This website uses cookies. Read about how we use cookies.

Close window

Complex patterns of future local warming

Doubling ratios in the HadGEM2-ES model

January 2015 - Although human emissions of carbon dioxide (CO2) are known to be warming the planet, it is less clear how much climate change is to be expected over specific regions of the world, under different future levels of greenhouse gas emissions. This information is needed to allow people to better manage the risks of future weather and climate. This article reports on some new Met Office research revealing a finer scale, more nuanced understanding of what climate models are predicting for future regional warming.

Climate models project changes in climate for every location on the globe, covering a vast range of different conditions. To understand and interpret these projections for clear, actionable policy advice, it is necessary to distil out the most important and robust information from climate models. We don't yet know how high atmospheric CO2 levels will rise but views of the future without any policy to limit emissions indicate they could rise to more than 4 times their pre-industrial levels. In many studies of climate impacts it is assumed that a quadrupling of atmospheric carbon dioxide levels from pre-industrial values would produce twice the local warming that would occur from a doubling of the atmospheric CO2. In other words, a first and second doubling of CO2 are often assumed to cause the same levels of local warming. Our new research shows where and why this is not the case. 

What happens in different regions?

Gaining a clear picture of warming in different regions is technically challenging. The new Met Office study has used a carefully-formulated set of climate model experiments, designed to isolate the difference in warming from the first and second CO2 doublings. Results, from five different state-of-the-art climate models, are given as 'doubling ratios': the warming from the second CO2 doubling divided by that for the first doubling.  
Different land regions can show quite different doubling ratios: for some regions, slightly less warming is found for the second doubling than for the first (doubling ratio less than 1), whereas other regions have much greater warming from the second doubling (doubling ratio greater than 1). In the Met Office HadGEM2-ES model, the western Amazon sees 80% more warming from the second doubling (i.e almost twice as much), although this is an extreme example (values of 30% are more commonly found elsewhere). The five models broadly agree on the range of doubling ratios to expect across the globe (Figure 1), but only partly agree on where large doubling ratios are to be found (e.g. over the Amazon, doubling ratios are somewhat smaller in non-Met Office models).Figure 1. The range of doubling ratios found over land in each of the five climate models. Curves show the area of land (y-axis) for which the doubling ratio exceeds a given value (shown by the x-axis).

Underlying causes

It is important to understand why different regions show different doubling ratios - to help determine how realistic these results may be. There are three main physical processes that cause doubling ratios to vary from place to place. These involve behaviour of sea ice, the Atlantic Meridional Overturning Circulation (AMOC - which includes the 'Gulf Stream') and evaporation of water from land. Over North-Western Europe, the behaviour of the AMOC appears particularly important (Figure 2).  Current climate models suggest that the AMOC will slow as CO2 increases. Over North-Western Europe, this would offset some future climate warming. However, the AMOC tends to slow less (offsetting less warming over NW Europe) for a second CO2 doubling than for a first, although this effect is highly uncertain. These regional effects (from sea-ice, the AMOC and land evaporation) combine with global-scale processes to give large doubling ratios at some locations.Figure 2. The importance of the AMOC over Europe. The additional warming from a second CO2 doubling after global-scale processes are factored out is plotted against a prediction based on change in the AMOC alone. Symbols show five climate models studied.

Implications of our findings

Given effective action to reduce emissions, CO2 concentrations by 2100 may be approximately double pre-industrial levels. Under business-as-usual conditions, CO2 could be roughly quadrupled. Therefore, the response to a second CO2 doubling may roughly be viewed as the climate consequences of not taking action to reduce emissions of CO2.  These results show that the climate consequences of failing to reduce CO2 emissions involve additional, nonlinear effects that are unimportant for climate change under low CO2 conditions. In the debate on the benefits of reducing emissions, such effects need to be taken into account.

More information

More details are found in the article by Good et al. (2014), published in Nature Climate Change on January 26, 2015.

Last updated:

Follow us on

Facebook, Twitter, YouTube, Instagram, Snapchat, or LinkedIn Facebook Follow @metoffice on Twitter YouTube Instagram Snapchat LinkedIn