AI-enhanced research shows regional warming will exceed critical thresholds faster than expected, with most regions exceeding 1.5°C by 2040. Fragile regions like South Asia face greater risks and require rapid adaptation action.
Three leading climate scientists analyzed data from 10 global climate models, using artificial intelligence (AI) to improve accuracy. Their results suggest that regional warming thresholds are likely to be reached sooner than previously estimated.
The study, published in IOP Publishing's Environmental Research Letters, predicts that most land areas are likely to exceed the 1.5°C warming threshold as defined by the Intergovernmental Panel on Climate Change (IPCC) by 2040 or earlier. Additionally, temperatures in some areas are expected to exceed the 3.0°C threshold by 2060, significantly earlier than previously estimated.
Regions including parts of South Asia, the Mediterranean, central Europe and sub-Saharan Africa are expected to reach these thresholds faster, exacerbating risks to fragile ecosystems and communities.
The study, by Colorado State University professor Elizabeth Barnes, Stanford University professor Noah Diffenbaugh, and ETH Zurich professor Sonia Seneviratne, used a cutting-edge artificial intelligence transfer learning method that integrates knowledge from multiple climate models and observational data to refine previous estimates and provide more accurate regional forecasts.
Main findings
The researchers used artificial intelligence-based transfer learning to analyze data from 10 different climate models to predict rising temperatures and found:
By 2040, 34 regions are likely to experience warming exceeding 1.5°C.
By 2040, 31 of these 34 regions are expected to experience 2°C warming.
By 2060, 26 of these 34 regions are expected to experience warming exceeding 3°C.
Elizabeth Barnes said: "Our research highlights the importance of incorporating innovative artificial intelligence techniques, such as transfer learning, into climate modeling, with the potential to improve and constrain regional forecasts and provide actionable insights to policymakers, scientists and communities around the world.
Noah Diefenbaugh, a professor at Stanford University and a co-author of the study, added: "It is important to focus not only on rising global temperatures but also on specific changes occurring locally and regionally. By limiting when regional warming thresholds are reached, we can more clearly predict specific impacts on societies and ecosystems." The challenge for research is that regional climate change is likely to be more uncertain, both because the climate system is inherently noisier at smaller spatial scales and because changing processes in the atmosphere, oceans, and land surfaces create uncertainty about how exactly specific regions will respond to global-scale warming."
Compiled from /ScitechDaily