Artificial intelligence (AI) predictions reveal that Earth may exceed crucial climate targets much sooner than expected, with temperatures in some regions possibly rising by 3°C as early as 2060. This timeline is far ahead of earlier projections, signaling urgent climate challenges.
The 2015 Paris Climate Agreement aimed to keep global warming below 1.5°C above pre-industrial levels by the century’s end. However, new research by leading climate experts suggests that this target is now unattainable.
Using machine learning, researchers analyzed 10 advanced climate models, forecasting that 34 regions will likely exceed 1.5°C warming by 2040. Of these, most will hit 2°C warming by 2040, and 26 regions could surpass 3°C warming by 2060.
Regions like South Asia, the Mediterranean, Central Europe, and parts of sub-Saharan Africa are at the highest risk. These areas could face intensified threats to ecosystems and communities, the study warns.
In 2024, global temperatures are expected to set new records, surpassing pre-industrial levels—a stark reminder of how burning fossil fuels continues to drive climate change.
The report emphasized, “Rising temperatures will worsen extreme weather events, threaten water resources, harm food security, and affect ecosystems, livelihoods, and human health. If regional warming continues at this pace, the risks will become more severe in the next two to three decades.”
Advanced AI and Regional Insights
The study, published in Environmental Research Letters, was led by climate scientists Elizabeth Barnes (Colorado State University), Noah Diffenbaugh (Stanford University), and Sonia Seneviratne (ETH-Zurich). They used AI techniques like transfer learning to combine previous climate models, providing more precise regional forecasts.
According to Prof. Diffenbaugh, understanding local and regional climate changes is crucial since global averages often mask the extremes faced by vulnerable areas.
A second study, also released on Tuesday by Profs. Diffenbaugh and Barnes, highlighted that global heat records could soon be broken. Their AI-driven research suggests a 50% likelihood of surpassing 2°C warming, even if net-zero emissions targets are met.
AI Models Reveal Future Risks
For the second study, published in Geophysical Research Letters, researchers trained an AI system to assess future warming based on greenhouse gas data and the rate of decarbonization. If net-zero targets aren’t achieved by 2100, they found a 90% chance that future years will see temperatures far exceeding the hottest year of 2023. Some regions may experience heat three times worse than 2023 extremes.
Both studies build on earlier research by Profs. Diffenbaugh and Barnes, which predicted when Earth could breach the 1.5°C and 2°C warming thresholds.
“We’ve witnessed escalating climate impacts like heatwaves and heavy rainfall in recent years,” Prof. Diffenbaugh said. “This research shows that even under optimistic scenarios, conditions will worsen significantly.”
Prioritizing Climate Adaptation
The scientists stress the importance of ramping up investments in decarbonization and climate adaptation measures to protect against rising heat, droughts, and floods. Current policies, such as President Joe Biden’s Inflation Reduction Act, focus more on reducing emissions, while adaptation measures lag behind in funding.
