Models Put 2023-Level Tropical Heat Every Year From 2043

Three quarters of the world's tropical forests endured extreme heat and drought at the same time for an average of six months during the 2023-24 El Nino, according to a study published in the American Geophysical Union journal Earth's Future on 7 October 2026. Projections in the same paper put conditions like those at annual frequency across most tropical regions by 2043.

Johanna Van Passel and colleagues at Ghent University's Q-ForestLab, with co-authors at Wageningen University, the University of Campinas and the Laboratoire des sciences du climat et de l'environnement, analysed 12 observational datasets of heat and drought covering the past 25 years. The study area ran to nearly 16 million square kilometres of tropical forest. The team searched for months-long spells roughly half a degree Celsius hotter than a normal month.

Every tropical forest region ran hot, and stayed hot for roughly 18 months on average. Eighty per cent were drier than normal, taking 22 to 56 millimetres less rain a month than usual across multiple months. Parts of the Amazon had two full years of anomalous drought. At the port of Manaus in Brazil, water levels fell to their lowest since record-keeping began more than a century ago.

What 24 models say about 2043

For the projections the team drew on 24 Earth system models across a range of emissions scenarios, then asked how likely the 2023 and 2024 conditions would be in any given year, El Nino or not. If the world reaches carbon neutrality by 2075, the high annual temperatures of those two years still become at least six times more likely across all tropical regions by 2100 than they were over the past 25 years. Under any less optimistic scenario they become the annual norm in most tropical regions by 2043.

Drought projections diverge more widely than temperature. Asia will probably see drought less often. For the Amazon under a high-emissions scenario, the low rainfall of 2023-24 becomes up to nine times as likely in any given year. The mechanism that worries the authors is circular: water released by trees through evapotranspiration supplies as much as 30 per cent of the Amazon's total rainfall, so when one region stops transpiring, it withholds rain from the next.

The ozone the fires leave behind

A second paper published the same day, in Communications Earth & Environment, puts a figure on an effect that had never been quantified against the Amazon's carbon budget. Flossie Brown, an atmospheric chemist at ETH Zurich, and co-authors ran a land surface model driven by hourly ozone concentrations from the UK Earth System Model to estimate how much carbon uptake is lost when nitrogen oxides from fires form ground-level ozone, which damages leaves well away from the flames.

Their central estimate is that the reduction in net annual carbon uptake from fire-derived ozone damage equals 24 per cent of the direct carbon emissions of the fires themselves. The paper reports an uncertainty range of 18 to 35 per cent, derived from observed variation in how susceptible tropical tree species are to ozone, and says the contribution should be interpreted as an order-of-magnitude estimate. Earlier studies had put fire-related ozone damage at roughly a 15 per cent cut in Amazon plant productivity without translating it into carbon emissions.

Both findings are modelled rather than measured, and the limits are stated in the papers. Brown's team notes that its simulations do not fully capture year-to-year changes in broadleaf forest extent, forest degradation or shifts in forest edges along active deforestation frontiers, which may introduce regional bias. One of its own results cuts against the headline: across the 1997 to 2024 study period ozone damage declined on the whole, because rising carbon dioxide concentrations let trees photosynthesise more efficiently and so draw less air, and less ozone, through their stomata. The authors' argument is that this gradual improvement is outweighed by what happens during the extremes.

Together the papers describe a pressure on the Amazon that forest policy has little direct purchase on. As Parallax Nexus reported, Brazilian Amazon deforestation more than halved between 2022 and 2025 on INPE's PRODES record. Clearance rates and sink strength are now moving in opposite directions, and the forest left standing is being asked to absorb carbon in a climate it did not grow up in.

The honest reading is that the number worth arguing over is 2043, not 2100. A projection for the end of the century is a claim about emissions policy, and can be discounted or deferred accordingly. A projection 17 years out is a claim about forests that are already standing, measured against a cycle that has started again: Inside Climate News reported on 7 October that a powerful El Nino is now driving drought in the region, with the agricultural burning season due to begin within weeks.