Sensitive climate shows model shortcomings

Antarctic ice sheet

A study published tomorrow in the journal Nature Geoscience shows that three to five million years ago, during the mid-Pliocene warm period, global temperatures were significantly higher than can be accounted for by levels of atmospheric carbon dioxide at the time. The implication is that long-term climate sensitivity must be taken into account when looking at climate change in the age referred to informally as the Anthropocene.

University of Bristol physical geographer Dan Lunt and his co-workers combine results from a global climate model with a reconstruction of the Earth’s environment during the mid-Pliocene, when temperatures were on average three to five degrees warmer than today.

What the researchers found is that atmospheric carbon dioxide concentrations caused more warming than would be expected from the estimates of climate sensitivity used for the future climate projections we hear about so much in the news today. The conclusion is that, in the mid-Pliocene, the Earth system had a long time to adjust to changing environmental conditions, whereas current climate models tend not to include feedbacks that result from slowly changing parts of the system, such as shifting ice sheets and vegetation patterns.

There is a good reason why ice sheet dynamics have so far not been included in the climate models that inform the work of official bodies such as the IPCC. Ice dynamics remain relatively little understood. To date, the IPCC has focused on the so-called Charney sensitivity, named after the late Jule Charney, who in 1979 chaired a committee of the US National Academy of Sciences tasked with investigating anthropogenic global warming.

Following this methodology, research groups have used Charney equilibrium scenarios to determine the degree of greenhouse gas emissions likely to lead to levels of climate change deemed dangerous. The new study suggests that the equilibrium climate change associated with an increase in atmospheric carbon dioxide is likely to be significantly larger than has been estimated using Charney sensitivity-based models.

Timescales are all-important. Vegetation, for example, can take several centuries to reach an equilibrium state, whereas ice sheets may take millennia. That said, evidence coming in from observational data and models shows that ice sheets may reach equilibrium much faster than this, especially when surface melt water enters crevasses, sinks to the bottom of the sheets and decreases the the amount of friction holding them in place.

Lunt and his colleagues say that, given uncertainties in the timescale for vegetation and ice-sheet responses to climate change, estimates of the impacts of long-term greenhouse-gas stabilisation should focus on Earth system sensitivity rather than the traditional Charney sensitivity.

As it is, the omission of ice dynamics from climate models is likely to result in over-conservative estimates of temperature rises resulting from greenhouse gas emissions, owing to the positive feedback mechanisms involved. As we learn more about the behaviour of ice sheets, their influence on climate should be incorporated into climate models.

Further reading

Lunt et al., “Earth system sensitivity inferred from Pliocene modelling and data”, Nature Geoscience (2009)