By plotting global mean surface temperatures against sea level rise, the team found that levels could rise by 59% more than current forecasts.

The researchers say the possibility of greater increases needs be taken into account when planning coastal defences.

The findings have been published in the online edition of the journal Science.

The team from Germany and the US found that for the timescale relevant to human-induced climate change, the observed rate of sea level rise through the 20th Century held a strong correlation with the rate of warming.

When applied to the possible scenarios outlined by the Intergovernmental Panel on Climate Change (IPCC), the researchers found that in 2100 sea levels would be 0.5-1.4m above 1990 levels.

This projection is much greater than the 9-88cm forecast made by the IPCC itself in its Third Assessment Report, published in 2001.

Troubled waters

The paper's lead author, Stefan Rahmstorf, from the Potsdam Institute for Climate Impact Research, Germany, acknowledges that projecting future outcomes is a challenge.

He writes in Science Express: "Understanding global sea level changes is a difficult physical problem, as a number of complex mechanisms with different timescales play a role."

These include:

• thermal expansion of water through heat absorption
• water entering the oceans from glaciers and ice sheets
• increased ice flows after the removal of buttressing ice shelves

Professor Rahmstorf said he decided to use observational data because computer models of climate significantly under-estimated the sea level rise that had already occurred.

"The fact that we get such differences using different methods shows how uncertain our sea level forecasts still are," he said.

Greenland's glaciers have been sliding faster towards the sea

He added that the main uncertainty was the response of large ice sheets in Greenland and Antarctica to rising temperatures, which was difficult to predict.

Simon Holgate, from the UK's Proudman Oceanographic Laboratory, said fellow scientists would have to take a closer look at Professor Rahmstorf's findings; until then, the jury would remain out.

"Normally, you would use the heat content of the surface layer of water rather than just using the surface temperature because the surface temperature is affected by a lot of other factors," Dr Holgate observed.

"When you try to do a similar exercise as Professor Rahmstorf but using heat content you do not get such a good correlation."

He added that only 40% of observed sea level rise was believed to a result of thermal expansion.

Dr Holgate said the next assessment report from the IPCC would include data based on more robust modelling, thereby reducing uncertainties surrounding models on sea level rises.

The IPCC is set to publish its much anticipated Fourth Assessment Report in February 2007.

http://news.bbc.co.uk/2/hi/science/nature/6179409.stm