Pesquisadores usaram fatias de abacaxi para pôr a amizade dos animais à prova
Descoberta pode dar pistas sobre as origens das normas humanas de justiça e amizade
27.out.2026 às 6h00
Taylor Mitchell Brown
Pasa e Yoyo são melhores amigas. Elas passam as tardes descansando juntas na grama do Santuário de Chimpanzés da Ilha de Ngamba, em Uganda, cuidando uma da outra e compartilhando comida. Às vezes, até tentam fugir juntas do local.
Elas também são menos propensas a roubar comida uma da outra, segundo os resultados de um experimento que usou fatias de abacaxi para pôr a amizade delas à prova.
O estudo é o primeiro a constatar que primatas podem priorizar amizades em vez de alimentos, revelando uma capacidade de conter impulsos egoístas até então não documentada nesses animais.
Chimpanzé cuida da higiene de outro no Santuário de Chimpanzés da Ilha de Ngamba, em Uganda – Innocent Ampeire/NYT
A descoberta, que pode oferecer pistas sobre as origens das normas humanas em torno da amizade, foi publicada recentemente na revista Evolution and Human Behavior.
“Não era óbvio para nós que os chimpanzés roubariam menos de seus amigos”, disse o psicólogo social e autor do estudo Oded Ritov, da Universidade da Califórnia em Berkeley (Estados Unidos).
A pesquisa mostra que os chimpanzés, assim como os seres humanos, têm expectativas sociais uns em relação aos outros. “Acreditamos que essas expectativas possam ser um elemento fundamental que, mais tarde, evoluiu para o senso de justiça e outros tipos de moralidade que observamos nos seres humanos.”
As noções de “amizade” entre primatas não humanos têm um histórico controverso. Muitos pesquisadores afirmam que atribuir esse termo às relações entre primatas como chimpanzés e bonobos é uma forma de antropomorfismo, que projeta ideias familiares a nós sobre animais evolutivamente distantes, aos quais elas não se aplicam.
“Durante muito tempo, era como se essa fosse uma palavra proibida”, disse a antropóloga e coautora do estudo Laura Lewis, da Universidade da Califórnia em Santa Bárbara (EUA). Nas últimas décadas, porém, os primatólogos vêm encontrando cada vez mais evidências de amizades entre primatas sociais, constatando que eles demonstram mais confiança e apoio em relação aos companheiros dos quais se sentem próximos.
Na natureza, as amizades entre chimpanzés se desenvolvem sobretudo entre os machos, que cultivam laços sobretudo para se defender de machos alfa tirânicos ou para contar com apoio em confrontos contra outros grupos. Entre os bonobos, as amizades são comuns entre as fêmeas, que precisam disputar poder na matriarquia predominante. Em cativeiro, em ambas as espécies, as amizades se desenvolvem de forma mais ampla dentro do grupo, com menos diferenças entre os sexos.
“Em cativeiro, de certa forma, eles são obrigados a conviver uns com os outros o tempo todo”, afirmou a ecóloga comportamental Zoë Goldsborough, especialista em primatas do Instituto Max Planck de Comportamento Animal (Alemanha), que não participou da pesquisa. Como o cativeiro elimina algumas das limitações normais da vida de um chimpanzé relacionadas à alimentação e aos conflitos, observa-se o florescimento de um número muito maior de relações, incluindo amizades.
Uma vez estabelecidas, as amizades proporcionam muitos privilégios. Amigos passam tempo juntos, dão as mãos, fazem ninhos juntos à noite, protegem-se em confrontos e consolam-se após brigas ou quando um filhote morre. Às vezes, um chimpanzé chega a sentir ciúme ao ver um amigo cuidando da pelagem de outro e interfere para impedir que eles criem laços, segundo Goldsborough.
No novo estudo, os pesquisadores queriam descobrir se as amizades entre chimpanzés estimulavam o autocontrole diante de certos comportamentos inadequados —neste caso, o roubo.
A equipe examinou minuciosamente 800 horas de gravações em vídeo do santuário em Uganda para identificar quais chimpanzés eram amigos. Em seguida, realizou testes nos quais os animais eram colocados em duplas. Um dos chimpanzés tinha a opção de agir como ladrão e roubar frutas de uma vítima, que podia ser um amigo ou um conhecido com quem não tinha laços especiais. Os chimpanzés ficavam em salas contíguas: em uma delas, havia bandejas de metal com fatias de abacaxi; na outra, cordas presas às bandejas, que o ladrão podia puxar para ficar com as guloseimas.
“Às vezes, os ladrões entravam e puxavam a bandeja com tudo”, afirmou Lewis. Em outras ocasiões, porém, eles entravam, examinavam a corda e depois a soltavam, em um gesto de protesto. Quando Yoyo teve a oportunidade de roubar de Pasa, rejeitou completamente o experimento: virou as costas para as cordas e se recusou até mesmo a entrar na sala.
De modo geral, os resultados revelaram que os amigos eram menos propensos a roubar uns dos outros e afastavam as bandejas com abacaxi em apenas 61% das vezes, em comparação com 70% quando se tratava de conhecidos.
Quando um chimpanzé de fato roubava de um amigo, o resultado era o caos. A traição levava o chimpanzé lesado a fazer coisas como esmurrar o chão, soltar guinchos altos ou se coçar em um ataque de ansiedade, mostrando que esperava mais do companheiro.
Mas, após essas traições, os pesquisadores constataram que os mesmos chimpanzés perdoavam os amigos mais rapidamente e, em testes posteriores, acabavam aceitando esse tipo de trapaça. Os que não eram amigos não recebiam a mesma clemência e eram tratados com o mesmo grau de raiva durante todo o experimento.
“O fato de eles conseguirem moderar esse comportamento de acordo com suas relações sociais é extremamente importante e acreditamos que isso seja fundamental para manter essas relações sociais ao longo do tempo”, afirmou Lewis.
Como seres humanos, “damos muita importância não apenas às coisas boas que as pessoas fazem por nós, mas também ao fato de não fazerem coisas ruins conosco”, disse Goldsborough. “Esse estudo mostra muito bem que os chimpanzés também agem assim”.
Pesquisadores de universidade nos EUA identificaram padrão em relatos de crianças em meio a 1.678 casos analisados
Resultados não comprovam a reencarnação, mas consideram que o fenômeno merece investigação, dizem os autores
26.out.2026 às 23h00
Felipe Espinosa Wang
Um novo estudo baseado no maior acervo do mundo de supostas memórias de vidas passadas identificou um padrão em relatos de crianças em meio a 1.678 casos analisados. Os pesquisadores afirmam que os resultados não comprovam a reencarnação, mas consideram que o fenômeno merece investigação, já que não pôde ser totalmente explicado por fatores culturais ou psicológicos.
É possível que a reencarnação tenha pouco espaço na ciência atual, já que ainda não se encontrou evidências de que ela realmente exista. Mesmo assim, a ideia de nascer de novo está longe de ser estranha para muitas formas de tentar compreender o mundo.
Para certas tradições do hinduísmo e do budismo —ainda que encarem esse processo de maneiras diferentes—, o ciclo de mortes e renascimentos não é uma mera curiosidade sobre o que acontece após a morte, mas sim um elemento central na concepção da existência.
Pesquisadores sobre reencarnação investigam as causas da mortes e o ‘tempo de intervalo’ entre uma vida e outra – Andrey Popov/Adobe Stock
Arquivo sobre reencarnação
A falta de evidências não impediu que alguns pesquisadores se aventurassem em um terreno pouco comum para a ciência. Entre eles estava o psiquiatra Ian Stevenson (1918-2007), que já na década de 1960 viajava por diversos países para investigar testemunhos de supostas vidas passadas.
Seu trabalho está ligado à Universidade da Virgínia, nos Estados Unidos, onde há mais de seis décadas são compilados relatos tão inusitados quanto controversos: crianças que, segundo os depoimentos, contam espontaneamente detalhes que parecem corresponder à vida de pessoas já falecidas.
A Divisão de Estudos Perceptivos, que vem compilando esses relatos desde a década de 1960, formou assim o maior acervo do mundo dos chamados “casos do tipo reencarnação”. E uma recente análise, publicada na revista Psychology of Consciousness: Theory, Research, and Practice, identificou nesse acervo um padrão que fatores culturais e psicológicos não conseguiram explicar completamente.
Para se ter uma ideia do tipo de depoimentos que esse arquivo reúne, é preciso conhecer dois casos descritos pela revista Popular Mechanics.
James, um menino do estado da Louisiana, tinha pesadelos com um avião em chamas. Aos três anos, ele afirmou ter sido piloto durante a Segunda Guerra Mundial e mencionou um porta-aviões chamado Natoma Bay, além de um colega chamado Jack Larsen. Os pesquisadores relacionaram suas declarações com a biografia de James Huston Jr., um piloto cujo avião foi derrubado no Oceano Pacífico em 1945.
Outro caso é o de Ryan, um menino de quatro anos do estado de Oklahoma. Ao ver uma fotografia de Hollywood dos anos 1930, ele afirmou se reconhecer em uma das pessoas, que nem sequer aparecia nos créditos. O relato acabou sendo associado a Marty Martyn, um agente de Hollywood já falecido. Dos detalhes que o menino forneceu sobre sua suposta vida anterior, os pesquisadores conseguiram confirmar 55 com a biografia de Martyn.
O que busca o novo estudo
O mais recente levantamento não se concentrou em determinar se esses relatos são memórias autênticas de outras vidas, mas sim investigar padrões entre os depoimentos. Para isso, uma equipe da Universidade da Virgínia, liderada pela pesquisadora em neurociência cognitiva Saeedeh Sadeghi, examinou 1.678 casos que tinham algo em comum: em todos eles, os pesquisadores haviam identificado possíveis correspondências entre o que uma criança contava e a biografia de uma pessoa já falecida.
Uma das questões levantadas foi quanto tempo separava os dois eventos: o falecimento da pessoa identificada e o nascimento da criança —o que, nessa área, é conhecido como “tempo de intervalo”. Segundo a Popular Mechanics, havia informações suficientes para calculá-lo em 1.319 casos: a média foi de 16 meses, embora o intervalo mais longo registrado tenha chegado a 150 anos.
Mortes inesperadas e intervalos mais curtos
O que realmente chamou a atenção foi a comparação das circunstâncias das mortes. Quanto mais inesperado fosse o falecimento, menor tenderia a ser esse intervalo. Nos casos de mortes completamente imprevistas, a média era de 12 meses. Por outro lado, quando as mortes eram mais previsíveis, esse intervalo chegava a 21 meses.
Ou seja, o intervalo entre a morte e o suposto renascimento era aproximadamente 43% mais curto quando o falecimento havia sido inesperado, de acordo com os dados coletados pela publicação Science Alert.
O arquivo também apresenta outra particularidade: entre os casos analisados, 57% das mortes foram classificadas como não naturais e 68% como completamente inesperadas. A idade média das pessoas falecidas era de apenas 30 anos, números que traçam um perfil pouco comum. Segundo a Popular Mechanics, a proporção de mortes não naturais e inesperadas supera amplamente a registrada nas estatísticas gerais de mortalidade.
Precedente de quase quatro décadas
A possível relação entre uma morte violenta e um suposto renascimento mais rápido já havia sido levantada décadas atrás. Em 1988, o psicólogo N. K. Chadha e o psiquiatra Ian Stevenson analisaram 326 casos e observaram uma tendência semelhante. No entanto, como lembra a publicação IFL Science, a associação só atingiu significância estatística na Índia e no conjunto dos casos estudados.
Justamente por causa dessas limitações, a nova equipe quis verificar se o padrão persistia em uma amostra muito maior e, sobretudo, se poderia ser explicado por fatores culturais ou psicológicos.
Uma das objeções mais evidentes é que as famílias que acreditam na reencarnação poderiam favorecer esse tipo de relato. Para verificar se esse poderia ser o caso, os pesquisadores examinaram possíveis fontes de viés, como o contato entre as famílias, a confiabilidade dos dados, as crenças religiosas e se as crianças poderiam ter tomado conhecimento prévio das informações que relatavam.
Mesmo após considerar esses fatores, a associação permaneceu: “Para nossa surpresa, a relação continuava existindo quando a analisamos em um número muito maior de casos, e os fatores sociais e psicológicos que examinamos não conseguiam explicá-la”, observou Sadeghi, de acordo com um comunicado da Universidade da Virgínia (UVA Health).
Na verdade, o efeito era mais acentuado entre famílias cujas religiões não aceitavam a reencarnação: “Esses resultados não corroboram a ideia de que as crenças sociais estejam determinando esse padrão”, afirmou o psicólogo social da UVA, Philip Cozzolino.
