Are climate change effects here already? - FreeAstroScience
A billion more people meet extreme heat each year than in the 1970s, but how much of that billion is the heat?Welcome to FreeAstroScience. Most writing about climate change effects predicts them, and this. . .
A billion more people meet extreme heat each year than in the 1970s, but how much of that billion is the heat?Welcome to FreeAstroScience. Most writing about climate change effects predicts them, and this piece counts the ones that already landed across every land cell on Earth between 1950 and 2024. The pair of numbers that stopped us was 0. 32 against 0. 27. Climate change effects on heat are now measured rather than projected, and for most severity thresholds the warming itself drives at least as much of the rise in human exposure as population growth does. The hottest nights of the year are warming at 0. 32 °C per decade, faster than the hottest days at 0. 27 °C. Roughly one billion more people meet at least one day of extreme heat stress annually than in the 1970s, on a feels-like scale where extreme begins at 46 °C. In 2003 Europe had its great heat wave, and Carlo Buontempo remembers what his colleagues were saying at the time. An extraordinary event now, they told each other, but by 2020 or 2030 a more normal one. Buontempo directs the Copernicus Climate Change Service today, and he gave la Repubblica that line this month as his answer to what the summer of 2026 taught us. Twenty-three years on, the record is finally long enough to say how right they were. Rebecca Emerton and three colleagues at the European Centre for Medium-Range Weather Forecasts haveput 75 years of feels-like temperature through that question, land grid cell by land grid cell. What comes back is a ledger of climate change effects that have already happened, on every continent at once. How much faster are the nights warming than the days?Start with the instrument, because it is not a thermometer. The Universal Thermal Climate Index is a feels-like temperature that folds humidity, wind and radiation in with the air. Values come out ofERA5-HEAT, a reanalysis that blends real observations with a weather model to fill the gaps where direct data is thin. Coverage runs hourly from 1940 onward at a quarter-degree resolution. Take the ten hottest days of each year and the ten hottest nights, average them across every land cell outside Antarctica, and fit a line from 1970 to 2024. Days give 0. 27 °C per decade. Nights give 0. 32, with the same uncertainty of 0. 05 °C on both and significance better than one in a thousand. Put plainly, that is a bit over a quarter of a degree every ten years on the worst days, and roughly a third of a degree on the worst nights. The night is where the margin is. Five hundredths of a degree per decade reads like a rounding error, so we did the multiplication. Over the five decades since the 1970s the hottest nights have gained about a quarter of a degree more than the hottest days, and roughly 1. 6 °C in total. That matters because of what a hot night does to a body. High nighttime temperatures limit recovery, and the mortality work the authors lean on links hot nights to deaths directly. A tropical night is one whose minimum never falls to 20 °C, and the overnight cooling window simply does not open. We went through that mechanism inour June piece on why tropical nights are Europe’s deadliest heat risk, which stayed inside one continent and one summer. The paper in front of us drew no such boundary around continent or season. 2026 has tracked above the 1991 to 2020 average all year while staying under the 2 degree line, and the gray band behind it is every year since 1940. Graphic by Dr Zack Labe from Copernicus Climate Change Service ERA5 data, current to 18 July 2026. Climate change effects you can put on a calendarAveraged across the northern extratropics, the belt lying outside the tropics, the first and last days carrying moderate heat stress now sit 15 days further apart than they did in the 1970s. Strong heat stress gains 12 days, very strong 9, extreme 6. May is doing June’s work now. Table 1 — days added to the Northern Hemisphere heat stress season, last decade against the 1970s (Emerton et al. ,Nature Climate Change, 2026)Heat stress levelUTCI thresholdDays addedModerate26 °C15Strong32 °C12Very strong38 °C9Extreme46 °C6Europe makes the shift easy to picture. Moderate heat