steering team member Dean Roemmich, of Scripps Institution of Oceanography in San Diego.

Whereas data from ships provide only scattered snapshots, Argo buoys are deployed about 180 miles apart and provide regular updates. Riser likens the network to the weather balloons and other meteorological instruments that take the atmosphere's pulse several times a day for weather forecasting.

"We now have a similar sort of grid over the ocean," he said.

For the southern hemisphere, a place where few ships travel, the number of deep measurements gathered by Argo exceeds the entire historical record.

All the data is available free on the Internet, where scientists can also track individual robots, or floats.

Like bemused dads, Riser and his colleagues marvel at some of the pickles their instruments land in.

One was trapped by tides in a shallow lagoon near New Guinea. "It's a very remote place, so we figured we'd never be able to retrieve it," Roemmich said. But 10 days later, the float popped to the surface — right on schedule — and winds blew it through a narrow passage, back to the open ocean.

Another vanished from the coast of Australia, then started transmitting from downtown Brisbane. A local researcher crisscrossed the city with a receiver, trying to pinpoint the signal as it jumped from place to place.

The mystery was solved only after she inquired at a local fishing dock. A trawler had dredged up the buoy, and the captain was driving around with it in his truck. He planned to use the shell as a mailbox.

Most of the buoys come to quieter ends, sinking to the bottom when their batteries die after four to five years.

 

Maintaining the network

Each buoy costs about $20,000 to build and deploy. To keep the network at its current strength will require about 700 replacements a year, at a cost of $14 million. More than 20 nations have contributed to the project so far, with the U.S. picking up half the tab.

"This is the most international experience in the history of oceanography," Roemmich said.

The National Oceanic and Atmospheric Administration pledged to continue its support, at least in the short term. But funding from other nations is more uncertain.

Steve Piotrowicz, who manages the program for NOAA, is relying on the "scream factor" to keep the project alive: The more useful Argo becomes, the more people will scream if it goes away, he said.

Already, shippers and seafood fleets are turning to Argo data to help plot currents, routes and promising fishing grounds. More than 1,000 scientific papers have drawn on measurements from the network.

For climate modeling, Argo is indispensable, said Trenberth, who is not involved in the project. Most of the heat from global warming is absorbed by the oceans, but the temperature increases aren't uniform, he said. This differential heating can change ocean circulation, which affects temperature and rainfall around the planet.

Changing weather also influences the ocean. Argo measurements have shown that much of the Pacific is becoming less salty, as a result of increased rainfall. Since salty water is denser than fresh, shifts in salinity will affect currents and circulation.

"The oceans are the big flywheels in the climate system," Piotrowicz said.

As oceans heat up, they expand, contributing to sea level rise — a process Argo can monitor with unprecedented accuracy.

Argo data also are pointing up weaknesses in the current understanding of climate change. Between 2003 and 2007, Argo floats measured no appreciable warming in the upper oceans — despite the fact that temperatures on land have continued to break records. At the same time, sea level is rising faster than can be explained by melting glaciers alone, said Josh Willis, of NASA's Jet Propulsion Lab.

"The lack of warming over a period of a few years isn't really that surprising, because of all the natural variability," he said. "It's a bit of a mystery what's going on with sea level."

Which is all the more reason to make sure Argo keeps running, Trenberth said.

A new generation of robots is already in the works, to gather more information at lower cost.

Riser is tinkering with sensors to gauge the oceans' biological productivity, measure wind speeds and estimate rainfall. Roemmich is shrinking the instruments down.

The scientists also have developed a system any parent would envy: a satellite cellphone link that will allow them to beam instructions to the robots, instead of having to wait for their creations to call in.