By Pesach Benson and Omer Novoselsky/TPS-IL
The blind spot could leave less time to prepare for increasingly dangerous heat events.
Even advanced climate models struggle to capture early atmospheric warning signals that precede extreme heatwaves in the Middle East, an Israeli study warns. This limitation could reduce lead time for forecasting and preparedness as the region faces increasingly frequent and severe heat events, the researchers said.
The study, conducted by a team from the Hebrew University of Jerusalem, suggests that while leading models can reproduce heatwave conditions once they are underway, they often fail to capture key physical processes that trigger them days in advance. This blind spot could leave less time to prepare for increasingly dangerous heat events.
The research was led by Andre Klif, Professor Chaim I. Garfinkel, Professor Dorita Rostkier-Edelstein and Dr. Assaf Hochman, one of the study’s lead researchers, who told The Press Service of Israel about the implications.
“Our results show that CMIP6 models [Coupled Model Intercomparison Project Phase 6], the latest generation of standardized global climate simulation models used in Intergovernmental Panel on Climate Change (IPCC) assessments, capture the mature stage of Eastern Mediterranean heatwaves reasonably well, but they systematically underestimate and delay the physical processes that initiate them,” Hochman explained.
“In particular, they delay the buildup of near-surface warming, upper-level ridging, horizontal warm-air advection and the associated tropical–extratropical teleconnections,” he said.
The team analyzed 11 state-of-the-art climate models used by the United Nations’ Intergovernmental Panel on Climate Change, comparing simulations with observed atmospheric data across one of the fastest-warming regions on Earth.
They found that major heatwaves in the Eastern Mediterranean are preceded by a complex chain of atmospheric processes stretching thousands of kilometers. These include circulation changes over Europe, Turkey, India and North Africa.
“These systems are not isolated events,” the researchers noted, adding that subtle shifts in high-pressure systems and wind patterns can begin up to a week before temperatures peak.
However, the models frequently delayed or weakened these precursor signals, suggesting that important early-warning dynamics are not fully represented in current climate simulations.
Blind spots
Researchers also identified a key role for a strengthening high-pressure ridge over Turkey, which, when properly represented, improved model accuracy in simulating heatwave intensity.
Another important finding was a long-distance atmospheric link between the South Asian monsoon and Eastern Mediterranean heatwaves, which none of the examined models successfully captured.
“We do not quantify a specific increase in forecasting lead time,” Hochman said. “However, the missing precursor signals appear up to seven–10 days before heatwave onset in the observations, while the models only converge toward the observed evolution during the final two–three days before the event.
“This suggests that improving the representation of these physical processes has the potential to extend useful sub-seasonal predictability, although the exact gain should be quantified in dedicated forecasting studies.”
He added that the findings have implications for both short-term forecasting and long-term projections.
“While many CMIP6 models reproduce historical heatwave statistics reasonably well, our study shows that they often do so for the wrong physical reasons, with compensating errors masking deficiencies in the underlying dynamics. This means that confidence in future climate projections should depend not only on whether models reproduce the observed warming, but also on whether they correctly simulate the physical mechanisms responsible for heatwave development,” he said.
The researchers propose a process-based framework for evaluating climate models, focusing on whether they reproduce the atmospheric dynamics behind extreme events rather than temperature outcomes alone.
Hochman said the next step is to refine these diagnostics and integrate them into model development and prediction systems.
The study was published in the peer-reviewed Weather and Climate Extremes journal.
