Writing in the December 1 issue of Geophysical Research Letters, Dr. D. Andrew Rothrock of the University of Washington, Seattle, and colleagues say the average draft of the sea ice (that is, its thickness from the ocean surface to the bottom of the ice pack) has declined by 4.3 feet (1.3 meters). This represents a reduction of about 40 percent as compared with the earlier period.

The decrease in sea ice occurs all across the Arctic Ocean and corresponds to previously reported evidence that the Arctic climate is warming, the researchers say.

The sea ice data in the 1990s were acquired by the Scientific Ice Expeditions (SCICEX) program, which consisted of six extended cruises by nuclear submarines. This study analyzed data from three autumn cruises: by USS Pargo in 1993, USS Pogy in 1996, and USS Archerfish in 1997.

The SCICEX cruises covered most of the deep Arctic Ocean basin. Measurements of the sea ice thickness showed a perennial ice cover of from 3 to 9 feet (1-3 meters) in mean draft, which was considerably thinner than previous estimates.

The earlier data, used for comparison, began with the first nuclear submarine, USS Nautilus, in 1958 and continued through a cruise of HMS Sovereign in 1976. Data from the earlier cruises were adjusted as necessary for the time of year they took place, to correspond with the autumn data acquired in the 1990s.

There are few data available from the 1976-1993 period. The researchers conservatively estimate the overall errors in measurement as less than one foot (0.3 meters).

Rothrock writes that the changes from the earlier period to the present are "striking in the uniformity of their sign and in their magnitudes." That is, every one of the 29 sites compared between the earlier cruises and those of the 1990s showed a decline in ice thickness.

In certain areas, such as the Nansen Basin and the eastern Arctic, the thinning is over five and a half feet (1.7 meters). Elsewhere, such as the Beaufort Sea and Chukchi Cap, it is around three feet (one meter), and at the North Pole and in the Canada Basin the decrease lies between those extremes.

However, the researchers note, this is not an instance of ice thinning in one area while thickening in another, which could be induced by a change in surface wind patterns.

The researchers say the available data are insufficient to provide answers about the cause of the ice loss. They suggest several hypotheses about the flow of heat from the ocean itself, from the atmosphere, and from shortwave radiation. Other avenues to be explored include the amount of precipitation and snow cover in the region and ice movement.

A key topic for future research is whether ice volume has reached a minimum in a multi-decadal cycle or whether the decline will continue into the future.

Regardless, the researchers say, the thinning of Arctic ice that has already occurred is "a major climatic signal that needs to be accounted for in a successful theory of climate variability."

To help fill the gaps between the earlier and more recent submarine observations, they call for the public release of other ice thickness data gathered by submarines over the past 40 years, which they believe would be "of immense help" in refining this climatic signal.

The study was funded by the National Science Foundation, the Office of Naval Research, and NASA.