The summer France changed climates

On August 11, 1999, France stood outside with cardboard glasses to watch a total solar eclipse, and plenty of people had a sweater on, because the afternoon topped out somewhere between 17 and 25 °C. On August 12, 2026, twenty-seven years later almost to the day, another eclipse crossed the sky, partial this time as seen from France, and several parts of the country were sitting at 40 °C.

Serge Zaka put the two temperature maps side by side on Bluesky and said almost nothing, since there was nothing to add^[1]. I've followed him for a while, and that post is a decent summary of why. He sounds the alarm without ever giving up on explaining.

The storm chaser who lives under police protection

Zaka is not a television weatherman. He trained as an agricultural engineer in Dijon, earned a doctorate in environmental biology at Poitiers in 2016 with a thesis on how temperature affects forage plants, and works as an agroclimatologist, which is to say he watches what the climate grows and what it kills^[2]. He also chases storms on his own time, and one of his shots, a lone bolt of lightning over the Bay of Cannes, won the public prize at the 2021 Weather Photographer of the Year competition. In March 2025 he and Jérémie Szpirglas published Orages sur le climat, or storms over the climate.

Then there is what it costs him. On December 31, 2025, he announced that the French interior ministry had placed him under police protection after a run of death threats, calls for his murder, messages urging him to kill himself, and racist abuse^[3]. He blames part of that violence on interests tied to fossil fuel finance, and he has not slowed down. Counting degrees ought to be one of the quieter jobs available, and it now takes a certain nerve to do it out loud.

Five heat waves and seventy days without a break

"Since May, I honestly feel like we're living under some other country's climate." Zaka wrote that on August 7, as he announced the fifth heat wave of the year^[4]. The month-by-month record backs him up.

It had begun in late May, which had never happened before. From the 21st to the 30th, a heat dome parked over Western Europe across the Pentecost weekend and France hit levels it had never seen that early in the year, with up to ten straight days above 30 °C south of the Loire. Météo-France chose to call it an early heat episode rather than a heat wave, a careful way of saying it had arrived before the season officially opened^[5].

June, from the 17th to the 30th, settled the vocabulary question. On the 24th and 25th, the twenty-four-hour mean temperature across the whole country reached 30 °C for the first time since records began, in any month, where the previous peak had stopped at 29.4 °C on August 5, 2003. Paris and Limoges each ran ten hot nights back to back^[6]. I was writing about how to keep an apartment livable right in the middle of that one.

July ran from the 4th to the 19th, sixteen days of heat wave, with 37 départements, more than a third of the country, under red alert on the 12th. The month closed on an outright record, a national average of 24.9 °C, the hottest month ever measured in France since 1900, ahead of August 2003 and its 24.8 °C, with afternoons averaging 5.2 °C above normal^[7]. A fourth wave started on the 28th, and a fifth heat wave on August 8.

The number that captures all of it sits elsewhere. On August 10, Zaka pointed out that France was living through the longest run of consecutive days above seasonal normals ever observed, and that the previous record had been set in the spring of 2026, with sixty-one days between January 12 and March 13^[8]. On August 14, after 102 new monthly records fell in a single day, he counted 726 since the start of the year across the stations that have been running for more than fifty years and are considered reliable, then added a line worth more than any chart, "I'll let you guess how many cold records"^[9]. The answer is zero.

They told him to stop saying exceptional

On July 29 he posted the thread that explains best why I read him. People had been telling him to stop describing temperatures as exceptional when they clearly are not anymore, and to drop the alarmism, since these episodes come back every year and we're all still here. His answer fits in one sentence. An exception that becomes frequent does not stop being destructive^[10].

The reasoning underneath deserves a minute. Statistics are a human invention, seasonal normals get recalculated every ten years, and a curve will always absorb whatever you feed it. The oak trees have not adjusted their scale. Our forests hold the same species they held fifty or a hundred years ago, and those species have not evolved at the speed the climate is breaking, so what looks routine on our graphs stays violently abnormal for the ecosystems living under it.