stress now begins in mid-May rather than early June and persists until nearly October, and strong heat stress starts in June instead of July. Africa records the largest single change of any continent, with an extreme heat stress season 28 days longer. Its shortest such season of the past ten years still runs 11 days past the longest one the 1970s produced. Watch that season-stretching on this year’s calendar. British summer 2026 arrived a month early: on 25 May the reading at Kew Gardens came within a fraction of 35 °C, the highest spring temperature ever recorded in the United Kingdom. By the end of July more than half of England had been declared officially in drought. Four consecutive heat waves have hit northwestern Europe and the Mediterranean since 24 May, with peaks between 45 and 46 °C in France, Spain and Portugal. Where the water went is its own story. On 28 June the German village of Coschen, on the Polish border, hit 41. 7 °C. In stretches of Hungary and Romania the Danube’s flow fell to a third of average, which forced two nuclear plants to shut reactors, and EDF stopped three more in the Ardennes. Buontempo names the drought itself as the aggravating factor. When there is no water, all the available heat goes into the air, which is the mechanism we traced inour account of the May 2026 heat dome over Europe. Repubblica’s map for 6 August puts Bologna at 38 °C and Ancona at 31. Rimini sits on the coast between them, which is where this article was written. The sea hit 33 degrees, and the grid stops at the coastThat coast is having its own summer. At about four on Wednesday afternoon a buoy off Sa Dragonera, in Majorca, read 33. 02 °C at the surface. If the reading is confirmed it beats the 31. 96 °C measured off Egypt in 2024, in a basin that normally runs 25 to 26 degrees in early August. Federico Grazzini runs the Arpae Emilia-Romagna operations room and trained as a physicist. He told Giacomo Talignani that the peaks are not the part that worries him. Marine heat waves, meaning long stretches of anomalously warm sea, have climbed without pause since 2022. His own measure of this summer is not a number. He swam in Liguria and at Nice and found no real difference between water and air. One detail in his account is worth keeping. Something similar probably happened at Capo Mele on the Ligurian coast, and nobody can prove it, because the wave buoy there was configured to read no higher than 29. 9 degrees. It has since been replaced with one that reaches 35. Whoever set up the old one did not expect the sea to do this. ISPRA’s network read 30. 5 at Anzio, 30 at Venice and 29. 3 at Trieste that same afternoon, and warm water has turned green with algae at Jesolo and Follonica. Copernicus has certified anomalies here reaching 8 degrees above average, on a basin mean of 27 in early August. Grazzini’s worry runs forward. Water near 30 degrees is a reservoir. While the high pressure holds, nothing comes of it. When wind and cooler air arrive, that stored energy feeds instability and evaporation, and evaporation is what clouds burn. He expects violent weather this autumn and does not think Italy is ready. None of it appears in the paper. Emerton and colleagues excluded ocean regions from every figure and table, so their grid stops where the water starts. What a Mediterranean at 30 degrees does to the overnight minimum on the shore is a question no source in front of us answers. What does extreme heat stress actually mean on a thermometer?The index sorts thermal stress into bands, set at feels-like values rather than air temperatures, which is why they look higher than a forecast. Moderate heat stress starts at 26 °C. Strong heat stress starts at 32 °C, the band counted whenever the paper says heat stress day. Very strong starts at 38 °C. Extreme starts at 46 °C, the point at which, on the health verification work Di Napoli and colleagues published in 2019, urgent action is needed to prevent severe harm. Measured against the 1970s, extreme heat stress now occurs two and a half times as often in Europe and South America, and twice as often in North America. Africa, Oceania and Asia come in at 1. 8, 1. 7 and 1. 2. Those multipliers reward the places that started cool, and Europe is the clearest example. Extreme heat stress there still touches 0. 2 percent of days and locations, a 138 percent rise on the 0. 08 percent of the 1970s and still almost nothing. Africa’s went from 4. 