If we pass safe climate limits, it’s a long way back

University of Victoria’s Andrew MacDougall in Canada’s Kluane National Park Credit: Nicolas Roux

If CO2 levels in the air pass the ‘safe’ limit, we’d have to take out up to four-fifths more than we originally emitted to get back under it. That’s the result from seemingly the first study to look at climate change’s reversibility with plausible scenarios, done by Andrew MacDougall from the University of Victoria (UVic), Canada. “With monumental effort and political will climate change is reversible within the millennium,” Andrew told me. “However, more carbon will need to be extracted from the atmosphere than was originally emitted to it. Meanwhile, changes in sea-level are effectively irreversible on the millennial time-scale.”

Andrew started looking at whether climate change could be undone in autumn 2012, after publishing a study showing that melting permafrost will speed up global warming. “The results were pretty grim,” Andrew said. “Combined with the failure of the political classes to implement controls on carbon emissions I began to wonder if there was a way to undo what humanity will do to the climate if we greatly exceeded the 450 parts per million (ppm) target.” That target comes because scientists say temperatures 2°C higher than the ‘pre-industrial’ average from 1850-1899 could become dangerous, and governments have agreed to keep warming below this level. Scientists also calculate that 450 CO2 molecules are allowable in every million air molecules to give us better than a 3/5 chance of temperature rises below 2°C.

After human emissions cease, current evidence suggests that natural processes would take tens of thousands of years to remove all of the fossil carbon from the atmosphere. Most of the warming will remain, even 10,000 years into the future. This sentence could be reduced by taking CO2 directly from the atmosphere, though this would be a huge effort, on the same scale as today’s fossil fuel industry according to one estimate. One method for doing that involves generating electricity by burning plants or trees that grew by absorbing CO2, and capturing and storing the CO2 from the burning. The other, known as air capture, uses machines to scrub CO2 right out of the air. However, this would need to be powered by clean energy and arguments over its cost are holding back research.

Slightly simpler climate model

One vision of technology to remove CO2 from the air involves giant scrubbers like this, where the greenhouse gas is removed by a chemical reaction. Credit: Carbon Engineering.

Andrew found that previous attempts to simulate using these methods in climate models weren’t looking at all the important parts of the climate system. They also didn’t remove the CO2 very realistically, with one even taking away all the extra CO2 in one giant gulp. So Andrew started to work out what it might realistically take to get back below 450 ppm alongside his main PhD studies, by himself. To do this, he turned to the UVic Earth-System Climate Model, a ‘climate model of intermediate complexity’.

“In most ways it is like other climate models, having a similar land surface model and ocean model, however, the UVic model has a simplified atmosphere,” Andrew explained. “The simplified atmosphere allows the model to run hundreds of times faster than a full climate model.” That low computational cost means UVic model users can look thousands of years into the future, and study big questions like Andrew’s. It also means they can develop and add in simulations of aspects of climate systems that other models ignore. In particular, these include the role of permafrost and sea level rise from loss of ice from ice-sheets in Greenland and the fact that the oceans expand as they get warmer.

“I began designing my own reversibility scenarios based on new scenarios used in the fifth assessment report of the UN’s Intergovernmental Panel on Climate Change (IPCC),” Andrew said. “This turned out to be a fairly straightforward model experiment to conduct. I used an existing version of the model and was able to conduct the experiments and write the paper in my spare time. To my knowledge my study is the first to look at reversibility with complete climate scenarios, including CO2, land use change, aerosols, methane and other non-CO2 greenhouse gases. Having the extra components is key to my study as permafrost and ice-sheets are expected to exhibit irreversibility and tipping point behaviour.”

Reversible, eventually

Change in surface air temperature (SAT) from the pre-industrial average (left) and change in CO2 levels in the air in billions of tonnes of carbon per year (PgCa^-1) (right) in the four scenarios Andrew used. The dotted lines are uncertainty limits. The numbers against each scenario refer to the ‘radiative forcing’ the amount of energy trapped by the CO2 emitted in the future it projects. Hence higher numbers refer to scenarios where more CO2 is emitted. Image copyright Wiley, used with permission.

In his experiment, published in Geophysical Research Letters on Monday, Andrew used four IPCC-based scenarios where CO2 in the air rose to different levels. He modified the original scenarios so that after CO2 levels and human land use rose to a peak, they fell as fast as they had risen. Other greenhouse gas emissions fell steadily while the cooling effect of pollution stopped immediately once CO2 levels began to decline.

In the scenario where CO2 peaks at the second lowest level, called MCP 4.5, that peak happens in 2130. Surface air temperature peaks two decades later, 2.8°C above the pre-industrial temperature, and is still 0.3°C above that level by 3000. None of the simulations show a full recovery to 19th century temperatures by the end of the 30^th century. But the sea level rise from melted ice sheets is much slower to recover. In MCP 4.5 Greenland ice sheet melt and the oceans expansion’s effect on sea level peaked in 2251, adding around 60cm to 1990 levels. By the year 3000 this had only fallen by around 30cm.

Sea level change from 1990 in the four scenarios Andrew used. Sea level rises both from melting ice, mainly from Greenland (left) and because water expands when it gets warmer (centre). Andrew’s modelling therefore looks at both these parts (right). The part from melting Greenland ice is slowest to recover. The dotted lines are uncertainty limits. Image copyright Wiley, used with permission.

The models also showed that greenhouse gas emissions from thawed permafrost will hamper efforts to restore CO2 levels to what they were before industrialisation. That means in total, we’ll have to take 15-80% more CO2 out of the air than we emitted to hit these levels. And even if we do take CO2 straight out of the air, we’re committing the generations that come after us to centuries of change. “Even under the scenario with the most aggressive reductions in emissions during the 21st century, removal of CO2 from the atmosphere would need to continue until at least the 24^th century,” Andrew stressed.

Andrew’s models show that to get back below safe levels once CO2 emissions have peaked would mean removing more CO2 from the air than we emitted to reach that level.

Journal references:

Andrew H. MacDougall, Christopher A. Avis, Andrew J. Weaver (2012). Significant contribution to climate warming from the permafrost carbon feedback Nature Geoscience DOI: 10.1038/ngeo1573
Andrew. H. MacDougall (2013). Reversing climate warming by artificial atmospheric carbon-dioxide removal: can a Holocene-like climate be restored? Geophysical Research Letters DOI: 10.1002/2013GL057467