As limitações da descoberta
Apesar do que foi observado no estudo, esses resultados estão longe de comprovar a reencarnação. A relação identificada é fraca: saber se uma morte foi inesperada permite explicar estatisticamente apenas entre 2% e 3% das diferenças na duração dos intervalos entre uns casos e outros. Além disso, por se basear em dados observacionais, o estudo não permite estabelecer uma relação de causa e efeito nem concluir que a consciência sobreviva à morte.
Outras explicações também não podem ser totalmente descartadas. Como aponta a IFL Science, as informações podem circular entre famílias e comunidades sem que os pesquisadores consigam reconstruir como elas chegaram até a criança. Soma-se a isso a possibilidade de identificar retrospectivamente uma pessoa falecida cujos dados se encaixem em parte do relato. Até mesmo o maior impacto das mortes violentas poderia influenciar os depoimentos coletados.
A capacidade de prever os intervalos é limitada. Mesmo assim, chama a atenção dos pesquisadores o fato de existir uma associação estatística entre as circunstâncias de uma morte e o nascimento de uma criança que supostamente se lembra dessa vida. Segundo os pesquisadores, ainda não foi estabelecido nenhum mecanismo convencional que vincule diretamente esses dois eventos.
A ideia também apresenta um curioso paralelo com algumas crenças religiosas. Os pesquisadores apontam que, em certas tradições hindus, budistas e indígenas da América do Norte, uma morte repentina pode estar associada a desejos, laços ou obrigações que ficaram para ser resolvidos. Segundo sua interpretação, essas questões pendentes poderiam favorecer um renascimento mais rápido. No entanto, não se trata de uma crença compartilhada por todas essas tradições nem de uma explicação científica para o fenômeno.
Questão que continua em aberto
Por enquanto, o mistério permanece. Os resultados não comprovam a existência da reencarnação, algo que a própria Sadeghi faz questão de ressaltar: “Isso é uma prova de algo metafísico? Os dados são observacionais, e tomamos muito cuidado para manter uma postura agnóstica em relação ao mecanismo”, afirma a pesquisadora.
No entanto, ela considera que a descoberta merece atenção: “O que posso dizer é que o padrão é real nos dados”, ressalta, e isso constitui um motivo “para continuar pesquisando, não para ignorar o assunto”.
Uma postura que é compartilhada pela pesquisadora Marieta Pehlivanova, que, segundo a UVA Health, considera que a falta de uma explicação clara oferece novos motivos para continuar estudando esses padrões.
No fim das contas, o trabalho levanta mais perguntas do que respostas. A correlação aparece nos dados analisados, mas ainda não está claro quais fatores a explicam. E enquanto não for encontrada uma explicação convincente, a reencarnação continuará despertando fascínio e curiosidade, mesmo que sua existência continue sem contar com provas científicas sólidas. Ao menos até agora.
Earth makes a tiny seismic rumble every 26 seconds.
Is the pulsating caused by ocean waves, volcanoes, or something else completely? Theories abound.
The “microseism” doesn’t seem to be hurting anything and has not been a high priority for seismologists.
Why is Earth pulsating every 26 seconds, and why can’t scientists explain it after 60 years?
This is an enigma wrapped in a periodically predictable mystery motion. It could be a harmonic phenomenon, a regular seismic chirp caused by the sun’s energy, or a beacon drawing scientists to its source to begin a treasure hunt.
In the early 1960s, a geologist named Jack Oliver first documented the pulse, also known as a “microseism,” according to Discover. Oliver, who worked at Columbia University’s Lamont-Doherty Geological Observatory at the time, heard the noise, but didn’t have the advanced instruments seismologists have now at his disposal.
Since then, scientists have spent a lot of time listening to the pulse and even finding out where it comes from: “a part of the Gulf of Guinea called the Bight of Bonny,” Discover says.
Some researchers think the pulse has a kind of prosaic cause. Under the world’s oceans, the continental shelf acts as a gigantic wave break—it’s the boundary off the very far edge of, for example, the North American continental mass where the highest part of the plate finally falls off into the deep abyssal plain. Scientists have theorized that as waves hit this specific place on the continental shelf in the Gulf of Guinea, this regular pulse is produced.
If that sounds improbable, consider all the different shapes of drums, from timpani to bass drums to bongos that you hit with your hands. It’s not impossible that just one shape of continental shelf “drum” would create the right harmonic bang to rattle the Earth. If that’s true, we’re probably lucky it’s just one.
But other researchers think the cause is a volcano that’s also very near the critical spot: “That’s because the pulse’s origin point is suspiciously close to a volcano on the island of São Tomé in the Bight of Bonny,” Discover explains. And there’s a similar volcanic microseism that’s already well documented in Japan.
It seems like reams of new scientific research emerge every day, but the mystery pulse is a good reminder that so much remains to be discovered. Scientists have studied the pulse and debate its origin, but it just hasn’t reached a tipping point of interest to be solved. Discover explains that researchers have likely been studying higher-priority seismic events instead, which makes sense.
Don’t worry, you have to wait an average of 16 months for another turn. Image credit: Lidiia/Shutterstock.com
A new study looking at reports of reincarnation claims to have found an odd correlation between the manner of death of a person and the claimed “rebirth” of an individual who remembers this “past life” in some way. Since this is October, the spookiest of all the months, we thought we’d take a look.
The first study
In 1988, psychologist N.K. Chadha and psychiatrist Ian Stevenson conducted a small study on reports of reincarnation, and focusing on the cause of death of the “previous person”, (PP) or the person whom it is claimed a second person is reincarnated from.
Looking at a pool of “326 cases suggestive of reincarnation”, the team claimed to have found that where the supposed “previous person’s” death had been violent, the time until a claimed “reincarnation” was shorter than if the claimed “previous person’s” death had been natural.
“The effect was found in all cultures studied (except Sri Lanka), but reached statistical significance in only one individual culture (India) and for the group taken as a whole,” the team wrote in their study.
“The age of the subject when first speaking about the previous life was also found to be significantly lower with violent death (compared with natural death).”
Let’s be frank; we’re not talking hard science here, and the researchers drew much broader conclusions than their data could necessarily support.
“We ourselves cannot think of any plausible normal hypothesis to account for our data on the intermission period, and so we provisionally interpret it to mean that, if reincarnation occurs, persons dying violently tend to be reborn more quickly than those dying naturally,” that team wrote in their paper, for example, whilst clarifying that this is if reincarnation occurs.
“This accords with traditions of Hinduism, according to which, every action in our lives, both of body and mind, is stored in the form of a Samskara or impression in our minds. Desires persist after death as strongly as before and ultimately lead to a new birth.”
Of course you might question why this effect would only reach statistical significance in one country, and wonder why reincarnation appears to be respecting Sri Lanka’s borders and not delivering those who died by violence in the numbers seen elsewhere.
But in short, it was quite a small study, and with so many variables in how these “cases” are collected and reported, it doesn’t tell you much. As the team explains, it is pretty difficult to eliminate information transfer between families and communities, so all that these are documents of is claims, not confirmed cases.
New investigations
Now a new team from the University of Virginia’s Division of Perceptual Studies have attempted to look into this claimed effect, to see if it holds up in a larger dataset.
The team looked at a database of reports put together by the Division of Perceptual Studies, which encompasses 2,254 cases, 74 percent of which the investigators felt that the details of a “past life” provided by the supposedly reincarnated person matched the details of another previously deceased person.
This includes plenty of details of variables in the cases, such as how the supposed PP died, information on whether the supposed reincarnated child had been fed details of the PP by the family or during investigation, as well as how many details the “reincarnated” supplied which the investigators deemed “matched” with a PP.
Small but significant effect
“There had been an earlier published report suggesting a relationship between violent death and a shorter time between death and birth,” lead author Saeedeh Sadeghi, PhD, said in a statement.
“We wanted to test that relationship more systematically and see whether it could be explained by cultural or psychological influences. To our surprise, the relationship was still there when we tested it in a much larger number of cases, and the social and psychological factors we examined did not explain it away.”
The effect they found was small, but statistically significant.
The team found that the median time between the death and claimed rebirth was 16 months, but with unexpected deaths the turnaround time to get back at human shenanigans could be cut down by 40 percent.
However, the team stresses that the association was small, accounting for approximately 2–3 percent of the variance in “intermission time” between death and supposed reincarnation.
“This is the largest systematic study to date of temporal patterns in reported past-life memories,” Sadeghi added.
“Individual cases can leave room for different interpretations, but patterns across many cases are much harder to explain as exaggeration, fantasy or fabrication.”
Interestingly (though please don’t take this as an endorsement of reincarnation) the team found that this effect still existed when the child involved did not make any claims about remembering their deaths.
So, that’s reincarnation confirmed?
No. It’s worth really stressing that these are not clean cut cases of people clearly being reincarnations of someone else.
The possibility of information transfer within and between communities is extremely difficult to rule out, and provides a much more plausible explanation than reincarnation in many cases, whilst you can’t really rule out biases either, as much as you might try.
Plus there is the problem that a person claiming to have been reincarnated could provide a few details, and with plenty of deaths out there to choose from you can post-select the person who sounds most similar.
News of violent deaths travels wider than natural deaths, and this could certainly be a factor that complicates the picture. If hints of this person then gets back to the claimant, for example by their family, well then you’ve doubled your investigative problems.
What can we conclude from this? As with the first study, you’re placing a lot of trust in what investigators have deemed a plausible match between supposed “parent persons” and the supposedly reincarnated. These cases are pretty murky, with plenty of possibility for memory contamination and other plausible explanations that don’t require reincarnation.
We know that sounds annoying to people who want to believe in reincarnation (and you’re absolutely free to) but as the researchers make clear, the scientific evidence here is not strong enough to make any claims of reincarnation. Nevertheless, claims of these things are always interesting, even if you suspect what they are really studying is cultural beliefs and the transfer of information in communities.
“We hope this work encourages researchers to look beyond individual cases and ask whether there are broader patterns that can be tested systematically,” Marieta Pehlivanova, PhD, added.
“The mechanisms underlying these findings are currently unclear, but that gives us even more reason to keep looking, not to look away.”
It seems like every day a new, potentially internet-shattering cybersecurity issue is discovered or exploited by artificial intelligence. It’s gotten bad enough that even OpenAI CEO Sam Altman, one of the people who is most responsible for pushing AI development forward despite the risks, now believes that simply patching the problems is no longer a viable option.
During an appearance at the Stanford Graduate School of Business, Altman told a group of grad students that the cybersecurity crisis is worse than anyone could imagine. “I no longer think we’re going to be able to just patch all the bugs,” he told the audience, in part due to the ongoing advancements of AI models and in part because the targets of cyber attacks have acted too slowly to get ahead of it.
“This cybersecurity…I don’t even know a dramatic enough word to use, but the problem coming towards us, I think, is still something that we have not, as organizational leaders or countries or individuals, taken seriously,” he said.
Altman pointed to the freakout when Anthropic announced its Claude Mythos model with the ominous warning that it can’t be made public because it is simply too powerful and could lead to the undoing of cybersecurity as we know it. “People definitely patched some bugs, but it was another one of these cases where people didn’t really intuit the exponential.”
Basically, Altman suggested that the organizations that did take action just patched up the problems that AI models were able to find and called it a day, as if no other future bugs or exploits would be found. “Largely speaking, I think most organizational leaders have somewhat moved on. And the models have kept going,” he said.
Altman seems to be scolding these organizations for not taking the situation seriously enough or considering that models might evolve and improve, but seems to have less interest in reflecting on whether or not he might have undersold the issue while pushing forward because fully acknowledging it might have led to a more serious push to slow or stop development while he’s got an initial public offering he’s trying to get to.
Notably, when OpenAI’s own agents broke containment and hacked Hugging Face, it was clear that its own procedures and security methods were insufficient. Despite pushing for the AI industry to adopt chain-of-thought monitoring—a technique for examining a model’s internal reasoning—as standard practice, OpenAI had not implemented it for its own training sessions. And, according to an investigation by AI safety nonprofit METR, OpenAI had already been alerted that its agents had escaped their sandbox and patched the exploit that allowed them to do so. But those measures failed to prevent another escape that led to the attack.
So we’re kind of in pot-calling-the-kettle-black territory here for Altman. But worry not, OpenAI’s founder has a solution for where patching fails: more AI. “I think we’re going to have to completely switch the defense model to using the same technology for a new kind of defense shield,” he said. OpenAI is apparently getting into the arms dealer model of selling to both sides of a conflict.
