Scientists say major predictions about sea level rise may be too low. 

Maximum sea level rises predicted by the most recent report from the Intergovernmental Panel on Climate Change (IPCC) are probably not high enough, according to Danish and Norwegian scientists.

They calculated future sea-level rise based on actual observations of changing sea-level in the recent past and compared them with the IPCC's computer model-based predictions. 

The IPCC estimates are too conservative, they conclude, because their models have never been tested against past patterns of sea-level rise to check that they are accurate. 

The predictions used by the IPCC are based on a jigsaw puzzle of models for ice sheets, glaciers, and the warming of the sea. The predictions suffer, however, from the fact that only a limited amount of data are sometimes available for the models to be tested on. 

There were practically no data on the melt-off rate for Antarctica before we had coverage from satellite observations in the 1990s. 

“We have better historical data for the sea-level rise in total, which, in principle, allows for a test of the combined puzzle of models”, says  Aslak Grinsted, associate professor at the Niels Bohr Institute. 

“However, it has not been part of the routine to make sea-level hindcasts at IPCC, so presently we are not able to tell if these models are capable of reproducing the historical sea level. 

“At the Niels Bohr Institute, we have used this situation as our starting point, and so we observe how sensitive the models are in reacting to warming. 

“We expect that if we compare observational data from the fairly short period of time from 1850 onwards with the sensitivity of the models, it should allow us to assess whether the models are realistic or not,” Dr Grinsted says. 

The findings are detailed in an article for the journal Ocean Science.

“You could say that this article has two main messages: the scenarios we see before us now regarding sea-level rise are too conservative – the sea looks, using our method, to rise more than what is believed using the present method,” Dr Grinsted says.

“The other message is that research in this area can benefit from using our method to constrain sea-level models in the scenarios in the future.”