Time scales of atmospheric circulation response to CO2 forcing

By Paulo Ceppi

An important question in current climate change research is, how will atmospheric circulation change as the climate warms? When simulating future climate scenarios, models commonly predict a shift of the midlatitude circulation to higher latitudes in both hemispheres – generally referred to as a “poleward circulation shift”. As an example, under a “business as usual” future emissions scenario the North Atlantic jet stream is predicted to shift northward during the summer months (Figure 1). If true, this would likely affect the average amount of precipitation, wind, and sunshine experienced in the UK and more generally across Western Europe.


Figure 1: Change in eastward wind speed at 850 hPa in the RCP8.5 (“business as usual”) experiment during the 21st century for June-August. The response is calculated as the mean of 2070-2100 minus 1960-1990. Grey contours indicate the wind climatology, while colour shading denotes the change (in m/s). Results are averages over 35 coupled climate models.

Such shifts of circulation being caused by global warming, it is natural to assume that the more the planet warms, the larger the circulation shift. But is that assumption generally true? More specifically: as the planet warms in response to CO2 forcing, do circulation shifts scale with the change in global-mean temperature? Here we are interested in the time evolution of the transient response to CO2 forcing, i.e. the period during which the climate adjusts to the change in CO2. This evolution is best represented in climate model experiments in which CO2 concentrations are increased abruptly and then held constant; since the forcing happens all at once, the various time scales of climate response are cleanly separated. Below I will present results from the so-called “abrupt4xCO2” experiment, in which a set of climate models were subjected to a sudden quadrupling of CO2 concentrations and then run for 150 years.

It turns out that as climate changes following a sudden quadrupling of CO2, circulation shifts do not generally scale with global warming. Instead, two distinct phases of circulation change occur: during the first 5 years or so, the planet warms quickly and the jet streams shift poleward; but thereafter the jets tend to stay at a constant latitude, despite the fact that the planet continues to warm substantially. This is summarised in Figure 2 below, where the curves indicate changes in the latitude of the jet stream (averaged over a set of 28 climate models). In the North Pacific region, the change in jet stream latitude even changes sign over the course of the experiment (Figure 2b).


Figure 2: Change in annual-mean jet stream latitude (measured as the latitude of peak eastward wind at 850 hPa) in climate models during the first 140 years following a quadrupling of atmospheric CO2, as a function of global-mean surface temperature. The curves indicate means across 28 climate models. Shading denotes the 75% range of responses. The jet shifts are in degrees latitude and positive anomalies are defined as poleward. Circles denote individual years until year 10; diamonds denote decadal means between year 11 and year 140. The black crosses indicate the means of years 5-10 and 121-140, respectively.

How can we explain this peculiar time evolution of circulation shifts? The evolution of changes in atmospheric temperature and circulation is mainly controlled by how the ocean surface warms in response to greenhouse gas forcing. Like the atmosphere, the ocean has its own circulation and in some regions deep, cold water rises to the surface – a process known as “upwelling”. Due to this and other processes, the ocean surface does not warm at the same rate everywhere; in particular, upwelling regions like the Southern Ocean experience delayed warming.

We find that the time scales of ocean surface warming determine the time scales of change in atmospheric circulation, via the changes in atmospheric temperature. In particular, the patterns of ocean surface warming before and after year 5 of the experiment are strikingly different; when imposed in an atmospheric climate model, we obtain circulation changes consistent with the results shown in Figure 2, confirming the role of the ocean surface in controlling the atmospheric response.

While the scenario of abrupt CO2 increase described in Figure 2 is unlikely to happen in the real world, further analysis shows that the two time scales of circulation shift are also present in more realistic scenarios of gradual greenhouse gas increase. This indicates that care must be taken when extrapolating transient circulation shifts to estimate changes in future warmer climates.


Ceppi et al., 2017. Fast and slow components of the extratropical atmospheric circulation response to CO2 forcing: submitted to Journal of Climate.

Zappa et al., 2015. Improving climate change detection through optimal seasonal averaging: the case of the North Atlantic jet and European precipitation: Journal of Climate, doi: 10.1175/JCLI-D-14-00823.1

This entry was posted in Atmospheric chemistry, Climate, Climate change, Climate modelling, Numerical modelling and tagged . Bookmark the permalink.

Leave a Reply

Your email address will not be published. Required fields are marked *