To think, we were so close to not having to live this nightmare. Altman revealed during the talk that he nearly went into investment banking because “it was what all the cool kids wanted to do, and I fell victim to peer pressure.” It’s not clear if he’s still as susceptible to peer pressure as he once was as a teenager, but we could all try bullying him just to see if it still works.
IPCC vice-chair Dr Ladislaus Chang’a and chair Sir Prof Jim Skea in discussion at the IPCC-65 meeting. Credit: IISD/ENB | Danny Skilton
Countries have failed – for the sixth time running – to agree on the timeline for publishing the Intergovernmental Panel on Climate Change’s (IPCC) next set of influential reports.
The IPCC’s 65th session (IPCC-65) was seen by many to be the final opportunity for countries to agree on a schedule where its seventh assessment report (AR7) would be published in time for the UN’s second “global stocktake” of climate action.
Before last week’s meeting in Addis Ababa, Ethiopia, IPCC chair Prof Sir Jim Skea had warned that releasing the reports in time for the stocktake in 2028 would depend on a timeline being agreed by the end of 2026.
But the week-long meeting descended into fraught disagreements between governments, resulting in Skea “suspending” the talks on Friday evening with no decision made.
Countries were also unable to agree on the IPCC’s budget for the rest of the AR7 cycle, leaving the body with a “serious funding crisis”.
The IPCC has pledged to reconvene the meeting at the “earliest opportunity”; however, it has not yet given an indication of when this will be.
In his concluding remarks, Skea described the Addis Ababa summit as a “profound disappointment” and “among the most difficult IPCC meetings”, according to reporting from the Earth Negotiations Bulletin (ENB).
Some countries have reacted with dismay to the breakdown of negotiations, warning that a failure to have the IPCC’s latest assessments on climate adaptation and mitigation ready for the UN stocktaking exercise would undermine global climate action.
A climate envoy from a Pacific island state said it was a “travesty that a handful of countries continue to delay the IPCC’s reports – the science we need back home to help us save lives and livelihoods”.
Below, Carbon Brief explores the key developments, debates and disagreements at the IPCC meeting.
How might the IPCC use AI and enhance gender inclusivity?
What timeline options were delegates debating?
Top of the agenda during the Addis Ababa meeting was a decision on the timeline of the publication of AR7’s three working group reports – also known as the “work plan”. The three reports cover climate science (WG1), impacts and adaptation (WG2) and mitigation (WG3).
Countries failed to reach a resolution on the scheduling of the influential reports at five consecutive IPCC meetings since 2024 – despite drafting being well underway.
Reporting from the ENB from inside these meetings has highlighted that countries are split, broadly, into two groups – those that want all three reports to be published in time for the UN’s second “global stocktake” in 2028, and those that one, or more, to be published afterwards.
A key pillar of the Paris Agreement, the global stocktake is a formal assessment of the world’s collective progress on tackling climate change. The exercise is designed to guide future global climate action, including decisions on national emissions targets and climate finance.
The second global stocktake is due to culminate in late 2028 at the COP33 climate summit.
The Addis Ababa meeting was the final IPCC plenary where countries would be able to agree on a stocktake-aligned timeline. In April, Skea told Carbon Brief that publishing reports by the end of 2028 would only be possible if a decision was taken before the end of 2026.
India, Kenya, Russia and Saudi Arabia have emerged as among the loudest voices calling for a timeline where the IPCC’s latest assessment of climate impacts and mitigation are published after the stocktake. They have argued that developing countries need more time to prepare, review and approve working group reports.
A coalition of developed nations and small-island developing states, meanwhile, have argued in favour of all three reports being published in time for the stocktake, so that the latest climate science can inform future climate action.
(For more on the standoff around the reports and why it matters for climate action, read Carbon Brief’s explainer.)
Ahead of the Addis Ababa meeting, the IPCC’s secretariat, on behalf of Skea, published a document setting out three options for a timeline for the working group reports. These are summarised below.
Option 1
Option 2
Option 3
Proposal set out by working group co-chairs at IPCC-61 in Sofia.
Revisited proposal presented by working-group co-chairs at IPCC-63 in Lima.
Submissions by India and Saudi Arabia to a survey sent out by IPCC in 2026.
Report approval
Plenary
Proposed month of approval
Plenary
Proposed month of approval
Plenary
Proposed month of approval
Working Group 1
IPCC-69
May 2028
IPCC-69
May 2028
IPCC-69
May 2028
Working Group 2
IPCC-70
June 2028
IPCC-70
July 2028
IPCC-71
April 2029
Working Group 3
IPCC-71
July 2028
IPCC-71
Sept 2028
IPCC-72
July 2029
Source: Table 1 and Appendix A in “Work plan of the IPCC seventh assessment report”, IPCC.
Over several days of fraught talks in Addis Ababa, the number of timeline options in contention ballooned to seven – before being reduced down to four.
How did the timeline discussions unfold?
On Monday, Skea established a “contact group”, chaired by Brazil and Ireland, to focus on the timeline issue.
By Friday morning, the group reported back to the plenary that they were still a “long way from consensus”, according to the ENB.
By that afternoon, four options still remained – the three set out by the IPCC ahead of the meeting and a fourth, proposed by Haiti, which would see WG1 approval in April 2028, WG2 in July 2028 and WG3 in October 2028.
The IPCC’s secretariat subsequently presented a “draft decision” for the timeline issue to plenary, informed by the contact group’s work. The text did not establish a formal decision on the timeline, but presented a number of options for how the IPCC could move forward on the issue, as well as proposals for addressing inclusivity concerns.
Two out of three options suggested that IPCC co-chairs continue work on reports in 2027 “without prejudging the panel’s agreement on the timing of the respective approval sessions” and for timeline discussions to be continued at a future session.
Another suggestion was that the IPCC fund a second delegate for developing countries and “economies in transition” for future approval sessions.
After one delegation said they were uncertain how to engage with the draft text, Skea clarified the text had three elements – a preamble, timeline and a section on inclusivity and avenues for providing input.
He said that “it was clear the panel would not be able to take a decision on the timeline at IPCC-65” and explained the idea was to include the four timeline options that emerged from the contact group in an annex “in order to capture progress made during IPCC-65”.
After countries were not able to agree on the draft decision, Skea suspended the meeting at 8:29pm on Friday.
He noted that it would no longer be inclusive for the meeting to continue, given that a number of delegations from developing countries had left.
Earlier in the plenary, Antigua and Barbuda’s delegation had emphasised that 6pm was its deadline and noted that representatives of other small-island developing states had departed. Nepal’s negotiating team had also stated it was leaving.
What were the main points of disagreement?
The ENB’s report details how the five-day meet was beset by disagreements between country delegations on the timeline issue.
As proceedings kicked off on Monday, a significant number of countries set out their position on whether publication of the reports should occur before, or after, the global stocktake.
Both camps cited inclusivity concerns. Those in favour of a stocktake-aligned timeline said that “inclusivity is a priority and should be addressed, but said delaying delivery of the report would further disadvantage the countries most vulnerable to climate change”, according to the ENB.
In discussions on Monday, the Pacific island nation of Vanuatu underscored that “every year of delay is a year in which communities on the frontlines bear losses that cannot be measured or claimed”, the ENB says. The delegation pointed to the International Court of Justice’s 2025 advisory opinion that countries have binding obligations to protect the climate.
The Netherlands delegation said it was “inconceivable” that two years before the conclusion of second global stocktake, the “body created to produce scientific advice accepts it will not deliver working group reports on mitigation and adaption, while the [IPCC co-chairs] say it is possible”, according to the ENB.
But other countries emphasised that the global stocktake could draw from other sources of science if the reports were not ready in time, the ENB notes.
China’s delegation opposed the IPCC holding more than three plenaries each year and said IPCC reports did not need to be linked to the global stocktake. Kenya said that publishing reports after the stocktake “did not mean science would not inform policy”.
Also among those to argue for a longer timeline were Iran and Iraq, with the latter underscoring the importance of participating in IPCC processes on an “equal footing” with other governments, the ENB says.
There were also disagreements over the specific mandate of the contact group set up by Skea on Monday to advance the timeline question. This debate endured well into the end of the week and days after the contact group had started its work.
Countries pushed back on each others’ characterisation of the different timeline options. Proponents for a longer timeline described plans where reports were published in 2028 as “rushed”, whereas those who wanted all reports published in time for the stocktake referred to the longer timeline as “delayed” or “significantly extended”, according to the ENB.
Countries also failed to reach consensus on the draft timeline decision presented to plenary on the last day of the meeting.
While some countries – including Antigua and Barbuda, Costa Rica, France, Germany, Hungary, Italy, Kiribati, Marshall Islands, Mexico, the UK and Uruguay – detailed their preference for particular passages of the IPCC secretariat’s text, others rejected it outright.
For example, Algeria, India, Iraq and Saudi Arabia stated the draft text could not serve as a basis for work or decisions, according to ENB. India, meanwhile, argued that it could not engage in the decision as the contact group reached “no clear conclusion”.
In an analysis of developments at IPCC-65, the ENB states that discussions in Ethiopia had revealed “room to compromise” between countries.
It continues that “most countries” in attendance had stressed the importance of not deviating from a tradition where the IPCC follows the advice of working group co-chairs when it comes to the schedule for production of assessment reports.
In this case, that would mean agreeing to publish all three working group reports in 2028 “on a timeline that the co-chairs consider feasible and meets all the criteria for IPCC reports” (options one and two in table above).
In a statement shared with Carbon Brief after the meeting, Marshall Islands climate envoy Tina Stege said that it was a “travesty that a handful of countries continue to delay the IPCC’s reports – the science we need back home to help us save lives and livelihoods”. She continued:
“We were here to work, but found every proposal dismissed – on inclusivity, on the relevance of the science for our national planning, on the necessity of evidence to inform the next global stocktake. Progress was blocked, down to the end – a shameful play for time when there is no time to waste.”
Hamid Abakar Souleymane, a negotiator from Chad, tells Carbon Brief the situation was “not helping” countries such as his, which are on the “frontlines” of climate change.
The African nation faces a host of climate impacts, including drought, heatwaves, severe rainfall, desertification and biodiversity loss, he says. As a result, it wants AR7 reports to be ready by the global stocktake so that “[climate] mitigation action can be accelerated”. He continues:
“It is only about two or three countries that are opposing every single element of the agenda – and this is really disappointing. We are in a very uncomfortable situation now because the global stocktake is coming. We don’t have a timeline [for reports] – and we haven’t yet adopted a budget.”
Abakar Souleyame also notes that Chad was at a disadvantage during negotiations because it was only able to send one representative to the meeting:
“A lot of meetings take place at the same time…The participation of a single delegate puts us in a very uncomfortable situation where we cannot contribute adequately to the ongoing process.”
Why is the IPCC in a ‘serious funding crisis’?
Inextricably linked with the AR7 timeline is the IPCC’s budget, which – among other things – allocates money for its author meetings and approval sessions, including support for experts from developing countries.
Throughout the meeting in Addis Ababa, governments and the IPCC leadership repeatedly emphasised that the IPCC is “facing a serious funding crisis”, the ENB says.
At the start of the meeting, the IPCC’s secretariat presented a proposed budget for 2026-29, which would cover the core work of AR7. This was submitted “against the backdrop of growing annual deficits and a progressively declining cash balance, raising significant concerns about the financial sustainability of the IPCC trust fund”, the document says.
The trust fund is made up of ongoing contributions from its parent organisations, the World Meteorological Organization (WMO) and UN Environment Programme (UNEP), in addition to voluntary contributions from its member governments and the UNFCCC.
The budget proposal notes that an “insufficient level of voluntary contributions” now “risks the successful completion of the AR7 cycle”.
Contributions in 2025 were 5.3m Swiss francs (£4.8m) lower than 2023, while costs have risen, the secretariat notes. As of the start of July, just 16 countries and organisations had contributed in 2026, totalling 2.8m Swiss francs (£2.6m).
These donors include the EU, UK, WMO, UNFCCC and UNEP, as well as Australia, Canada, Hungary, Luxembourg, New Zealand, Norway, Peru, South Africa, Sweden, Switzerland and Trinidad and Tobago.
Despite being its largest donor historically, the US has not contributed financially to the IPCC since 2024 – before the second term of US president Donald Trump.
Chart showing the largest direct contributors to the IPCC since its inception in 1988, with the US (red bars), EU (dark blue) and UNFCCC/WMO/UNEP (mid blue) highlighted. Grey bars show all other contributors combined. Figures for 2026 are January to June inclusive. Figures for 1988-2003 are reported per two years, so these totals have been divided equally between each year. Source: IPCC (2025) and (2010). Contributions have been adjusted, as per IPCC footnotes, so they appear in the year they are received, rather than pledged.
