A new study in Nature Communications modelled six of the deadliest heatwaves on record and found something the field has been slow to acknowledge: every single one of them killed thousands of people while staying below the wet-bulb temperature long treated as the threshold between danger and death. Not one of the six events crossed the 35°C line. All six remained beneath it. All six produced mass mortality.
For roughly two decades, that single number has anchored scientific and policy discussions about the upper limits of human heat tolerance: 35 degrees Celsius on the wet-bulb thermometer. Above this threshold, the theory holds, the body’s primary cooling mechanism (the evaporation of sweat from the skin) becomes physically incapable of removing heat fast enough to prevent core temperature from rising to lethal levels. Given sufficient time above 35°C wet-bulb, any person without external cooling will die. This is not contested biology. The number has appeared in peer-reviewed climate projections, in government risk assessments, and in the background assumptions of infrastructure planning across Europe and beyond.
What the study published in Nature Communications has now established is that this number has been systematically misread, not as a physiological fact but as a policy threshold. The implication is precise, if uncomfortable: the threshold that climate science has treated as the boundary between danger and death is not where the dying actually begins. The dying begins much earlier. The 35°C line marks the theoretical extreme, not the practical edge.
What wet-bulb temperature actually measures
Wet-bulb temperature is not the same as the air temperature shown on a standard thermometer. It is a composite measurement that accounts for both heat and humidity, specifically the temperature a surface reaches when water evaporates from it. In practical terms, it approximates the cooling limit achievable by sweating: if the wet-bulb temperature is high, the air already contains enough moisture that evaporation slows and the cooling effect of perspiration diminishes. The higher the wet-bulb temperature, the less effective the body’s thermoregulation becomes.
The 35°C wet-bulb figure was established through laboratory research on young, healthy adults under controlled conditions. At that level, even a resting, unclothed individual with access to unlimited water cannot prevent their core temperature from rising. The figure is physiologically correct for those conditions. The problem, as the Nature Communications study makes clear, is that those conditions are not the conditions under which people die in actual heatwaves.
Real heatwave mortality is concentrated in populations who are not young, not healthy, not at rest, and frequently not in conditions where sweating is the primary challenge. The elderly lose thermoregulatory efficiency with age, independently of ambient conditions. People with cardiovascular disease face compounding risk as their hearts struggle to redirect blood flow for cooling. People living in buildings that retain daytime heat, which includes most urban social housing stock across Europe, experience effective wet-bulb exposures that far exceed what weather station readings suggest. The 35°C line was derived from a best-case scenario and then applied to a worst-case population.
The six events and what they showed
The heatwave events modelled in the research span geographies and decades but share a common characteristic: they are among the most lethal heat events in the historical record, accounting collectively for tens of thousands of confirmed excess deaths. The modelling reconstructed the biophysical experience of people exposed to those events using frameworks that account for clothing, activity level, building heat retention, solar radiation, and age-related reductions in thermoregulatory capacity, rather than just raw atmospheric readings.
In each case, the modelled conditions crossed the threshold for physiological unsustainability (the point at which the body cannot maintain safe core temperature) at wet-bulb readings substantially below 35°C. For the most vulnerable populations modelled, the proportion of six-hour windows crossing the survivability threshold ranged from roughly 12% of the event period in Mecca to 24% in Phoenix, in conditions the 35°C wet-bulb standard would have classified as safe. The people dying in these events were not exceeding the safety line. They were dying well short of it, in conditions the existing framework characterised as dangerous but survivable.
Every single event stayed below the line.
The researchers are careful to note that this does not invalidate the 35°C figure as a physiological ceiling. It invalidates the practice of treating that ceiling as the relevant planning threshold. What the data shows is that the effective danger zone, the zone in which population-level mortality risk rises sharply, begins well below the theoretical maximum, and that the gap between the two is where most heat-related deaths actually occur.
Why this matters for European policy
European governments have spent the past three years significantly revising their heat response frameworks, driven by the summer mortality data from 2022 and 2023. A study published in Nature Medicine estimated excess heat mortality across the continent at over 61,000 in summer 2022 alone, a figure the WHO has since cited in its own regional reporting. Countries including France, Spain, Portugal, and the United Kingdom have updated their heat emergency protocols, expanded early warning systems, and in some cases introduced mandatory cooling provisions for vulnerable populations.
The Nature Communications findings suggest that much of this policy revision may still be calibrated to the wrong baseline. If early warning systems are designed to trigger at thresholds derived from the 35°C ceiling, and the actual mortality risk rises sharply at significantly lower wet-bulb readings, the systems are providing less lead time than planners believe they are. The populations most at risk, the elderly, the chronically ill, people without access to cooling, may be experiencing dangerous conditions before the official response apparatus has been activated.
This has direct implications for how European cities think about heat as an urban infrastructure problem. The EU Mission on Adaptation to Climate Change has pushed member states toward heat resilience planning, but the specific thresholds embedded in those plans may need to be reconsidered in light of the new research. What counts as a heat emergency is, on the evidence of this study, a larger category than current frameworks acknowledge.
Implications for the technology and infrastructure sectors
For technology companies and the wider infrastructure sector, the study carries a set of practical implications that extend beyond public health policy. Data centres already face significant operational challenges during heat events due to cooling system demands and electricity grid stress. The study doesn’t address infrastructure planning directly, but its core finding, that the 35°C wet-bulb ceiling underestimates real-world risk, raises an open question for any contingency planning that uses that ceiling as a baseline: if the relevant danger threshold sits lower than assumed, the planning window for high-risk periods may be larger than current models suggest.
The same applies to logistics and cold chain infrastructure, urban mobility networks, and the broad category of outdoor-dependent operations (construction, agriculture, utilities maintenance) that employ large numbers of people across southern Europe. Occupational heat safety standards, which define permissible exposure limits for outdoor workers, often rely on the same kind of single-threshold thinking the study critiques. The paper doesn’t evaluate any specific occupational standard, but its findings suggest a reasonable question for regulators: whether exposure limits built around the 35°C ceiling are conservative enough for outdoor workers who are older, less able to access shade, or already managing chronic illness — the populations the study identifies as most at risk well below that line.
The threshold has been read as a safe boundary. The study suggests it was always an extreme endpoint, and that the gap between the two is where the actual risk lives.
What comes next
The Nature Communications paper is likely to prompt significant follow-up research, particularly around the identification of population-specific effective thresholds: the wet-bulb levels at which mortality risk becomes elevated for elderly adults, for people with specific comorbidities, and for people in different housing types. That research will take time to accumulate and to be translated into revised policy guidance. In the interim, the study’s core finding, that the 35°C line is an endpoint rather than a boundary, is available to policymakers, infrastructure planners, and regulators who choose to act on it.
The authors do not prescribe a replacement threshold. The data does not generate a single clean number that could be substituted for 35°C in every planning context. What it generates is a more honest picture of where heat kills, which is, consistently, in conditions that existing frameworks have characterised as manageable.
So which city goes first? Which mayor stands up before next summer and concedes that the cooling centres open too late, that the warning sirens trigger at a number derived from a laboratory ceiling no real heatwave has ever needed to cross? The six events in this study killed tens of thousands of people without ever touching 35°C. The threshold has already failed. The only remaining question is how many more summers planners will spend waiting for it.
