Three prominent climate scientists used Artificial Intelligence to analyze the results of ten global climate models, and their forecasts are even more alarming than those drawn by humans, surpassing in several regions, including the Mediterranean, the warming thresholds that would trigger extreme weather events.
The study, published in Environmental Research Letters by IOP Publishing, predicts that most land regions, according to the definition of the Intergovernmental Panel on Climate Change (IPCC), will surpass the critical threshold of 1.5°C by 2040 or earlier; and several of them are on track to exceed 3°C by 2060, whereas the IPCC assessment report, based on over 30,000 scientific studies by 270 authors from 67 countries, estimated that following the current trend would lead to a temperature rise between 2.3 and 2.7°C by 2100.
According to AI, the worst-case scenario is reserved for southern Asia, the Mediterranean, central Europe, and parts of sub-Saharan Africa, exacerbating risks for ecosystems and vulnerable communities.
According to the latest IPCC report, if warming exceeds 1.5°C, we will witness extreme phenomena such as severe storms, heatwaves, longer droughts, heavier rainfall, and massive forest die-offs. Additionally, many glaciers will completely disappear or lose most of their mass.
With a 2°C warming, food security in sub-Saharan Africa, South Asia, or Central America will be significantly affected. Between 800 to 3,000 million people worldwide will suffer from chronic water scarcity. And although the first species extinctions have already been confirmed, the risk would be 10 times higher with a 3°C increase.
The research, conducted by Elizabeth Barnes, a professor at Colorado State University; Noah Diffenbaugh, a professor at Stanford University; and Sonia Seneviratne, a professor at ETH-Zurich, used a cutting-edge artificial intelligence learning approach that integrates knowledge from multiple climate models and observations to refine previous estimates and provide more precise regional predictions.
According to AI results, 34 regions on the planet could exceed 1.5°C of warming by 2040; and in 31 of them, 2°C will be reached; while by 2060, 26 regions will surpass 3°C.
"Our research highlights the importance of incorporating innovative AI techniques, such as transfer learning, to improve and potentially constrain regional forecasts, and provide valuable information for policymakers, scientists, and communities worldwide," notes Elizabeth Barnes.
On the other hand, Noah Diffenbaugh, co-author and professor at Stanford University, points out that "by limiting the timing of when regional warming thresholds will be reached, the specific impacts on society and ecosystems can be anticipated more clearly. The challenge is that regional climate change can be more uncertain, both because the climate system is more intricate at smaller spatial scales and because processes in the atmosphere, ocean, and land surface create uncertainty about how a specific region will respond exactly to global-scale warming."