According to its proposed budget, the secretariat says it needs approximately 33m Swiss francs (£30m) to complete AR7 – equivalent to just over 8m Swiss francs (£7.2m) per year. At the level of contributions received in 2025, the IPCC trust fund would be “depleted by the end of 2028, preventing the IPCC from delivering the AR7 reports”, the document warns.
As a minimum, the IPCC needs 5.8m Swiss francs (£5.3m) per year in contributions to “enable the successful completion of the cycle”, the secretariat adds, although this would leave “no financial buffer for unforeseen circumstances”.
At the meeting, “many countries expressed concern” about the IPCC’s financial situation and “called for existing donors to increase their contributions”, says the ENB.
Iceland and Ireland said they “anticipated” making contributions soon, the ENB says, while others – including Azerbaijan, China, Finland and the EU – “assured” the IPCC of their continued support.
However, according to the ENB summary, the meeting struggled to agree on a number of points, including whether budget discussions should be paused until a timeline for AR7 had been agreed, and the work of the “Financial Task Team” (FiTT).
An FiTT is assembled for each assessment cycle for “reviewing income and expenditures, assisting in preparing the budget proposals and developing other recommendations related to finance for consideration by the panel”, according to the IPCC.
The IPCC established the FiTT for AR7 at its 60th session in Istanbul in January 2024. It is co-chaired by Kenya and Australia, with core team members from the Bahamas, Norway, Saudi Arabia and the US. Other government representatives can also participate.
Midway through the meeting, FiTT co-chair Patricia Nying’uro from Kenya reported to the plenary that they had been “unable to advance” on reviewing the budgets beyond 2026, with disagreements on whether “budget lines should remain or be deleted”, reports the ENB.
The attempts to remove whole budget lines drew expressions of “deep concern and disappointment” from other countries, the ENB says, with many noting that the work of the FiTT is “purely technical and has never been controversial”.
Belize, France, Germany, Panama and the Netherlands specified that “two delegations have threatened to have no budget if their conditions are not met”, the ENB says, with even budget lines for activities already agreed being “taken hostage”.
The ENB notes that both Saudi Arabia and India expressed concern that certain lines in the 2027 budget “favoured one option” for the AR7 timeline. Both countries opposed using the proposed budget as a basis for discussion in the FiTT, the ENB says.
The ENB quotes Prof Skea, who noted that most of the budget lines apply to all options for the timeline and was puzzled that members seemed to “disagree about things they don’t disagree about”.
A way forward was found for the FiTT to continue its discussions – despite Saudi Arabia’s accusing the secretariat of “blatant bias”, says the ENB. However, further disagreements arose over budget lines for approval meetings in 2027 to extend financial support for developing countries to bring two representatives rather than one.
Skea noted that despite “almost universal” support for increased participation of developing countries in IPCC sessions, the plenary discussions had continued for two and a half hours without agreement. He asked the timeline contact group to take up the issue.
Discussions on the rest of the budget continued in the FiTT, but “there were no further reports to plenary” before the meeting was suspended, the ENB says.
In the evening of the meeting’s final day, with no agreement on the timeline, Skea asked delegates to consider approving the proposed budgets for 2026 and 2027 to allow work on AR7 to continue. According to the ENB, Saudi Arabia, India and Russia objected.
The Democratic Republic of the Congo warned that the delay to agreeing the budgets was an “immediate danger to the viability” of the panel.
Credit: ENB
How have scientists reacted?
According to the ENB, Skea said in his closing address that if there were stronger words than “profound disappointment,” he would use them.
He noted that recent months had been “some of the most turbulent for the panel” and characterised IPCC-65 as “among the most difficult IPCC meetings”, it says.
It remains to be seen when discussions on the AR7 timeline and budget will restart.
A statement from Skea released after the meeting detailed plans to reconvene the plenary “at the earliest opportunity so the panel can effectively conclude its business”.
The IPCC declined Carbon Brief’s request for an interview with Skea, with a spokesperson noting it could not “discuss a plenary that had not yet finished its work”.
On Monday after the meeting, nearly 120 scientists currently drafting the AR7 reports had signed an open letter where they pledged to “keep working” on the reports. By Wednesday, more than 200 scientists had added their names.
The report authors said they remained committed to delivering the assessments “in a manner that is rigorous, timely and inclusive”. The letter also stresses the “world should not have to face” the second global stocktake without the IPCC’s latest reports.
“Urgent agreement is now needed to give planning security to the voluntary [report] authors, more than half of which are from developing countries and to safeguard a timeline that sees all WG reports published in 2028, in time to inform the second global stocktake.
“A resumed IPCC-65 session still this year, whichever way possible, is essential to at the very least agree to the IPCC’s 2027 budget to allow work to continue.”
Prof Fabrice Lambert, a contributing author of WG1 and Chilean government delegate, wrote on LinkedIn that there was now a “clear need for professional negotiators and experienced facilitators” from UN bodies to help organise and run panel sessions. He continued:
“Discussions have moved too far beyond the scientific questions that should be the IPCC’s mainstay. Bringing in that expertise could facilitate breakthroughs that have eluded us for so many sessions during AR7.”
The IPCC has also not yet confirmed the date and location of its next plenary session, IPCC-66.
What other issues remain unresolved?
After every IPCC meeting, the organisation writes a report summarising the discussion. However, the IPCC has failed to sign off on the reports for its past four sessions.
The first item on the agenda for last week’s IPCC meeting was the approval of these four draft reports. To open the discussion on Monday, Skea proposed developing guidelines for producing these meeting reports, the ENB says.
Belgium, France and the Netherlands suggested publishing full transcripts of the meetings in addition to the meeting reports. Panama supported the idea, “emphasising the need for transparency”, the ENB says. However, it adds that Saudi Arabia opposed the suggestion.
Saudi Arabia said it objected to the “highly selective manner” in which the draft meeting reports outlined certain topics in more detail than others, arguing that this “distorted [their] balance”.
Meanwhile, Belgium and Germany opposed the use of the passive voice in the meeting reports, arguing this gave an impression that all opinions are “equally supported”. They said that this resulted in a “skewed way of presenting discussions”, the ENB says.
IPCC vice-chair Prof Diana Ürge-Vorsatz led a huddle to continue the discussion. She reported back at the end of the week that the group had reached a “preliminary agreement” for the “principles for producing meeting reports”, the ENB notes:
“She said the discussion focused on how to record diverging views, as some panel members felt separating out two or three countries invited unwelcome attention or misinterpretation.
“Ürge-Vorsatz said the huddle agreed to group arguments, first identifying countries representing one perspective and elaborating on their argument, then identifying countries representing the other side of arguments. The secretariat will decide on the granularity of reporting.”
The panel requested that the IPCC secretariat prepare revised draft meeting notes in accordance with the new guidance. It decided to “submit the revised draft reports for consideration by the panel at its next working session”.
The IPCC also agreed to defer consideration of a review of its principles and procedures to a future meeting, after India, Iran, Russia and Saudi Arabia argued that there was not enough time to discuss the issue thoroughly, the ENB reports.
These documents are typically reviewed every five years. However, the last review was delayed due to the Covid pandemic.
WG2 co-chairs Prof Winston Chow and Prof Bart van den Hurk explained to the conference floor that the second order draft of the special report was circulated for review earlier this year and received more than 25,000 comments. They added that the fourth lead author meeting was held in India in August and a final draft version is currently being prepared.
In his closing address on Friday, Skea said the progress reports presented by working-group chairs were evidence of the “ongoing achievements” of the IPCC, according to the ENB.
Noting that the IPCC’s scientific work was progressing, Skea reportedly stated that, despite challenges on the timeline issue, “the engine room of the IPCC is working – and it’s healthy”.
The sentiment was echoed by WG1 co-chair Dr Robert Vautard, who wrote on LinkedIn after the meeting that disagreements across country delegations on the AR7 timeline were “so disconnected from the actual collaborative spirit across authors in the report making”.
Nevertheless, the ENB’s meeting summary cautions that the ongoing inability of governments to reach agreement on “critical issues” was now “spilling over to authors”, with “serious potential consequences both for the IPCC and for the science the world depends on”.
It continues that, “while the work of scientists and governments is separate, the friction that has characterised AR7 plenary sessions is now directly affecting the ability of the scientists to carry out their work”.
How might the IPCC use AI and enhance gender inclusivity?
He added that the IPCC bureau – made up managerial roles including the IPCC chair, vice-chair and co-chairs of working groups – decided to establish an “informal group” on AI that would provide advice on the “responsible use of AI and emerging digital technologies in IPCC processes”, reports the ENB.
India and Russia “expressed concerns about lack of transparency in [the] use of AI in scientific research”, according to the ENB. Burundi and Iran highlighted the need for disclosure and human oversight in AI-produced material, while Congo asked how AI has so far been used by the IPCC.
After discussions about the terms of reference for the informal group on AI, the panel agreed that any guidance or recommendations that it produced will be “subject to subsequent confirmation by the panel”, reports the ENB.
Separately, vice-chair Ürge-Vorsatz reported on the progress of the IPCC’s gender action team, explaining that the group had finalised the report of the IPCC expert meeting on gender, diversity, equity and inclusivity.
The gender action team “aims to enhance gender equality in IPCC processes through monitoring, accounting for gender in decisions, including diversity in nominations and appointment”, according to the IPCC website.
Dozens of countries commented on the update. Luxembourg, supported by nine other countries, invited the panel to consider expanding the mandate of the gender action team to include “gender diversity, equity and inclusion”, reports the ENB.
However, Saudi Arabia opposed expanding the mandate of the gender action team, arguing it risked “diluting” the team’s work.
To close the discussion on Tuesday, Skea suggested the gender action team consider expanding its mandate and bring suggestions to IPCC-66.
Supersized heat waves used to be rare. No longer. A Reuters data-driven examination of global temperatures reveals the emergence of what we call ‘heat storms.’ They now swarm the planet, upending life around the world from Texas to Ivory Coast to Poland.
Firefighters battle Europe’s largest wildfire on record, which spread during a heat storm in summer 2023 in Greece’s Dadia National Park. REUTERS/Alexandros Avramidis
Maps illustrate the growing size and reach of heat storms across the globe.
In less than a lifetime, about six and a half decades, heat waves have evolved from sporadic, isolated episodes to massive, cascading weather events that can blanket vast swaths of the planet for months or even years at a time.
The largest heat wave of the 1960s reached about 6% of the planet as it drifted for 46 days over the Atlantic Ocean, parts of Latin America and West Africa.
The largest one this decade swept across almost 85% of the globe during its 367-day lifetime, stretching into the polar latitudes and touching virtually every human being on Earth.
An animation shows a heat wave moving around the mid-Atlantic, spanning areas between Central America and West Africa.
To continue calling such sprawling phenomena “heat waves” would be like calling a monsoon a rain shower. To understand the new reality of these immense expanses of heat, Reuters consulted with two dozen meteorologists and climate researchers – and analyzed temperatures for every hour of every day since 1960 for more than one million grid squares covering the entire globe. Out of those results we developed a new concept: the “heat storm.”
Simply put, heat storms are supersized heat waves. Reuters uses a conservative benchmark to define an ordinary heat wave: five or more consecutive days of abnormally high temperatures. Then we define the largest heat waves – those that at their peak cover at least 1% of the globe, roughly half the size of Canada – as heat storms.
Applying this definition, Reuters identified a startling transformation. Heat storms were rare in the 1960s, when they averaged about five active storms per year. This decade, heat storms have swarmed the planet, averaging about 102 a year.
One of them caused this summer’s deadly heat in Europe, notching the highest June temperatures ever recorded in Bordeaux, Budapest, and Bilbao, the latter of which hit 108.9 degrees Fahrenheit. (Show Celsius.) Roasting Europe for most of June and July, the heat storm warped steel railroad tracks, drove what European governments estimated were more than 10,000 excess deaths in a single week, and prompted Belgian zookeepers to cool their tigers by feeding them raw meat encased in ice.
The largest heat storm on record – the one that lasted for 367 days over 2023-2024 and reached almost 85% of the globe – sent temperatures in parts of the Philippines up to 100.5 F, immiserating students in the country’s mostly unairconditioned schools. In Manila, math teacher Joel Arrojado recalled a student who was face-down at his desk, crying. When Arrojado coaxed the boy into looking up, Arrojado saw his face and chest covered in blood. The boy, Arrojado said, had suffered a spontaneous nosebleed, a commonproblem in extreme heat. As more blood pumped from his nostrils, Arrojado recalled, the child pleaded with Arrojado to call his mother.
