Study Using AI Warns We Are 10 to 15 Years From Breaching 1.5°C Paris Agreement Goal
A new study using artificial intelligence (AI) to forecast global warming timelines has found that Earth will exceed the target of limiting warming to 1.5 degrees Celsius above pre-industrial levels within about 10 to 15 years, according to a press release from Stanford University.
Emissions goals designed to keep warming below the more ambitious threshold may be needed to keep our planet’s average temperature from exceeding two degrees Celsius, the study said. Less warming is still possible for the planet in the future, however.
“We have very clear evidence of the impact on different ecosystems from the 1C of global warming that’s already happened,” said Stanford University climate scientist Noah Diffenbaugh, who was co-author of the study, as The Guardian reported. “This new study, using a new method, adds to the evidence that we certainly will face continuing changes in climate that intensify the impacts we are already feeling.”
The study, “Data-driven predictions of the time remaining until critical global warming thresholds are reached,” was published in the journal Proceedings of the National Academy of Sciences.
A type of AI known as a neural network was used by the researchers to estimate a new “time to threshold” using recent global temperature data. The AI found that, no matter the fluctuations in greenhouse gas emissions in the next decade, the global warming trajectory shows we will reach 1.5 degrees Celsius in the early 2030s.
“Using an entirely new approach that relies on the current state of the climate system to make predictions about the future, we confirm that the world is on the cusp of crossing the 1.5 C threshold,” Diffenbaugh said in the press release.
AI predicted there is a 50 percent chance that our planet will heat up by an average of two degrees Celsius by mid-century if emissions continue to be high over the coming few decades, with a four-in-five chance by 2060.
It also predicted that Earth would warm up by two degrees Celsius even with a reduction of emissions.
“Our AI model is quite convinced that there has already been enough warming that 2 C is likely to be exceeded if reaching net-zero emissions takes another half century,” said Diffenbaugh, who co-authored the study with Colorado State University atmospheric scientist Elizabeth Barnes, in the press release.
If the thresholds of 1.5 and two degrees Celsius are breached, the goals of the 2015 Paris Agreement — a legally binding international treaty on climate change signed by 193 countries and the EU — will not have been met.
Every fraction of a degree of warming will ramp up the effects of climate change for humans, animals and ecosystems. As the Earth gets hotter, tipping points where large-scale consequences are felt — like forest die-offs and the melting of enormous polar ice sheets — become more likely, and scientists predict the severity of impacts will be more pronounced and widespread as our planet warms beyond two degrees Celsius.
Even where emissions are quickly reduced to net-zero by 2076, the AI predicted a 50 percent chance of crossing the two degree Celsius threshold by 2054, and a two-in-three chance of breaching it between 2044 and 2065.
With the net-zero goals of countries like China, India and the U.S., as well as the EU and non-state entities like universities, there is still a possibility that we can quickly reduce the greenhouse gases being pumped into Earth’s atmosphere and avoid the more severe impacts of climate change.
“Those net-zero pledges are often framed around achieving the Paris Agreement 1.5 C goal,” said Diffenbaugh in the press release. “Our results suggest that those ambitious pledges might be needed to avoid 2 C.”
To test how accurate the AI model was, the researchers asked it to predict the current 1.1 degree Celsius level of warming based on annual temperature anomaly data from 1980 to 2021. The model correctly forecast that the current level of warming would be reached this year, with a predicted range of 2017 to 2027.
It also correctly modeled the pace of decline in recent decades that led to 1.1 degrees Celsius.
“This was really the ‘acid test’ to see if the AI could predict the timing that we know has occurred,” Diffenbaugh said in the press release. “We were pretty skeptical that this method would work until we saw that result. The fact that the AI has such high accuracy increases my confidence in its predictions of future warming.”
Diffenbaugh said he hoped the results of the study will motivate humanity while there is still a sliver of time to avoid more severe effects and prepare for future consequences.
“Managing these risks effectively will require both greenhouse gas mitigation and adaptation. We are not adapted to the global warming that’s already happened and we certainly are not adapted to what is certain to be more global warming in the future,” Diffenbaugh said, as The Guardian reported.