How do climate feedbacks affect model projections?

As greenhouse gas emissions build up in the atmosphere, less radiation is emitted to space and the Earth’s temperature increases until the energy balance in its climate system is restored. Most of the excess energy is stored in the ocean, but some also heats the atmosphere and land, leading to melting of snow and sea ice. There are also changes in clouds and water vapour, plus other processes that can create further changes in the Earth’s energy balance.

These “climate feedbacks” generally strengthen climate change, but their influence can vary over the globe and through time, in what is often referred to as the “pattern effect”.  Gaining a better understanding of feedback patterns, and their implications, is important for improving the climate model projections that help us to plan for what lies ahead.

Recent CONSTRAIN research, led by Dr Lawrence Jackson from the University of Leeds, used a climate emulator – a simple climate model that can run hundreds of simulations at a fraction of the cost of more complex models – to investigate how the pattern effect and associated changes in climate feedbacks affects both temperature and rainfall projections across the globe.

Keeping the feedback broadly constant meant that the emulator projections followed those of more complex climate models which do not explicitly vary the climate feedback.  Alternatively, plugging in different strengths for climate feedback resulted in differences in global temperatures of up to 1°C by 2050.  This is important for climate policy: with global temperatures already more than 1°C above pre-industrial levels[i], opportunities to keep warming within the Paris Agreement 1.5°C ambition could be even narrower than we think.

When it comes to adaptation, the rate of warming can be as important as the total temperature change.  Using the emulator found that variations in feedbacks could affect the rate of warming by up to 50% in coming decades, with strong negative feedbacks decreasing warming rates, and weaker feedbacks increasing them.

The impact on temperatures was also found to vary regionally: the eastern tropical Pacific, Southern Ocean and the far North Atlantic all seem to be particularly affected, perhaps through changes in the heat taken up by the oceans and changes in the solar radiation reflected back to space by low-level marine clouds.

Meanwhile, as well as temperature, changing feedbacks also appear to affect rainfall patterns: global rainfall is currently projected to increase, but continued strong negative feedbacks could reduce the amount of increase and also change the pattern of rainfall over the globe. Some regions, such as the Amazon and equatorial Pacific, appear to be particularly sensitive to changing feedbacks.

This work, which builds on other CONSTRAIN research into patterns of climate change across the globe, is a first step in developing a climate emulator that can better reflect the effects of climate feedback variations across the globe and through time.  Further work will provide greater clarity on how temperatures and rainfall will respond, both globally and regionally, to future greenhouse gas emissions within a complex and changing climate system.


Projections of global mean temperature for one representative climate model during the historical period and, from 2015, under SSP2-4.5. The projection with fixed feedback (in blue) closely follows the complex model (in black). Projections with varying feedbacks (in orange and green) show the sensitivity of global temperatures to variations in climate feedbacks