As the same heat storm encased southern Louisiana for months, it exacerbated a historic drought. The state burned, suffering what the governor called “the worst wildfire we’ve seen since at least the Second World War” – and close to 600 other fires in August alone. Two months later, smoke from a smoldering marsh fire developed into a blinding “super fog” that caused a highway pile-up of at least 158 cars, killing seven people.
The same heat storm helped fuel a 2023 Mediterranean cyclone that dropped massive rains over Libya. In the dead of night, two dams burst near the coastal town of Derna, washing away entire neighborhoods. A week later, on her hands and knees, Sabreen Blil clawed with her bare hands at the rubble of her brother’s house, hoping to dig out the dead. As she wept, she recited their names. “Taym, Yazan, Luqman, Salmah, Tumador, Hakim and his wife. Oh my God. My family, where are you?” she wailed. “Oh God. Even just one – my God – just let me find even one body.”
During the annual Hajj pilgrimage in Saudi Arabia, the heat storm drove temperatures at Mecca’s Grand Mosque to 125.2 F, killing more than 1,300 pilgrims. Bodieslay in the streets. And as Europe’s largest recorded fire spread in Greece during the span of that same heat storm, the inferno killed migrants traversing the thick Dadia forest. Their corpses remained in the woods for days; even the vultures that might have fed on the bodies had fled the flames.
These far-flung and seemingly unrelated catastrophes all shared a common connection: one single heat storm.
After heat storms battered Bolivia in October 2023, Lake Titicaca, one of South America’s largest lakes, edged towards record low levels. REUTERS/Claudia Morales
Heat storms inflict “a major disruption to global civilization,” said James Renwick, professor of physical geography at New Zealand’s Victoria University of Wellington. “Heat is the biggest killer from climate change,” he said, noting that it can lead to “famine, hunger, death, conflict over resources.”
Heat storms constitute a new way to conceptualize how Earth’s climate is changing as humanity pumps greenhouse gases into the atmosphere.
The determination by climatologists that the average global temperature in 2025 was 2.6 F above the pre-industrial level is useful to scientists and policymakers, but to many people it’s so abstract as to be meaningless. What people do understand are heat waves – in Spanish, an “ola de calor”; in Arabic, a “mawjat hararia”; in Mandarin, a “rè làng.”
Yet no universal definition of a heat wave exists. For example, authorities differ over duration: The US National Weather Service defines a heat wave as “abnormally and uncomfortably hot” weather typically lasting “two or more days,” while the meteorological agencies of Australia, France, Spain and the United Kingdom set the threshold at three days. Some of those countries also have different rules for what temperatures qualify as a heat wave.
Moreover, what constitutes a heat wave in Anchorage, Alaska, would not be considered a heat wave in Kuwait City. The climate in those places differs sharply, and a heat wave is an unusual period of hot weather for a particular place.
Heat storms intensify wildfires. Here, a tree burned in Mandra, Greece, in July 2023, when heat storms struck the area for most of that month. REUTERS/Louiza VradiVolunteers gave water to a sheep saved from a burning farm during a wildfire in the village of Hasia, near Athens, in August 2023. REUTERS/Alkis KonstantinidisA man walked on the cracked ground of a reservoir that was drying up during the extreme drought in Catalonia, Spain, in 2023. REUTERS/Nacho Doce
Such dueling definitions made it all but impossible to gauge whether heat waves are more common globally now than in the past. But the system Reuters has devised delivers the precision needed to do so.
We began by determining the historical temperature range for each day of the year from 1961 to 1990 for each of 1.04 million grid squares around the globe. The temperature data comes from the European Centre for Medium-Range Weather Forecasts, and our detailed methodology is here. A heat wave occurs when the daily high temperature for any location reaches at least the 90th percentile of the historical range – in other words, the hottest 10% – for five days in a row.
So a July heat wave in Istanbul is five or more consecutive days of at least 78.1 F, whereas one in Phoenix, Arizona, needs to reach 104.5 F.
Heat storms constitute a new way to conceptualize how Earth’s climate is changing.
In Reuters’ definition, a heat wave can happen in any season, so in Istanbul a January heat wave would be five consecutive days of at least 52.2 F, and in Phoenix the threshold would be 69.4 F.
To measure the geographical extent of heat waves, we map neighboring heat wave grid squares that form a cohesive area. Because very large heat waves exert an expanding effect on the planet’s rising temperatures, Reuters created a further definition: If at any point a heat wave covers 1% of the planet, we designate it a heat storm and give it a name, as is done with tropical storms and hurricanes.
Naming events is “a powerful way for people to understand that something is happening,” Anthony Leiserowitz, director of the Yale Program on Climate Change Communication, told Reuters. Yet “we don’t currently do that with heat waves,” he said, even though heat has been estimated to kill almost half a million people a year. Defining and naming heat storms is an important innovation, he added, and could help people to track impacts and consequences.
As their numbers have swelled over the decades, heat storms and heat waves have not only caused individual catastrophes. They have also helped drive lasting, systemic effects.
They exacerbate droughts and intensify fires – from Chile to Europe, the largest conflagrations on record happened during heat storms — and if they hit while a woman is pregnant, studies have found that she’s more likely to suffer everything from stillbirth to preeclampsia, a leading killer of expecting and post-partum women.
Warmer temperatures can usher in benefits, at least for some regions. Siberia now produces almost as much wheat as Russia’s traditional breadbasket near the Black Sea. That’s in part because increasing heat storms and heat waves have helped open far northern latitudes to agriculture while simultaneously contributing to droughts around the Black Sea that reduce production there.
Tens of thousands of ordinary heat waves flicker in and out of existence every year. Like heat storms, they too have escalated in number, more than doubling from the 1960s to this decade. The vast majority are small, however, never even covering 0.01% of the planet, about the size of a square that you could drive across in less than two and a half hours. Only a minuscule fraction of heat waves ever gets big enough to become heat storms, which currently number about 102 annually.
Yet heat storms now claim the lion’s share of Earth’s unusually hot days, those in the top tenth of the historical range. In the 1960s, heat storms averaged 3% a year for any given spot on the planet. (The remainder comprised ordinary heat waves and short bursts of heat that didn’t last long enough to become a heat wave.)
This decade, heat storms have accounted for almost 43% of such days.
This basic pattern recurs in cities around the world.
As greenhouse gas emissions have increased, so too have heat storms and ordinary heat waves. From the 1960s to this decade, the amount of carbon dioxide in the atmosphere has increased by more than 30%. Over that time, the average portion of the planet likely to be covered by heat storms and heat waves has risen more than 12-fold. El Niño – the periodic episodes of unusually warm water in the Pacific Ocean that drive major weather patterns – has acted as an accelerant, intensifying heat storms and heat waves.
Reuters has also examined the opposite of heat waves and heat storms – cold snaps and cold storms. They have dwindled. From the 1960s to this decade, the total area in a cold snap or cold storm on an average winter day has shrunk by almost 80%.
Despite their mammoth scale, the phenomenon of heat storms has been largely invisible to the public. Until recent decades, heat storms were rarer and smaller, so easier to overlook.
But perhaps as importantly, they hadn’t been defined and named in the sharp way that typhoons and hurricanes are. Yes, people in Montreal, Perth and Hong Kong knew they were baking in a stretch of hot days – but they did not know that they were all experiencing one unitary heat storm.
As part of this series examining heat storms and their effects, Reuters has integrated an online Heat Storm Tracker into our regular weather coverage. With this tool, people can see what had previously gone unnoticed and, using that knowledge, tally the impact of heat storms as they rage across regions.
The rapid evolution of heat has not only made individual heat storms larger. Now, in sharp contrast to even 20 years ago, massive heat storms often run into, merge with, and split off from one another. What has become common are heat storms that run consecutively, simultaneously or both – heat storm cascades.
The duration of these cascades swelled from an average of roughly two months in the 1960s to nine months in the first part of the 2000s. Then, on July 6, 2008, a cascade started – and has not stopped. For more than 18 years straight, 6,681 days, at least one heat storm has been roiling the planet, and often more than one.
Many of them have been colossal. In the almost half century between 1960 and the start of the current 18-year cascade, only five heat storms had ever reached a peak size bigger than a tenth of the planet, about the area of Europe and Asia combined. The ongoing cascade has included 56 such mega heat storms.
The largest of these was one we have named Heat Storm Vera, the storm that killed the Hajj pilgrims, intensified the Louisiana wildfires, and touched nearly 85% of the planet over its 367-day lifespan.
See original article for an animated map showing a cascade of heat storms evolving across the globe from 2023 to 2024.
What follows is a biography of Vera, from its birth on July 29, 2023, through its last day on July 29, 2024, that shows how the storm connected and often devastated communities scattered across the globe.
The sun rises as the Austin High School football team holds an early morning practice to avoid peak summer heat in Austin, Texas, this August. REUTERS/Callaghan O’Hare
TEXAS
AUGUST 2023
Friday Night Lights
On a sweltering evening in August 2023, as Heat Storm Vera coiled around the American South, Justice Trumpler led his high school football team onto the field for the season’s first game. In hopes of playing college ball, the 17-year-old captain had invited coaches and recruiters to watch. As the game kicked off at 7:30pm, the temperature was 100 F.
Texas’s passion for high school football – the American version, not the one that transfixed billions during this summer’s World Cup – was captured in the iconic movie Friday Night Lights, about a team gunning for the state championship. But players of American football suffer 11 times as many heat-related injuries as athletes in all other sports combined, according to data from the University of Connecticut Korey Stringer Institute.
That’s because the football season starts in the summer, the sport demands intense physical exertion, and many of its players, such as the 255-pound offensive lineman Trumpler, have large bodies that are harder to cool down. There’s another reason: In contrast to most other warm-weather sports, American football features heat-trapping helmets and heavy padding.
Justice Trumpler hoped college recruiters would see him shine in the season opener. Handout via REUTERS
Meanwhile, Texas is spending ever more time in extreme heat. Across the entire three decades from 1960 to 1989, the city of Austin – home to Trumpler’s Bowie High School – experienced a total of 557 days in heat waves and heat storms. This decade, our analysis shows, that number has already surpassed 700. During the football season, Heat Storm Vera enveloped Austin for 39 days.
By halftime, the Bowie Bulldogs led 14-0. As the teams broke apart, Trumpler’s parents said they watched him walk to the field’s edge and vomit. A trainer hosed it away.
Through the second half, they saw Trumpler stop to dry heave at least twice, they said, and he seemed slow to get off the turf after plays. At one point, his parents saw him walk toward the opposing team’s huddle, a sign of mental confusion.
Trumpler’s father, Jason Trumpler, told Reuters he began shouting from the bleachers for the coaches to pull his son out of the game. When that didn’t work, he went to the railing separating the stands from the field and continued yelling. Trumpler kept playing.
Bowie’s coach did not respond to multiple requests for comment. The school’s principal referred Reuters to a school district spokesman, who did not respond to messages seeking comment.
Other players struggled, too. So many boys fell to the ground after plays that Bowie’s game commentators declared it a “cramp epidemic.” By the fourth quarter, the score was tied. The coaches could have ended the game, but the referee announced it would go into overtime.
“Both teams are going to start on the 25 (yard line) of their opposition and, uh, keep on going until somebody dies,” said one commentator. “Don’t say that,” replied the other.
A fan cools a person on the sidelines after a football game between Austin High School and Bowie High School in Austin, Texas, on September 25. REUTERS/Callaghan O’Hare
The “unfortunate” phrasing, the first commentator told Reuters, was intended to convey how hard the teams were playing under very tough conditions.
Just before the second-to-last play, a leg cramp brought Trumpler to the ground. His coach finally pulled him out. “Texas has that mentality of ‘you gotta be tough, you gotta play through pain,’” said Jason.
Jason took Trumpler home, but the teen couldn’t stop throwing up. After an ice bath failed to cool Trumpler down, Jason said, he drove his son to the hospital. When Trumpler’s mother, Jennifer Norman-Wolfe, arrived, he couldn’t recognize her voice, she said, and couldn’t tell her how many brothers he had.
Reuters was not able to review Trumpler’s medical records, but his parents said tests showed their son’s kidneys were failing, and his brain was swelling so much that doctors said they might need to drill a hole in his skull to relieve the pressure. Doctors couldn’t assure them their son would survive, they said.
Twelve high school football players in the US have died from heat over the last decade, according to statistics compiled by the National Center for Catastrophic Sport Injury Research. Trumpler was released after two days in the hospital.
