A new study published in the Proceedings of the National Academy of Sciences takes a fresh look on just how much clouds are impacting climate models. While conventional climate prediction models estimate cloud physics with numerical algorithms, these models often have to depend on imperfect assumptions about weather and climate. That means that their results aren’t always the most accurate when factoring in climate change and other variables.
"Clouds play a major role in the Earth's climate by transporting heat and moisture, reflecting and absorbing the sun's rays, trapping infrared heat rays and producing precipitation," said co-author Michael Pritchard. "But they can be as small as a few hundred meters, much tinier than a standard climate model grid resolution of 50 to 100 kilometers, so simulating them appropriately takes an enormous amount of computer power and time.”
“A lack of supercomputer power, or the wrong type, means that these [simulations] are still a long way off," Pritchard said. "Meanwhile, the field has to cope with huge margins of error on issues related to changes in future rainfall and how cloud changes will amplify or counteract global warming from greenhouse gas emissions."
In order to compensate for this, the researchers, who collaborated from the University of California, Irvine, the Ludwig Maximilian University of Munich and Columbia University, decided to put technology to work. Using deep machine learning, they took advantage of data to create computer algorithms called neural networks that mimic the human brain and can produce more accurate climate predictions. To do so, the neural networks utilize micro cloud-resolving models along with planetary-scale weather patterns in a fictitious ocean world to calculate cumulus. The scientists call the technology “The Cloud Brain.”
So far, The Cloud Brain has been able to accurately predict multiyear simulations of precipitation extremes and tropical waves. Lead author Stephan Rasp explains the significance of the neural network: "The neural network learned to approximately represent the fundamental physical constraints on the way clouds move heat and vapor around without being explicitly told to do so, and the work was done with a fraction of the processing power and time needed by the original cloud-modeling approach.”
The researchers hope to use this technology to improve data-driven climate and weather models, with specific focus on cumulus calculations for climate change impacts. "I'm super excited that it only took three simulated months of model output to train this neural network," Pritchard said. "You can do a lot more justice to cloud physics if you only need to simulate a hundred days of global atmosphere. Now that we know it's possible, it'll be interesting to see how this approach fares when deployed on some really rich training data."