Machine Learning Could Help to Improve Climate Forecasts

by Nicola Jones, Nature

As Earth-observing satellites become more plentiful and climate models more powerful, researchers who study global warming are facing a deluge of data. Some are now turning to the latest trend in artificial intelligence (AI) to help trawl through all the information, in the hope of discovering new climate patterns and improving forecasts.

“Climate is now a data problem,” says Claire Monteleoni, a computer scientist at George Washington University in Washington DC who has helped to pioneer the marriage of machine-learning techniques with climate science. In machine learning, AI systems improve in performance as the amount of data that they analyse grows. This approach is a natural fit for climate science: a single run of a high-resolution climate model can produce a petabyte of data, and the archive of climate data maintained by the UK Met Office, the national weather service, now holds about 45 petabytes of information—and adds 0.085 petabytes a day.

Researchers are combining artificial intelligence (AI) and climate science to create deep-learning analyses of weather patterns, and a September conference in Colorado will evaluate the state of climate informatics.

Because deep-learning systems develop their own rules, researchers often can’t say how or why these algorithms arrive at a given result. That makes some people uneasy about relying on these ‘black boxes’ to forecast imminent weather emergencies such as floods. “I’m reluctant to use [AI] as an answer machine,” says William Drew Collins, a climate modeller at the LBNL. “If I can’t explain what the machine is doing, then there’s a problem.” Instead, Collins says that AI algorithms are best suited to help test the next generation of climate models.

Last year, a team at the Lawrence Berkeley National Laboratory (LBNL) reported on the first use of a deep-learning system to identify tropical cyclones, atmospheric rivers, and weather fronts, demonstrating the algorithm could replicate human expertise. The researchers plan to apply similar methods to analyze a broader range of extreme weather events, including those as yet uncategorized. The team’s objective is to better rank and predict shifts in these phenomena related to climate change.

Similar work by George Washington University’s Claire Monteleoni has led to machine-learning algorithms that produce weighted averages of about 30 climate models used by the Intergovernmental Panel on Climate Change. LBNL’s William Drew Collins envisions AI algorithms being used to test next-generation climate models, and some scientists are using them for weather forecasts.  Read the report.

DCL:  They talk of having to analyze petabytes of data. This should be real-time analysis but I suspect it is actually “after-the-fact” analysis. Certainly the real challenge, particularly with fast moving natural disasters, is real time event analysis and prediction. These guys don’t seem to have started to tackle this yet – and they will need CEP!