Later, his parents said, they sent his medical records to Bowie High School and tried to discuss what had gone wrong. But they said the school’s athletic staff infuriated them by suggesting they hadn’t done enough to alert coaches about his condition during the game, Trumpler’s father told Reuters. A month later, Trumpler’s parents saw a video of an official for the governing body for Texas high school sports, the University Interscholastic League, referring to their son’s injury as a “questionable” heat-related incident.
Neither the school district nor the UIL official responded to requests for comment. In an interview, UIL Executive Director Jamey Harrison said that the organization relies on schools to report the details of medical events and avoids instituting rules based on individual incidents.
Bowie High School junior Alec Zarate, 16, wears a cold towel around his neck while a misting fan sprays behind him during a game against Austin High School in Austin, Texas, this September. REUTERS/Callaghan O’Hare
Heat storms are starting to reshape Texas football. State filings show that since the beginning of this year, at least four school districts in the state have registered to build indoor or covered practice spaces, two with price tags in excess of $10 million. And this year, the UIL adopted new rules requiring rest breaks, ice baths and shortenings or cancellations of practices in high heat. Yet these rules are not mandatory for football games themselves.
Two weeks after his hospitalization, as Heat Storm Vera continued baking Austin, Trumpler resumed playing football. He declined to speak to Reuters, but wrote on X at the time that “it felt so good to return.”
His heat injury, however, seemed to throw off college recruiters. Division I teams, those with the biggest and best programs, stopped calling, Trumpler’s father recalled. His mother said Trumpler himself seemed tepid about the prospect, which had once been his dream, and he is not currently attending college.
“Our son, Justice, is tough and built for football, mentally and physically,” she wrote at the time. “This heat was tougher.”
Olives in Spain’s Carcabuey district, which Heat Storm Vera battered in 2023. REUTERS/Marcelo del Pozo
SPAIN
AUGUST 2023
“Pirates of liquid gold”
Early in the morning of August 30, 2023, a team of thieves carefully maneuvered a tanker truck through the winding roads of the small Spanish town of Carcabuey, intent on an audacious heist.
Around 2:30 a.m., according to a local police official, the thieves pulled into a gated driveway on the edge of town. They shut off lights and cut through the rickety gate’s padlock. They left no fingerprints because they wore gloves. Once inside, they beelined for their unusual loot: extra virgin olive oil, which they siphoned from industrial-sized tanks.
Over the previous several months, a series of heat storms had helped fuel a drought that devastated the olive crop and sent the price of olive oil skyrocketing, raising the value of the stolen 56 metric tons to a staggering $546,000. Journalists nicknamed the thieves the Pirates of Liquid Gold.
By dawn, when the olive mill’s manager, Martín Parras, arrived, all that remained was an oil streak on the driveway. Parras was so overcome, he said, that he walked into a nearby olive grove, sat beneath the trees and wept.
Scene of the crime: Manager Martín Parras stands in the warehouse where the extra virgin olive oil was stolen. REUTERS/Marcelo del Pozo
In Spain’s southern region around Carcabuey, houses are painted white to repel the sun, and residents take long breaks during the hot afternoons. Olive trees, which evolved to have formidable heat tolerance, have thrived here for centuries.
But the trees have struggled in the new heat storm era. Between March 12 and August 1, when Vera first reached Carcabuey, the little town endured eight other heat storms that rolled one into the next with almost no break. The combination of early heat and scarce rain burned the olive flowers, preventing them from opening and pollinating, Parras said.
In the past five to 10 years, locals say, the heat that once lasted from June to August can now stretch from April to November, and it stays hot later into the evening. “People used to go out around 7 or 8 in summers but now it can be 9 or 9:30 and there’s no one in the streets,” said Juan Lara, who has been running a pastry shop near Carcabuey for 46 years.
Olive groves in Spain’s Carcabuey district, where heat storms exacerbated a drought that sent the price of olive oil skyrocketing. REUTERS/Marcelo del Pozo
Heat Storm Vera spent a total of 127 days over Carcabuey. At first, Vera teased the city, brushing it briefly before smothering the town for three weeks until a couple of days before the theft. Temperatures surpassed 104 F several times.
Parras said about half the olives at his Marín Serrano El Lagar mill ended up unusable, with scant flesh and barely any oil. “Olive trees had a yellow color like when grass dries, and leaves fell,” he recalled. “It was monstrous.”
The price of olive oil spiked by 44% that year, largely because the withering heat and drought reduced output by roughly half in Spain, the world’s largest producer. Some Spanish supermarkets placed alarms on bottles to prevent theft, but no one thought thieves would target a mill’s huge metal tanks. At Marín Serrano El Lagar, the only alarms were in the office and a room storing oil that was already bottled.
Spanish police arrested a 50-year-old man who had crushed olives at the mill for more than two decades until he left in 2020 after an argument over his salary. Authorities suspect the man conspired to steal the oil with the owner of a rival mill just down the road, the local police official said. The owner was also arrested. Both men were later released, but remain under investigation. They told Reuters they did not steal the oil, with the 50-year-old suspect saying he’s eager to clear his name.
The real culprit, according to mill manager Parras, has not been indicted. “Climate change has completely affected olive fields,” he said. “Has it been the cause of the theft? Totally.” If the heat hadn’t driven the price so high, he added, “There would have been no theft.”
A dead river dolphin floating in Lake Tefé. In 2023, Heat Storm Vera helped heat the lake to jacuzzi temperatures. REUTERS/Bruno Kelly
BRAZIL
SEPTEMBER 2023
Death of the dolphins
On September 28, 2023, an environmental scientist named Ayan Fleischmann sat in a small boat at the heart of a vast Amazonian lake and watched a dolphin die. The creature moved in jagged circles, its body shuddering. Suddenly, it sank.
As Fleischmann watched the dolphin vanish into the darkness of Lake Tefé, smoke from nearby wildfires fouled the air. An algal bloom stained the waters blood-red. Dozens of dolphin corpses dotted the lake’s beaches. Photos taken by Fleischmann’s team show the dolphins’ skin wrinkled like melted plastic or peeling to reveal yellow flesh. This, Fleischmann recalled thinking, might be what Hell looks like.
With its distinctive pink color, the river dolphin is one of the Amazon’s most iconic and mythologized creatures. According to legend, they would transform into humans at night to seduce locals. The dolphin was put on the endangered species list eight years ago, after poaching and rainforest destruction caused its numbers to plummet. Now, heat storms have emerged as an urgent new threat.
Heat Storm Vera settled over the Amazon in late July and spent a total of nearly seven months over Lake Tefé, simmering it for 42 days in a row before Fleischmann watched the dolphin die. The heat storm helped warm the lake to a high of 105.8 F – the temperature of a standard jacuzzi. For the dolphins, such temperatures were catastrophic: Fleischmann’s research team later concluded that the creatures had effectively been cooked alive.
Fleischmann, 34, had been studying how the waters of Lake Tefé were changing for several years. “It became a friend. I see it every day,” he said. But as the heat storm festered, he recalled, so much of the lake dried up that when he and his colleagues visited local villages, they were pelted by whirlwinds of sand whipped up from newly exposed beaches.
Retrieving dead dolphins from Lake Tefé on October 2, 2023. REUTERS/Bruno KellyLake Tefé’s water level dropped during Heat Storm Vera in 2023, exposing wide swaths of dry land. REUTERS/Bruno Kelly
Later, Fleischmann’s research institute autopsied the dolphins that washed up on the lake’s shores. Temperatures were so high that most of the corpses rotted faster than they could be studied, said Hilda Chávez-Pérez, a biologist who oversaw the study of the bodies. But in those they could examine, she said, the researchers kept finding that the dolphins’ brains were bruised and their organs speckled with blood: telltale signs of heat stress. Many had empty stomachs, indicating that they had lost the ability to feed in the heat.
The lake’s normal population is at least 700 river dolphins, said Fleischmann, but during September and October, when Vera scalded the area for all but one day, more than 200 died. “Nobody had ever seen that magnitude of mortality,” he said. “It was like a death circus.”
Hikers in Poland’s Tatra National Park, which Heat Storm Vera heated to spring-like temperatures in January and February 2024. REUTERS/Kacper Pempel
POLAND
FEBRUARY 2024
Weird winters
In Poland’s Tatra Mountains, the first months of 2024 were full of strange signs.
In mid-February, bears woke from hibernation weeks early and began roaming the ranges. Frogs also emerged early, ranger Tomasz Zając remembered, rising from the rotting leaves and underwater silt beneath which they had hunkered through the cold months. Although snow still clung to the high peaks, it was so slushy that it frequently slipped in mini-avalanches that left black scars of exposed rock.
While heat storms are hottest in summer, they can also mellow typically frigid months. Heat Storm Vera, which enveloped the Tatras for 31 days in January and February, sent the temperature up to 62.6 F. In neighboring Germany, Vera prompted the country’s weather service to tell people to leave behind their warm coats as they celebrated the annual winter carnival, while in the UK, television hosts joked about unusually “hot and steamy” Valentine’s Day temperatures in the English town of Lover.
But winter heat storms can also cause unexpected harm. In January of this year, bone-chilling cold descended on eastern Canada and the US, causing Toronto to endure 23 straight days of sub-freezing cold and gripping parts of Minnesota in temperatures that plunged to minus 43 F. Paradoxically, those frigid temperatures were largely caused by winter heat waves in the northern polar regions and Greenland, where residents basked in unseasonably balmy days that reached 52.3 F. That northern mass of warm air shunted a cold front down over eastern North America, in what climatologists call a polar vortex.
Tennessee got so cold that freezing rain encased trees in ice, causing many to collapse under the weight. Forest silence was occasionally broken by a sound resembling gunshots as sap froze and expanded, splintering tree trunks. Almost half of Pittsburgh’s snowplows stopped working, forcing the city to leave roads uncleared for days and trash on the streets for a week. The extreme cold killed at least 100 people across the US, caused power losses for more than a million customers, drove more flight cancellations than any time since the COVID 19 pandemic, and forced school closures from Milwaukee to Atlanta. Even in Cuba, farmers woke to frost on their crops.
Heat Storm Vera spent weeks over Europe in January and February, effectively transforming winter into spring. By mid-February, local media estimated that a quarter of European ski resorts were closed due to lack of snow. It was so warm that many resorts couldn’t even make artificial powder. Amid the unseasonal heat, intense rain battered parts of Europe and neighboring countries; as Vera settled over the southern Turkish city of Antalya, flash floods inundated entire neighborhoods, with the sudden flow submerging fire trucks and buses.
Poland’s Tatra Mountains are famous for their spring crocuses, which carpet the valleys in purple. Agencja Gazeta/Marek Podmokly via REUTERS
In the Tatra Mountains, winter’s end is marked by the blossoming of crocuses. More than a hundred million typically bloom starting in mid- to late-March, turning the valleys into purple oceans. Rangers have to patrol Chochołowska Glade – where crocuses bloom most densely – to prevent tourists stampeding the flowers into the ground. In late February, that frenzy was supposed to be weeks away. But on February 29, ranger Zając noticed a flash of color amidst Chochołowska’s freshly-uncovered grass. He squatted to get a better look at the purple petals unfolding around an orange stamen. Earlier than he could ever remember, the first crocus had arrived.
Learning first aid for people who pass out from the heat, students at Sagkahan National High School in the Philippines practice carrying an unconscious person. REUTERS/Eloisa Lopez
THE PHILIPPINES
APRIL 2024
Collapsing children
Partway through April 2024, on the top floor of Saghakan High School in the Philippines’ Tacloban City, Jeffrey Polido was giving a lesson on typhoons to 50 sweating 14-year-olds in a concrete classroom with two rickety fans. Interrupting him, a student burst through the door with news: Yet another child had collapsed.
Polido – a 37-year-old science teacher who at that time ran the school’s health clinic – raced outside to the stricken girl and took her vitals. He and the student who had alerted him carried the child down several flights of stairs to the clinic to recover. Then, Polido walked back to his roasting classroom and resumed teaching, knowing it would happen all over again. After all, he told Reuters, he had already seen dozens of students collapse.
Polido, along with the teacher who comforted the student with the gushing nosebleed, didn’t know it but they were on the frontline of Heat Storm Vera. The heat was intensified by its most dangerous accomplice: humidity. As a result, a temperature of 91.4 F in steamy Tacloban felt like 116.5 F, while hotter parts of the Philippines reached “feels like” levels of 127.5 F – at which point the country’s weather bureau warned, “Heat stroke is imminent.” For the millions of children in schools cooled only by fans, the thick muggy heat was especially dangerous.