Then comes the example that makes it land. The Bordeaux-Mérignac station has been measuring temperatures since 1920. Through 2025, it had recorded five days at or above 40 °C. By August 13, 2026, it had seven for that summer alone^[11]. No rolling average and no argument about word choice survives those two numbers set next to each other. The 40 °C mark has stopped being an event anyway, having gone in ten years from a handful of readings a year to 114 in 2026, a fiftyfold jump^[12].

What the summer took

On August 6, the 2026 drought became the most intense ever observed in France since record keeping began, past 1976 and past 2022^[13]. By the 14th, the country's soils held only 10% of their capacity, with no usable water left at all between La Roche-sur-Yon and Mâcon, and the Vendée, the worst-hit department, had dropped below 1% in its farm soils^[14].

Plants took the first hit. From the Vendée to Burgundy, leaf activity fell by as much as half under the combined weight of June's heat and drought, which means half the photosynthesis of an entire region shut down^[15]. Livestock followed. In late June, Zaka reported excess mortality of 1,000% in laying hens, 200% in pigs, and 45% in cattle, several million farm animals dead from heat, and authorities in Brittany forced to allow carcasses to be buried in the ground of the farms themselves^[16]. He filed the episode alongside 2003 as one of the two largest animal die-offs in recent French agriculture.

That leaves the forests, which burned everywhere. The Fontainebleau massif lost 2,050 hectares starting July 12, unheard of in the Paris region, and water bombers flew against a regional forest fire there for the first time^[17]. In the Gironde, the megafire that started at Saumos on July 22 destroyed 42,000 hectares and 240 buildings, and forced 224,000 people out of their homes at the height of the crisis^[18].

The worst part arrives once the heat leaves

On the evening of August 18, Zaka finally posted good news. After seventy straight days above normal, thousands of broken records, five heat waves, and three heat episodes, the end of the month was due back at seasonal temperatures. The good news lasted one sentence, because the rain returning to the Channel coast and the southeast will not rewet the soils for weeks, and across a wide central band of the country the drought carries on without a pause. In his words, we're not out of the woods^[19].

Two days earlier he had posted something much worse. The planet is now very likely headed for the most intense El Niño since anyone started watching, and probably for centuries^[20]. El Niño is a periodic warming of surface water in the equatorial Pacific that redistributes heat around the globe and mechanically pushes world temperature up for roughly a year. It does not cause climate change. It stacks on top of it, which is the problem.

American forecasts agree. NOAA's Climate Prediction Center puts the odds of very strong conditions between October and December at 81%, with an expected anomaly around +2.66 °C in the Pacific reference zone, where the strongest event ever measured since 1950, the winter of 1982-1983, topped out at +2.4 °C. The episode has a 97% chance of holding into early spring 2027^[21]. The first effects are already visible, since August 2026 may end up the hottest August ever measured on Earth and the Mediterranean just broke its own record with a surface average of 28.5 °C, across all months. Zaka's conclusion runs to one line I could have done without, which is that we are only at the beginning of a year of planetary records.

Back in June I wrote a whole article here about surviving a heat wave, the fan pulled back from the window, the chalk wash on the glass, the ice block behind the turbine. Two days later, in his thread about the dead livestock, Zaka dropped a sentence I haven't been able to shake. Our debates fixate on air conditioning, and that's what makes us miss the actual problem. He hadn't read my article. He was still right.

I don't regret writing it, because you have to get through the summer you're given. What I see more clearly now is the gap between adapting and watching. Zaka does the second job. He counts the records nobody counts anymore, refuses the word "normal" once it stops meaning anything, and does it under police protection, because reporting the temperature has turned into taking a side.

If you want someone to follow who helps you understand what's happening rather than just dread it, his account is there^[22]. We'll probably have to redraw the scale on those graphs again next year, and he'll be there to redraw it.