Mystery of the missing heat: upper ocean has cooled slightly in recent years, despite warming climate.

Between 2003 and 2005, the top layers of the world's oceans cooled slightly, but scientists aren't sure where the heat went.

According to climate data gathered worldwide, 2003, 2004, and 2005 are three of the five warmest years since reliable record keeping of global air temperatures began more than a century ago. However, oceanographic surveys suggest that on average, the upper 750 meters of the world's icefree oceans cooled about 0.03[degrees]C during that 3-year period.

This cooling reverses an oceanic-warming trend observed since the 1950s, oceanographer John M. Lyman and his colleagues report in the Sept. 28 Geophysical Research Letters. Between 1993 and 2003, the average temperature of the upper layers of the icefree ocean rose about 0.09[degrees]C, they note.

The newly documented cooling occurred throughout the top 750 m of ocean and seems to have extended to deeper waters as well, says study coauthor Josh K. Willis, an oceanographer now at NASA'S Jet Propulsion Laboratory in Pasadena, Calif. Data used in the new analysis were gathered by buoys tethered in deep water, instruments towed by or dropped from ships, and an armada of robotic probes, says Lyman, who's at the National Oceanographic and Atmospheric Administration (NOAA) laboratory in Seattle.

While the top layers of the ocean have cooled slightly overall, some limited areas have warmed, says coauthor Gregory C. Johnson, also of NOAA in Seattle. The cooling trend, as well as its patchiness, probably results from variations in climate cycles such as the Pacific Decadal Oscillation, he notes.

"Even within a long-term warming trend, you can have short-term drops in [ocean] temperature due to year-to-year variability," says Lyman.

Scientists are working to identify where the heat went. One possibility: It may have moved to the deepest layers of the ocean. The cooling of surface waters would cause them to contract, triggering a small drop in sea level, says Willis. But satellite data suggest that sea level is still rising. So, the missing heat may have gone deep, causing waters there tO expand and prevent a decline in sea level. However, "it's hard to envision a way to put that much heat down deep so quickly," says Willis.

In another scenario under consideration, the missing heat may have radiated into space. However, satellite observations don't support that notion, says Bruce A. Wielicki, a physicist at NASA'S Langley Research Center in Hampton, Va. Yet another possibility is that the heat warmed some of the waters in polar regions and promoted melting of the ice cover there, he notes.

"We have a few more pieces to unravel" about where the heat has gone, comments Sarah T. Gille, an oceanographer at Scripps Institution of Oceanography in La Jolla, Calif. "It's a real conundrum."