In Manila, Jenita Simangan, 49, watched her 50-strong class of five-year-olds break out into rashes, complain of headaches, and say the heat made their skin feel like it was burning, she recalled. Partway through the heat spell, she said, one five-year-old staggered into class and began vomiting uncontrollably. As the child stood in a puddle of watery bile speckled with half-digested chunks of lunch, Simangan felt herself panic. It was the worst moment she’d experienced in two decades of teaching, she said.
A student is carried on a chair after feeling unwell during a first aid training at Sagkahan National High School in the Philippines this July. REUTERS/Eloisa Lopez
Heat Storm Vera hit Tacloban particularly hard, tormenting the city for a total of more than nine months. Standing in a rare spot of shade at Sagkahan High School earlier this year, Cloe Novella, 17, recalled sitting in her class and trying to catch the weak breeze from one of the classroom’s old fans when she realized one of her best friends hadn’t said anything since they sat down. Turning around, Novella saw her friend appeared to be sleeping, tapped her, and quickly realized that her friend had lost consciousness. Novella’s teacher told one of her classmates to fetch Polido, who came and carried the student to the clinic, where she recovered. She was one of more than 40 who Polido estimated fainted during Vera.
The Philippines soon suspended in-person classes across the country and has since changed its school calendar to avoid exposing students to the most extreme temperatures. Yet for Novella and her fellow students, Heat Storm Vera kickstarted a new era of near-perpetual heat.
LJ Corsanes, 17, said he frequently plunges into temperature-induced panic attacks. “It’s like going into a fight,” he said of going outside. “It’s like opening a door to hell.”
Over the entire 1960s, heat storms hit Tacloban for only 7 days, and the city experienced 245 days of ordinary heat waves. This decade, ordinary heat waves have touched Tacloban for fewer than 40 days – but heat storms have engulfed the city for more than 1,470 days. That’s over 200 days a year.
Students learn first aid for people stricken by heat at Sagkahan National High School in the Philippines.
Tropical archipelagos such as the Philippines are particularly vulnerable to heat storms, Reuters’ analysis shows. The warm ocean heats the air from below while sunshine not only beats down from above but also reflects off the surrounding water. Other tropical islands, such as Guam, Haiti and much of Indonesia, spend most of the year in heat storms.
The implacable heat is forcing students to learn grim new skills. In July, Polido gathered 80-odd teenagers for first aid training with the Red Cross. Over two days, they practiced how to begin chest compressions, how to lift an unconscious person to safety, and how to position someone who has passed out so they don’t choke on their tongues.
For some, the heat has become eerily normal. Novella said it makes her feel “like I am burning in pain and fire.” But, she added, “It is not strange, because I am used to feeling the heat already.”
Cocoa pods on the farm founded by the late Brou Kouassi, village chief of Digrikro, Ivory Coast. REUTERS/ Luc Gnago
IVORY COAST
NOVEMBER 2023-JUNE 2024
Death of a chocolate chief
In his half-century as a farmer of cocoa, the raw ingredient of chocolate, Chief Brou Kouassi earned enough money to marry five times, send his 12 children to school, and float a little cash to his neighbors. He even had a nickname – “À Nous Deux”, or “Just Between Us” – a phrase people used when they came seeking assistance on the hush-hush.
In 2023, Heat Storm Vera started stalking Digrikro, the village in western Ivory Coast where Kouassi had been chief since 2000. At first, the heat storm came and went, lurking for up to two weeks at a time. Then, beginning November 1, Vera besieged the area for more than four months straight, sometimes sending temperatures above 100.4 F. “La secheresse” – the drought – also began in November, according to local farmers and United Nations rainfall data.
His sons had taken over the hard labor, but every morning Kouassi still pulled on his boots and went out to survey his 22-hectare farm. As Heat Storm Vera shriveled the pods and sent brittle leaves falling from the branches, Kouassi would shake his head and lament that he had never seen so many trees dead or dying.
His sons would go out in the morning with their machetes to cut down the cocoa pods but return just a few hours later because, as Kouassi’s son Jules put it, “the trees didn’t give us anything.” Jules said the cocoa yield on his portion of the land for February to April 2024, an important harvesting time, was less than half of normal.
Ivory Coast and its neighbor Ghana produce about 50% of the world’s cocoa. But in the wake of Heat Storm Vera, the harvest in those two countries fell by half.
Jules Kouassi inspects his cocoa trees two years after Heat Storm Vera helped decimate his harvest.
Vera had collaborators, among them disease, which preys more easily on heat-and-drought-weakened cocoa trees, and an El Niño, which likely intensified the heat storm. But Vera and the drought it exacerbated were relentless, with the heat storm lingering over one cocoa region of Ghana for a total of almost 11 months. In 2024, the price of cocoa almost tripled.
In West Africa, where almost all the cocoa is produced by smallholder farms such as Kouassi’s, the stakes were existential. Kouassi urged his sons to abandon cocoa for heat-resistant crops such as rubber or cashews. “He kept saying, ‘If you don’t replace the cocoa trees, you’ll have nothing to eat,’” recalled Jules.
In June 2024, as Heat Storm Vera took its last licks at Digrikro, Kouassi fell ill. After a medicinal broth of leaves didn’t work, his family sent him to the hospital, where Jules recalls doctors saying his father had malaria and anemia. He spent four days in the hospital, at a cost of 320,000 CFA, or about $525. But his illness persisted, so they sent him back again, paying another $525. They considered a third trip for the same high cost, but they couldn’t afford to keep sending him back because the devastation of the cocoa crop had beggared the family.
Remembering the late chief Brou Kouassi, relatives gathered in August at the home he built in his ancestral village of Lekissou, Ivory Coast. REUTERS/Luc Gnago
Kouassi’s one-story concrete farmhouse offered little respite from the clammy heat. As his illness worsened – swelling that began in his feet spread all the way up to his thighs – he struggled to sleep more than a few hours each night. With money scarce and Western medicine failing, he asked to go to the village where he was born to visit a traditional healer or, if that failed, to be buried next to his parents.
Kouassi traveled by bus, which typically takes more than six hours to cover the 200 kilometers, much of it on a choppy dirt road. Years earlier, before heat storms shattered his livelihood, Kouassi had built an eight-bedroom house there with a large patio, where he relished visits from relatives and neighbors, some whispering hush-hush stories. His sons hoped the more comfortable environment and traditional remedies might help him regain his health.
The healer took 90,000 CFA, a little less than $160, then refused any more payments even though he kept bringing clay pots filled with medicinal concoctions. But his father continued to decline, Jules recalled, and finally pulled his son aside and asked him “to take care of the family because he didn’t know if he would survive.” Kouassi died on October 27, 2024, aged 68. No official cause of death was given.
Jules Kouassi holds a photograph of his late father, Brou Kouassi, dressed in his chief’s regalia. Heat Storm Vera ravaged his cocoa farm in the village of Digrikro, Ivory Coast. REUTERS/Luc Gnago
His beloved cocoa farm has not recovered. Many of the trees are dead or sick, Jules said, and revenue remains in a rut. Kouassi himself lies buried under a blank headstone, his family unable to afford the finishing tribute: a name plaque. Nor have they been able to raise enough money for a traditional funeral, a fact that torments Jules. In the culture of their Baoulé ethnic group, the failure to hold that ceremony – complete with a sheep sacrifice, drumming and dancing – means his father cannot go peacefully into the afterlife. “It’s like his soul is still walking around here,” Jules said, “unable to rest.”
FEBRUARY, 2026
A new massive heat storm
Over its lifetime, Heat Storm Vera drifted over nearly 85% of the planet, more than any heat storm on record. Its peak size, covering 31.3% of the Earth on January 5, 2024, was also a record. Until this summer.
Heat Storm Raimundo, which broiled Europe and stoked massive wildfires in Spain and France, extended over 33.8% of the globe on August 28. Since its inception on February 28, 2026, Raimundo has touched almost three quarters of the planet’s surface.
Both heat storms spent time over a melting Nepalese glacier before its collapse on August 26 and the catastrophic flooding that followed, killing thousands. Early indications suggest that warming temperatures contributed to the collapse of the glacier and the bedrock beneath it.
The grim aftermath of the collapse of a glacier, which unleashed flash floods and mudslides on Devighat, Nepal, this August. REUTERS/Francis MascarenhasA school bag lies amid debris and mud following the collapse of the glacier above Devighat, Nepal, this August. The glacier had spent more than half of this decade in heat storms and heat waves. REUTERS/Adnan AbidiKalpana Thapa Magar, 43, displaced by the flash floods that followed the glacier collapse in Nepal, cries while sitting outside her temporary shelter this September. REUTERS/Adnan Abidi
Up until the glacier gave way, it had spent about 58% of this decade in a heat storm or heat wave, up from 7% during the 1960s. For the six days immediately preceding its collapse, the glacier had been blanketed by Heat Storm Raimundo.
Raimundo isn’t yet finished. It has been heating the earth for 235 days and counting.
By Maurice Tamman in New York; Ben Welsh in New York; Pete McKenzie in Sydney and Tacloban, the Philippines; Charlie Szymanski in Phoenix, Arizona; Delgerzaya Delgerjargal in New York and Ulaanbaatar, Mongolia; Robbie Corey-Boulet in Dakar, Senegal, and Daloa, Ivory Coast; Loucoumane Coulibaly in Abidjan and Daloa, Ivory Coast; Jeff Horwitz in San Francisco and Austin, Texas; and Joan Faus in Barcelona and Carcabuey, Spain
Data analysis: Maurice Tamman, Ben Welsh, Charlie Szymanski and Delgerzaya Delgerjargal
Graphics: Adrián Blanco and Feilding Cage
Photography: Callaghan O’Hare, Marcelo del Pozo, Eloisa Lopez and Luc Gnago
Photo editing: Corinne Perkins and Maye-E Wong
Video: Zachary Goelman, Francesca Lynagh, Louis Johnson, Holly Murtha, Joshua Sommer, Media Coulibaly and Adrian Portugal
Art direction: John Emerson
Visual Editing: Feilding Cage
Additional reporting: Anna Wlodarczak-Semczuk in Warsaw, Gleb Bryanski in Moscow and Hina Suzuki in Tokyo
On an afternoon in Seoul in March 2016, I watched a program I helped build put a stone on the fifth line of a Go board in what looked like a gift to its human opponent. Move 37 in game two of the five-game match looked so absurd that some commentators thought it was a programming glitch.
It wasn’t. AlphaGo won the game, ultimately triumphing 4-1 over Lee Sedol, one of the greatest professional Go players of all time. “I thought AlphaGo was based on probability calculation and that it was merely a machine,” Lee said afterwards. “But when I saw this move, I changed my mind. Surely, AlphaGo is creative.”
When Deep Blue defeated then reigning world chess champion Garry Kasparov in 1997, it did so by looking six to eight moves ahead per player and evaluating 200 million chess positions per second, using rules hard-coded by humans. Go is a vastly more complex game. A stone’s worth depends on how distant groups and territory unfold over dozens of moves. Computing even a fraction of the possible outcomes would take a supercomputer billions of years. To win, AlphaGo had to sense who was ahead at a glance and even invent moves no human had thought to play.
That is why many accounts of AlphaGo’s match against Lee portray move 37 as a flash of pure machine intuition. But that is a misunderstanding. It was actually AlphaGo’s powers of reasoning that made this creative choice—and these are powers that today’s AI lacks. If we want future AI systems to produce trustworthy results and really novel insights in fields like science and medicine, we need to equip them with genuine reasoning capabilities of this kind.
AlphaGo is made up of two systems. The first, its policy network, was trained to guess what move a strong human would play. This “intuitive” part regarded move 37 as nothing special—a play that had a roughly one in 10,000 chance of being made by an expert human player. What made AlphaGo choose it was the program’s search machinery, which looked beyond immediate plausibility and weighed the future consequences of proposed moves. It explicitly constructed and searched a game tree with thousands of branches, each representing a different possible future.
A well-known theory in the behavioral sciences, popularized by Daniel Kahneman, distinguishes between two modes of human thought: System 1 is fast, gut-level, effortless; system 2, slow, step-by-step, and deliberative. AlphaGo offered a striking machine analogue of that split. Its networks supplied the hunches—this move looks promising, this position looks won—and its search supplied the deliberation, testing those hunches against the moves and countermoves that would follow. As in human cognition, neither half works alone. Intuition alone would never have opted for move 37, and brute-force search would have struggled to sieve through all the many possible moves.
This is strikingly different from the way today’s AI models work. A large language model picks the next token, over and over. That amounts to system 1 in action—fast, associative, and surprisingly good pattern completion across almost every subject people write about.
