A striking new study has raised eyebrows this week with its alarming conclusions about a possible consequence of future climate change.
Under an extreme climate change scenario, the study found that huge tracts of stratocumulus clouds in the Earth’s atmosphere—which help to reflect sunlight away from the planet and cool the climate—could disintegrate.
If that happens, global temperatures could skyrocket by 8 degrees Celsius, or more than 14 degrees Fahrenheit, the study suggests. And that’s on top of the global warming that would have already occurred by that point.
It’s a jaw-dropping possibility, one that would likely leave large expanses of the planet unrecognizable.
Still, there are a few important caveats.
First, the circumstances required to cause such an event are fairly extreme as far as climate scenarios go—although not impossible. The study finds that the cloud breakup would probably start to occur once atmospheric carbon dioxide levels reach about 1,200 parts per million, or triple their current levels.
But scientists generally suggest that under a business-as-usual climate scenario, in which no action is taken to curb global greenhouse gas emissions in the future, CO2 levels will probably be approaching 1,000 ppm around the end of the century.
The new study also seems to imply that the 1,200 ppm represents a kind of sudden tipping point, at which the clouds will all begin to break up at once. Other scientists have questioned this result in comments to the media this week. They say that while clouds may indeed begin to break up in response to high levels of CO2, it may happen at different times and speeds around the world.
What may be most important is the attention the study has drawn to the issue of clouds and climate, which is becoming one of the biggest priorities in climate modeling.
Scientists increasingly suggest that clouds may be among the most important—although also some of the most complex—regulators of the global climate. Depending on local conditions, clouds may enhance warming by trapping heat, or they may help cool the climate by reflecting sunlight back into space. The broad stratocumulus cloud layers examined in this week’s new study, for instance, have an overall global cooling effect.
But clouds are notoriously difficult to model, even on a small scale. There are many factors that affect where and when they form and how big they grow—and it’s especially hard to simulate their behavior all over the globe.
It’s even more difficult to project their responses to future climate change. Global warming may affect a lot of aspects of the atmosphere that can change the behavior of clouds, from the temperature and moisture in the air to changes in global wind patterns.
So finding better ways to capture clouds in climate models is one of the fastest-growing priorities among climate scientists. The new study represents one approach to the problem, using a technique known as a “large-eddy simulation.” It models the behavior of tiny particles and other fine details that affect the formation of individual clouds, which regular climate models have difficulty capturing. The study conducted an eddy simulation modeling the formation of clouds over one specific patch of the ocean, and then extrapolated those results up to a global scale.
Large-eddy simulations are currently one of the most useful ways to model the physics of individual clouds. But while they’re improving, scientists can still only run the models at relatively small scales. They can’t reproduce these fine physics in a global-scale model.
So scientists are working to develop even more cutting-edge approaches. And the use of artificial intelligence may be leading the way.
Climate scientist Tapio Schneider of the California Institute of Technology, lead author of the new cloud study, is also the principal investigator on a new project known as the Climate Modeling Alliance, a consortium of researchers working to build an earth system model that represents fine features like the behavior of clouds.
Dubbed the “Climate Machine,” the model will incorporate machine learning, which will allow it to use observational data and fine simulation techniques, like eddy simulations, to continuously improve itself over time. The project just kicked off last year.
Other researchers are on a similar track. Scientists from Columbia University; the University of California, Irvine; and the Ludwig Maximilian University of Munich are working on using deep learning—a kind of machine learning method—to try to better represent clouds in large-scale climate models. The “Cloud Brain,” as they call their project, involves a neural network that learns to predict the outcomes of models that specifically simulate clouds. This technique can then be used to represent cloud behavior in larger-scale models, the researchers say.
Finding better ways to represent the behavior of clouds in global climate models may be critical for making accurate projections about future climate change, scientists increasingly warn.
Trouble simulating even the specific stratocumulus clouds in this week’s study “percolate into the global climate response,” the researchers wrote. “[U]ncertainties in the response of stratocumulus and other low clouds lead to large uncertainties in the predictions of global temperatures and climate impacts.”
Reprinted from Climatewire with permission from E&E News. E&E provides daily coverage of essential energy and environmental news at www.eenews.net.