1 percent to 7. 4, the highest frequency on any continent. Widen the band to strong heat stress and the gap stops being subtle: 70 percent of days and places in Africa over the past decade, against 63 in South America, 55 in Oceania, 24 in Asia, 15 in North America and 9 in Europe. Africa carries this, not Europe. Heat is alreadythe leading cause of weather-related death worldwide, and the continent bearing most of the exposure is the one appearing least often in the coverage. The ocean carries its own version of this article’s argument: by 18 July 2026 the red line had climbed above both 2023 and 2024, and 2024 is the warmest year on record. Graphic by Dr Zack Labe from Copernicus Climate Change Service ERA5 data, region 60 south to 60 north. Where the biggest numbers are also the weakestTwo things here deserve a second look, and neither is hidden. Both sit in the extended data, which is where the interesting arguments usually live. One is a table of multiplication factors for the IPCC’s reference regions, and some are spectacular. Central North America comes out at 48. 5 times more extreme heat stress than in the 1970s. The South American monsoon region reaches 87. 4. Read across to the significance marks and both carry the symbol for p greater than 0. 05, meaning the result is not statistically significant. Both marks are printed in their own table, to their credit. A press release would quote the 87 and drop it. Read the significance marks first. The other is a number that does not match itself. Emerton and colleagues write that Africa’s longest compound events in the Southern Hemisphere now occur 2. 8 times more often than in the 1970s. Those are the runs of 181 to 270 days. Extended Data Table 2 puts that bin at 1. 85 and hands 2. 84 to the next one up, 271 to 365 days. One of the two is mislabeled, and nothing in the paper lets us say which. We flag it rather than pick a side, because we cannot check it. A compound event, in this work, is a heat stress day followed straight by a tropical night, with no cooling in between, and Europe has gained more of them than anywhere on Earth. Single-day pairings are 73 percent more common than in the 1970s. Runs of 15 to 30 days are 3. 37 times more common. Even the longest, stretching to 120 consecutive days, have almost doubled in frequency. Anyone can walk these fields inThermal Trace, ECMWF’s interactive map on the same dataset. Is it the heat, or is it just that there are more of us?The honest answer is both. What matters is the proportion, and this is the part of the paper we found most worth reading twice. Emerton and colleagues ran a counterfactual. Hold the heat stress fields at their 1970s values, move only the population forward to 2020, and whatever exposure comes out is what demography alone would have delivered. Whatever is left belongs to the climate. Take at least 30 days a year of very strong heat stress: population growth adds 5. 3 percentage points of the world’s people, and the change in heat stress adds 10. 2. Nearly twice as much. One headline number behaves differently. Exposure to at least one day of extreme heat stress rose from 15. 9 percent of the world’s people to 21. 9 percent. Population growth supplies 4. 5 points of that rise, and the heat supplies only 1. 4. We ran the arithmetic ourselves, because a round billion is the kind of figure that gets rounded into existence between a paper and a headline. In 1975 the world held 4. 06 billion people, and 15. 9 percent of them is 0. 65 billion. By 2020 the total was 7. 77 billion, and 21. 9 percent of that is 1. 70 billion. The difference is 1. 05 billion, so the extra billion is real and nobody is inflating it. Most of that particular billion, at that particular threshold, is us. Broaden the threshold and the balance tips back. In the 1970s, 55 percent of the world lived through at least 90 days of strong heat stress a year, and it is 70 percent now, with population growth contributing 7. 5 points and the heat 6. 