Climate sensitivity on the rise?

Climate sensitivity – the global temperature response to a doubling of atmospheric carbon dioxide – is a useful measure of the sensitivity of the climate system to a given forcing.  It is the result of various climate feedbacks in the system, such as changes in cloudiness and atmospheric temperatures.  Yet the impact of these feedback processes – and so climate sensitivity – is difficult to quantify, which in turn leads to large uncertainties in our projections of future climate change.

Recent work has shown that climate feedbacks depend on the pattern of surface warming.  This matters, because if the past warming trend is different to what we might expect in the future, then climate sensitivity, estimated from the observed historical record, may not apply to the future. A new study, led by Dr Timothy Andrews of the Met Office Hadley Centre, along with an international team of collaborators, investigated this and found that estimates of climate sensitivity based on observations are likely to be biased low, and the real climate sensitivity may be larger and more uncertain than first thought.

Patterns of temperature change

Recent decades have seen cooling over the eastern tropical Pacific and Southern Ocean while temperatures rise globally.  Climate model simulations do not show these cooling features in the long-term response to sustained increases in atmospheric carbon dioxide.  This could lead to a difference in climate sensitivity between the past and future, since cloudiness – a key climate feedback – depends on the warming pattern in these regions.

To investigate this, Dr Timothy Andrews and his collaborators used a suite of world leading climate models forced with observed historical temperature changes. They used this to look at the models’ climate feedbacks and sensitivity to observed temperature patterns in contrast to their response to long-term carbon dioxide changes.

The study showed that when climate models are forced with observed temperature patterns they robustly simulate a smaller climate sensitivity than for the long-term response to increased carbon dioxide.

Climate sensitivity

In light of this, the study puts forward a way to account for this ‘pattern effect’ when trying to estimate climate sensitivity from historical temperature and energy budget changes. It revealed that climate sensitivity might be larger and more uncertain than we thought.  For example, it increases the long-term climate sensitivity (Equilibrium Climate Sensitivity or ECS) median estimate and 5-95% confidence interval from 1.9 °C (0.9-5.0 °C ) to 3.2 °C (1.5-8.1 °C) in one commonly used dataset.

Ways forward

The results of this study put a spotlight on our understanding of the causes of regional temperature change and how they evolve with time.  In particular, how credible are the patterns of temperature change simulated by climate models in response to long-term carbon dioxide increases, and why and where are they different to what we have observed in the past?  Further progress in constraining climate sensitivity could come from improved understanding of the causes of surface temperature change patterns in observations and climate model projections.

The paper, Accounting for changing temperature patterns increases historical estimates of climate sensitivity (doi: 10.1029/2018GL078887) is published in Geophysical Research Letters.