Global sea levels could rise by more than double the current best estimate, according to a new analysis of climate change in Antarctica.
The modelling assessment says that Antarctic melting alone could contribute more than a metre to sea level by the end of this century.
By 2500, according to the study, the same source could cause levels across the world to rise by 13m.
The authors say that rapid cuts in carbon emissions could limit this risk.
In 2013, the Intergovernmental Panel on Climate Change (IPCC) predicted that, without any restrictions on carbon emissions, the seas around the world likely rise by up to 98cm by 2100.
However, the IPCC estimates contained a minimum contribution from Antarctica.
Other analyses since then have projected bigger increases, with a recent study suggesting that the oceans were rising faster than at any time in the past 2,800 years and by 2100 they could be up to 1.31m higher.
The exact level of Antarctica's impact on these projections has been vigorously debated. Late last year, a research paper suggested that projections of a contribution of a metre or more were not plausible.
But this new study argues that by 2100 the world could see 1.14m of sea-level rise from Antarctica alone.
Additions to the model
The scientists say that their model is able to provide a more accurate prediction because it incorporates the impacts of some physical processes for the first time.
While other models have focussed on the impact of warmer waters melting the ice shelves from below, this new study also includes the effect of surface melt-water and rain trickling down from above and fracturing supporting ice, hastening its slide to the sea.
The model also calculates the impact of the disintegration of floating ice shelves. If this happens, it will reveal walls of ice so tall that they cannot support their own weight.
The scientists involved expect that these extra factors will kick in over the coming decades, as warming from the atmosphere (not just from warmer waters below) becomes the dominant driver of ice loss.
"One reason that other models didn't include the atmospheric warming is because it hasn't started to happen just yet," said co-author Dr David Pollard from Penn State University, US.
"In Antarctica, around the edges at sea level, it's just beginning to get up to the melt point in summer.
"With that warming, the flanks of Antarctica will start to melt drastically in about 50 to 100 years - and then it will start to kick in according to our model."
The authors believe that they have demonstrated the accuracy of the new model by correctly replicating sea-level rise in warm periods, millions of years into the past.
"Recently, we looked at the long-standing problem posed by geological evidence that suggests sea level rose dramatically in the past, possibly up to 10 to 20 metres around 3 million years ago, in the Pliocene," said Dr Pollard.
"Existing models couldn't simulate enough ice-sheet melting to explain that."
If the world continues to emit "business as usual" levels of carbon dioxide over the coming decades, the scientists argue that sea-level rise will be double what has already been estimated for the coming 100 years.
"If these processes do kick in and they end up being as important as we think that they could be, then they really do have a big impact," said Prof Robert DeConto from the University of Massachusetts, Amherst.
"West Antarctica is responding very soon in these simulations and that ends up having a big impact on North America in particular."
Other researchers have praised the development of the new model for including impacts such as surface melt water and ice-cliff collapse, but they are uncertain about the conclusions.
"I have no doubt that on a century to millennia timescale, warming will make these processes significant in Antarctica, as well as Greenland, and drive a very significant Antarctic contribution to sea-level rise," commented Prof David Vaughan of the British Antarctic Survey.
"The big question for me is, how soon could this all begin, and could it be early enough to drive substantially higher sea levels by 2100? I'm not sure, but these guys are definitely asking the right questions."
The authors believe that there is "good news" in their report. If global emissions of carbon are curtailed significantly then the extra factors that substantially boost Antarctic melting will be avoided.
Seas will continue to rise, but not at the runaway rates suggested by this paper, which has been published in the journal Nature.