5. Sub-Saharan Africa, South and Southeast Asia, the Arabian Peninsula and the Mediterranean carry the heaviest exposure of all. Our June piece on accelerating warming asked whether we are ready for it. This one asks what it has already done, and to whom. The record stops in 2024, and it never saw a cityFour limits sit on everything above, and the authors state three of them plainly. ERA5-HEAT cannot resolve urban heat islands or microclimates, so the exposure estimates come out conservative, and the paper says so specifically for cities. That is not a small caveat. Most of the exposed population lives in one, on asphalt that holds the afternoon well past midnight. UTCI also models a standardized reference person, with no age and no acclimatization, so an eighty-year-old and a fit thirty-year-old on the same street get the same number. Nighttime categories are the softest part, and the authors say so themselves. Those four bands were built for daytime conditions, so overnight minima get an indicative measure of severity rather than a physiologically exact one. They call for proper nocturnal thresholds. Nobody has them yet. One more limit hides in the methods. Every day here runs from 00:00 to 23:59 UTC with no adjustment for time zones, so for much of the planet a single counted day straddles two local nights. And the record ends in 2024. It cannot say what the summer of 2026 did to any of these curves, and the sea numbers above are no substitute, because they measure a different quantity in a place the grid throws away. The same ERA5 stream feeds this paper andthe daily global temperature curve in Climate Pulse, so 2026 will enter it on schedule. We are leaving two things out on purpose. Attribution, which asks how much likelier warming made any particular 2026 heat wave, is a different method with a different literature. So is the urban heat island work that would fix the blind spot above. Each deserves its own piece. What five hundredths of a degree per decade adds up toFive decades of it comes to a quarter of a degree, which is how much further the hottest nights have run than the hottest days since the 1970s. Africa spends 70 percent of its days and places in at least strong heat stress while Europe spends 9, a ratio missing from almost every European account of this summer. The extra billion is real at 1. 05 billion, and 4. 5 of its 6 percentage points are population growth rather than warming. Everything above stops in 2024 and never sees a city. We wrote it out this way for you because the version circulating this week hands you a multiplier with the significance mark stripped off, and the mark is what you actually need. Reason asleep is where bad numbers rule. Argue with the paragraph above if you think we misread the table; being corrected beats being agreed with early. ERA5-HEAT updates in near real time, so we will run this summer against these values the day 2026 lands in it. FreeAstroScience, Rimini. Gerd Dani. SourcesEmerton, R. , Nicolas, J. , Lombardi, A. & Di Napoli, C. (2026). Global heat stress intensification and its expanding footprint on the human population. Nature Climate Change, 16, 989-996. Published online 22 June 2026. https://doi. org/10. 1038/s41558-026-02670-5Di Napoli, C. et al. (2020). Thermal comfort indices derived from ERA5 reanalysis. Copernicus Climate Change Service (C3S) Climate Data Store. https://doi. org/10. 24381/cds. 553b7518Di Napoli, C. et al. (2020). ERA5-HEAT: a global gridded historical dataset of human thermal comfort indices from climate reanalysis. Geoscience Data Journal, 8, 2-10. Dataset:https://cds. climate. copernicus. eu/datasets/derived-utci-historicalDi Napoli, C. , Pappenberger, F. & Cloke, H. L. (2019). Verification of heat stress thresholds for a health-based heatwave definition. Journal of Applied Meteorology and Climatology, 58, 1177-1194. Roye, D. et al. (2025). Short-term association between hot nights and mortality: a multicountry analysis in 178 locations considering hourly ambient temperature. Environment International, 203, 109719. Chambers, J. (2020). Hybrid gridded demographic data for the world, 1950-2020. Zenodo. https://doi. org/10. 5281/zenodo. 3768003World Health Organization (2024). Heat and Health. https://www. who. int/news-room/fact-sheets/detail/climate-change-heat-and-healthFraioli, L. (2026). L’Europa soffoca, ma e solo l’inizio. “Rimpiangeremo questa estate”. la Repubblica, Cronaca, 6 August 2026. Talignani, G. (2026). Il Mediterraneo tocca 33 gradi, mai cosi calda l’acqua del mare: “L’aria fresca inneschera disastri”. la Repubblica, Cronaca, 7 August 2026. Copernicus Climate Change Service (2026). From evidence to understanding: game-changing visualisations powered by C3S data. C3S News, 27 July 2026. Climate Pulse:https://pulse. climate. copernicus. eu/