Not long after ChatGPT debuted, the field realized that language fluency alone falls short of true usefulness. The apparent solution was to make models that deliberate: Instead of answering immediately, they can now generate intermediate steps that decompose a problem, carry forward partial results, and influence subsequent reasoning—a process known as chain of thought.
The gains have proved real, above all in mathematics and coding. But unlike AlphaGo’s search, this does not introduce a genuinely separate reasoning mechanism: The intermediate reasoning is still produced by the same next-token prediction process, iterated for longer before the model commits to an answer.
Three shortcomings prevent what chatbots do from qualifying as reasoning (in a way that a scientist might recognize). First, these models typically maintain no explicit, persistent, and inspectable epistemic state. There is no open ledger that lays out the hypotheses a model is considering, the confidence it has in various explanations, the evidence it’s weighing, and the unresolved questions it’s holding onto—all things that should be systematically revised as new information arrives.
Second, they lack a clean separation between what the system knows and how it manipulates that knowledge. Knowledge and reasoning are inextricably interwoven in the weights of the neural network—there is no independent, explicitly represented set of beliefs. Third, while the chains of thought chatbots produce look like deliberation, research has demonstrated that the bots often concoctthem after the fact, reaching an answer by one route but reporting another.
This is a problem because in the high-stakes applications we all care about, such as medicine, engineering, and scientific research, it matters not only what a system concludes but also how it arrives at its conclusion. When mistakes happen—for example, in medical diagnosis and treatment—we need to be able to pinpoint what went wrong: Was the system’s reasoning at fault, did it draw on invalid evidence, or did it make incorrect assumptions?
This is why I recently left my position at Google DeepMind. I believe we need a fresh approach to machine reasoning—one that draws on AlphaGo’s architecture. AlphaGo maintains a record of what it knows about a given position: the game tree. This data structure contains all the variations, the possible futures, that AlphaGo has considered, each move and position being annotated with judgments made by its neural networks. As its reasoning progresses, AlphaGo updates the game tree and eventually synthesizes the information in it to decide which move to make.
Similarly, for general reasoning a system should maintain an epistemic state that represents what the system holds as settled, what it doubts, what it has ruled out, which questions stay open. Reasoning can then be understood as a sequence of moves that change the epistemic state to advance knowledge and reduce uncertainty: deducing consequences, breaking problems into parts, and—crucially—deciding what question to ask, calculation to perform, or experiment to run next.
Of course, open-world reasoning is harder than playing a board game such as Go or chess. In the real world the current state of affairs is only partially known, the set of available actions is large and variable, and the consequences of actions are stochastic or unknown.
But recent advances in LLMs and other neural models now give us the capability to take on such general reasoning problems. For example, LLMs can suggest ways of tackling a problem on the basis of what is known and what resources are available. They can interact with tools via APIs or code and help assess whether a claim is supported by available evidence.
Most important, to keep the system honest, an independent part of the system must evaluate each move by how much it actually resolves uncertainty, updating beliefs only when the change is backed by evidence. Once these rules are enforced, the model can accumulate certified knowledge and improve its reasoning policy by learning from past reasoning experiences. You can think of such a system as the scientific method on steroids, with the purpose of producing knowledge that can withstand scrutiny.
I do not think we reach trustworthy machine intelligence by making system 1 bigger. Scale sharpens intuition, but it does not make intuition more deliberative. Move 37 mattered because a machine held a position, weighed the possible futures, and chose the move its artificial instincts would likely have rejected.
Society needs such creative moves in drug discovery, materials, climate, diagnosis—fields where the board looks nothing like a Go board and nobody hands us the rules. We will get such insights only from systems that reason—systems whose conclusions arise from an auditable sequence of evidence, inference, and belief revision rather than from a convincing story told after the fact.
Thore Graepel is chair of machine learning at University College London. He was a core member of the AlphaGo team at DeepMind and works to ensure that AI benefits human flourishing.
Essas avaliações do IPCC (Painel Intergovernamental sobre Mudanças Climáticas) serviram de base para todos os grandes acordos climáticos das Nações Unidas nas últimas três décadas. Porém, o impasse sobre as datas de publicação pode levar os países a decidir a próxima fase de ação sem um consenso sobre questões cruciais —incluindo até que ponto parte do aquecimento global poderia ser revertida—, segundo entrevistas com mais de uma dezena de cientistas atuais e antigos do IPCC, negociadores governamentais e observadores das reuniões.
Prédios cobertos com lama após as inundações que causaram centenas de mortes no Nepal, em agosto deste ano – Athit Perawongmetha/Reuters
Na próxima segunda-feira (12), representantes dos 195 Estados-membros do IPCC devem se reunir em Adis Abeba, capital da Etiópia, para uma reunião plenária que muitos consideram a última oportunidade de definir datas de publicação anteriores às negociações climáticas da ONU de novembro de 2028. É nessa ocasião que os governos deverão avaliar o progresso em direção às metas estabelecidas no histórico Acordo de Paris, de 2015.
A possibilidade de relatórios do IPCC não ficarem prontos para as negociações de 2028 preocupa países que defendem ações mais contundentes, pois as discussões orientarão financiamento climático e metas de redução de emissões.
“Os dados científicos recentes […] permitem uma abordagem justa, equitativa e baseada em evidências para enfrentar o maior desafio de nosso tempo, em vez de uma abordagem conduzida por aqueles cujos interesses arraigados são incompatíveis com a redução dos gases de efeito estufa”, afirmou a Coalizão de Alta Ambição em um comunicado, ao qual a Reuters teve acesso, divulgado nesta sexta-feira (8).
O comunicado é assinado por mais de 30 países, entre os quais da Europa, América Latina e pequenos países insulares.
As revisões do IPCC, publicadas a cada 5 a 7 anos, são redigidas por especialistas e aprovadas pelos Estados-membros. Índia, Quênia, Arábia Saudita, Rússia e China pedem mais tempo para garantir a participação de cientistas com menos recursos, que enfrentam dificuldades para publicar estudos ou deixar seus empregos.
“Se o processo avançar muito rapidamente […], ele não será inclusivo”, disse Joyce Kimutai, cientista climática que representa o Quênia no IPCC.
O impasse inédito ocorre sob pressão de governos críticos a políticas climáticas que, segundo eles, prejudicam a competitividade e restringem o uso de combustíveis baratos por economias emergentes. Os Estados Unidos, maior financiador do IPCC, deixaram a organização em janeiro.
O IPCC publicará três relatórios fundamentais: um sobre a dimensão e as causas das mudanças climáticas, outro sobre vulnerabilidades e como os países podem se adaptar, e um terceiro sobre as opções para impedir o agravamento do aquecimento. Os cientistas que presidem os grupos responsáveis pela elaboração dos relatórios propuseram um plano de trabalho segundo o qual os três seriam aprovados até setembro de 2028.
Robert Vautard, copresidente francês, disse que a proposta se baseou na carga de trabalho prevista, e não na data da cúpula da ONU de 2028. Os cronogramas dos relatórios costumam ser aprovados em “questão de minutos”, acrescentou.
Os opositores da proposta dos presidentes querem adiar o segundo e o terceiro relatórios para 2029, argumentando que o cronograma é curto demais para analisar milhares de estudos, chegar a um acordo sobre a redação e concluir as revisões por especialistas e governos.
Os participantes trocaram acusações de politização do trabalho do IPCC. Alguns dos que defendem a publicação antes da cúpula de 2028 observam que grandes nações produtoras de petróleo estão entre as defensoras mais enfáticas de um prazo maior e as acusam de tentar evitar pressões para reduzir mais rapidamente as emissões de CO2.
O Ministério da Economia da Rússia diz que apresentar o debate como um confronto com produtores de petróleo era uma “deturpação deliberada” e que sincronizar o ciclo de avaliação com as negociações da ONU “levanta sérias dúvidas sobre a preservação da natureza independente do IPCC”.
Thiagarajan Jayaraman, cientista que representa a Índia no IPCC, afirmou ser “míope” sugerir que o valor do trabalho do órgão depende exclusivamente de sua contribuição para as negociações da ONU em 2028. Representantes da Arábia Saudita e da China apresentaram argumentos semelhantes em reuniões anteriores. Seus governos não responderam às perguntas para esta reportagem.
Equidade e inclusão
Países em desenvolvimento argumentam que suas perspectivas não são refletidas de forma suficiente nos relatórios do IPCC porque não há especialistas desses países em número suficiente e porque seus pesquisadores dispõem de menos recursos para publicar estudos revisados por pares.
Thiagarajan questiona, por exemplo, o que considera uma falta de consideração pelas contribuições históricas dos países ricos para o aquecimento global e pelas necessidades de desenvolvimento das nações mais pobres nos modelos usados pelo IPCC para avaliar quanto as temperaturas podem subir sob diferentes cenários de emissões.
O IPCC afirma estar tomando medidas para tornar as avaliações mais inclusivas.
Os próximos relatórios incluirão, pela primeira vez, capítulos sobre financiamento climático e formas de lidar com danos irreversíveis infligidos a alguns países, disse em junho o presidente do IPCC, Jim Skea.
Os países que defendem esse cronograma no IPCC afirmam que, de outra forma, os negociadores teriam de se basear em relatórios publicados em 2021 e 2022. Muita coisa mudou desde então.
Controlled fires burnt at Mount Solitary in the Blue Mountains range in Australia in 2018.Credit: lovleah/Getty
The suppression of Indigenous burning practices in Australia after British colonization probably transformed swathes of the southeast into dense, flammable forests that contribute to bigger, hotter bush fires today, according to a reconstruction of fires in the region over the past 1,000 years published in Nature1.
The researchers suggest that restoring Indigenous approaches to forest management could lower the risk of devastating fires, particularly as the climate warms. “It gives us another variable when we’re trying to tackle this problem, rather than just focusing on climate change,” says Michael-Shawn Fletcher, a Wiradjuri scholar and palaeoecologist at the University of Melbourne, Australia.
In 2022, a team that Fletcher was part of used pollen records to show that forests in the region tended to be more open and grassy before British colonization at the end of the eighteenth century2. Soon after the colonization, forests changed, says Fletcher. “You can see right across the landscape there was an increase in woody fuels following the British invasion,” he adds.
A millennium of wildfire
To investigate this pattern in more detail, Fletcher and his colleagues collected a sediment core from a lake at the southeastern tip of Australia in Croajingolong National Park, part of the traditional lands of the Gunaikurnai people. Reconstructions of the landscape as well as historical evidence show that they regularly burnt the land until their violent removal by the British in the mid-nineteenth century.
Like in the 2022 work, the researchers analysed pollen in the extracted sediment to reconstruct the make-up of plant species and the structure of the surrounding forest over the past 1,000 years. This record showed an open woodland, with grasses and few trees for 800 years. Then, around 1840, the researchers found that the coverage of eucalyptus (Eucalyptus spp.) had doubled and woody shrubs had spread across the area.
The team also analysed charcoal in the sediment to estimate the temperature of fires over the study time frame. They identified molecules produced by burning plants at different temperatures in the laboratory, then searched for those molecules in the sediment samples. They found an increasing fraction of high-temperature charcoals from the mid-nineteenth century onwards, suggesting more-intense fires.
The researchers say that the changing vegetation and fire regime are linked: hotter fires spur the growth of more-flammable vegetation, more fuel and a denser canopy, creating what they call a “catastrophic wildfire loop”. Eucalyptus trees rapidly sprout after a blaze clears out other competing plants, says Fletcher.
The authors attribute this change to the end of regular, small fires ignited by Indigenous people to protect certain areas from burning or to promote vegetation growth in other areas for animals to eat.
Wildfire loop
The idea of a feedback loop is “compelling”, says Christy Briles, a palaeoecologist at the University of Colorado Denver. However, she says that more sites need to be studied to understand how common the pattern is across southeastern Australia.
Others recognize this dynamic beyond Australia. “This is very convincing in my opinion,” says Clay Trauernicht, an ecologist at the University of Hawaii at Manoa. He adds that the change in wildfires seen at the lake is similar to that in other regions where Indigenous management was replaced with management oriented towards private property, such as Hawaii and the western United States.
There have been some efforts to restore cultural burning and other Indigenous management practices in southeastern Australia. But Fletcher says they have been too small and infrequent to have much effect.
As temperatures continue to rise and the planet warms, more frequent, low-temperature burns might reduce the chance of later, out-of-control fires, says Fletcher. “We need landscape-scale experiments to test these things before it’s too late,” he says.
Stephanie Arnett/MIT Technology Review | Getty Images
Você precisa fazer login para comentar.