You are here : Home > News > Current climate models may underestimate long-term climate change

Press release | Scientific result | Environment | Earth sciences

Current climate models may underestimate long-term climate change

Fifty-nine researchers from seventeen countries, including French specialists from the CNRS, Université de Bordeaux, Université PSL, the CEA, and UVSQ (*), have published an analysis of past climate periods published in Nature Geoscience. Climate variations in the past help us understand the implications of a rise of 2°C in global temperatures, and can be used to test our ability to simulate how the climate functions.

Published on 28 June 2018

​Climate conditions during certain periods in the past were at least as warm as now. Information revealed by the researchers' study is of interest for modeling global warming and how it may evolve.

An analysis of past climate intervals, undertaken by an international team of researchers from seventeen countries, is published in Nature Geoscience on June 25, 2018. The team concludes that global warming, even if it can be kept to no more than 2°C above pre-industrial levels, as recommended in the Paris Agreement, will cause rapid displacement of climate zones and their associated ecosystems. There will be significant reductions in the polar ice sheets lasting several thousand years. Rapid warming at the poles will release additional greenhouse gases, and sea level will rise by several meters in the course of the coming millennia. Observations also show that there is a risk that many current climate models, used to simulate changes over the 21st century, underestimate long-term changes.

A number of periods have been identified over the past 3.5 million years during which climate conditions were 0.5 to 2°C warmer than during the pre-industrial era. They show more significant warming at high latitudes than in tropical regions, which is similar to the results of simulations based on climate models for 2°C global warming by 2100. Although these warmer periods in the past were not all caused by increased levels of atmospheric CO2, analyzing them provides useful insights for studying the effects of warming comparable to the limit recommended in the Paris Agreement.

Ecosystems and climate zones will migrate

The study confirms that ecosystems and climate zones will generally migrate toward the poles, or to zones at higher altitudes. It also confirms that melting permafrost will release additional carbon dioxide and methane, which in turn will cause further warming. Observations of past warm intervals suggest that, if warming is kept below 2°C, as proposed in the Paris Agreement, the risk of catastrophic runaway warming linked to high levels of greenhouse gas emissions is relatively low. Nonetheless, even if that were the case, the additional CO2 released from the permafrost soils also has to be taken into account.

Long-term sea level rise of more than 6 meters

Even moderate warming within the range of 1.5° to 2°C above pre-industrial levels will be enough to cause substantial melting of the Greenland and Antarctic ice sheets in the long term, in turn causing a rise in sea level of more than 6 meters, which will last for thousands of years. It is therefore probable that sea level will rise at faster rates than has been the case in the last few decades.

Past warming intervals were warmer than shown by simulations obtained using climate models

Comparing data on the past with digital simulations suggests that climate models underestimate long-term global warming and the impact of polar amplification. While climate model predictions appear to be reliable insofar as regards moderate changes in scale over the next few decades, these models probably underestimate future climate change, especially in the case of long-term feedback forecasts for scenarios in which greenhouse gas emissions continue to increase. 

Top page