22,014 dengue cases in northern Peru that climate change put there

A small team of US, Peruvian, and Ecuadorian researchers published a paper in One Earth on 20 March 2026 that does something the climate-and-disease literature has been trying to do for a decade. They took Peru's worst dengue outbreak on record — the 2023 epidemic that killed more than 380 people in six months — isolated a specific weather event inside it, and calculated, district by district, how many of those cases would not have existed in a world without the rainfall that climate change made more likely. The number they arrived at is 22,014.

That is roughly 60% of the dengue cases recorded in the most-affected districts of northwestern Peru during the three-month window the study covers. The senior author, Erin Mordecai of Stanford's Center for Innovation in Global Health, put the methodological point plainly: this is "the first time scientists have been able to pinpoint the role of climate change and precisely measure the impact of a particular storm on dengue." The lead author, Mallory Harris — now a postdoctoral scholar at the University of Maryland, who conducted the work during her Stanford doctorate — added a sentence worth lingering on: "The magnitude did surprise me."

For Mosticare's readers — most of whom have been told for years that "climate change drives dengue" and accept it as background knowledge — the news is not the headline. It is that the science can now write that sentence in numbers.

What Cyclone Yaku did

In March 2023, Peru's normally-arid northwestern coast — the desert strip that runs from the Ecuadorian border down through Tumbes and Piura — was hit by an unusual atmospheric configuration. A weak tropical cyclone named Yaku stalled offshore, drawing in moisture, while a coastal El Niño event warmed the eastern Pacific. The combination delivered weeks of torrential rainfall on a coast whose drainage infrastructure was built for a few centimetres a year. Streets flooded. Standing water sat. Aedes aegypti — the day-biting urban dengue vector — bred everywhere there was a container, a pothole, or a clogged storm drain.

What followed was Peru's largest dengue outbreak on record. National case totals for 2023 ran more than ten times higher than the regional historical average across the most-affected coastal departments. Hospitals in Piura and Tumbes were overwhelmed for months. The death toll for the year topped 380, and dengue was Peru's most-discussed public-health story for two full quarters.

The question Harris and Mordecai's group set out to answer was not whether the rainfall mattered — every Peruvian epidemiologist on the case knew it did. The question was how much.

The method that produced the number

Attribution of a single disease outbreak to a single weather event is harder than it sounds. The temptation is to compare cases in flooded districts against cases in dry ones and call the difference the attributable share. The trap is that flooded districts are also typically the hotter, lower-elevation, more urbanised, and more Aedes-friendly districts in the first place. Any naive comparison will overstate the rainfall signal.

The team used a technique called generalised synthetic control, borrowed from causal-inference econometrics. The method constructs a synthetic counterfactual for each affected district by weighting and combining historically similar but unaffected districts elsewhere in Peru and Ecuador — districts that resemble Piura and Tumbes in their baseline ecology, urbanisation, and dengue history but were not hit by Cyclone Yaku's rainfall. The difference between the observed case count in the real, flooded district and the predicted case count in its synthetic, unflooded twin is the attribution. Run across every district and aggregated, the difference came out to 22,014 cases.

The team then pushed the analysis one step further. Using climate-model output, they asked how much more likely the rainfall conditions that triggered the outbreak have become because of anthropogenic warming. Extreme precipitation events of the kind that hit northwestern Peru in March 2023 are now 31% more likely than they would be in a pre-industrial baseline. When the analysis is widened to the combined precipitation-plus-temperature configuration that the Aedes-friendly conditions of 2023 required, the probability has nearly tripled since pre-industrial times.

Each of these steps is, on its own, well-established. What is new is that the chain — climate change to weather event, weather event to vector ecology, vector ecology to human cases — has been made with the numbers showing, in a peer-reviewed journal, for a single outbreak. The methodology is more important than the headline figure, because the methodology is what makes the next attribution paper writeable.

The honest reading

There is a real risk, with a study this striking, of reading more out of it than is in it. Two cautions.

The first is that the 60% figure is the climate-amplified share of cases in the most-affected districts during a specific three-month window. It is not "60% of Peru's 2023 dengue outbreak." Districts outside the rainfall corridor saw their own elevated case counts driven by transmission patterns the synthetic-control method does not directly attribute to the cyclone. The 22,014 figure is a precise, defensible subset of a larger national epidemic, not a redefinition of it.

The second is that climate attribution does not change clinical or operational defences. The mosquitoes are still the mosquitoes. The standing water in the bottom of a discarded tyre is still the breeding site it always was. Aedes aegypti still bites by day, still favours the human-built environment, still responds to coordinated source-reduction work in exactly the same way it did before climate scientists had a paper that quantified its rainfall-elasticity. Climate change is amplifying the hazard. It is not changing what works against it.

What it does change is the framing of preparation budgets. A municipal flood-resilience programme in Piura is no longer just a flood-resilience programme; on the strength of Harris and Mordecai's numbers it is also a dengue-prevention programme, and the cost-benefit arithmetic for upgrading storm drains and clearing solid-waste accumulations should now include the avoided disease burden of the next coastal El Niño year. The study's own recommendations close on exactly this point: targeted vector control, vaccination programmes where feasible, and urban flood-resilience infrastructure as a public-health intervention.

What this method will produce next

The most editorially significant fact about the Harris paper is not the Peru number. It is that the same method can now be turned on every other major arboviral outbreak whose proximate weather event is identifiable.

The 2024 dengue cases that ran double the 2023 global total — five million cases globally in 2023, more than doubled the following year — are full of candidates for the same kind of analysis. The 2024 Brazil dengue surge that drove Sabin and Butantan to accelerate vaccine rollouts has a clear El Niño signature. The 2024–2025 Indian Ocean chikungunya resurgence, anchored on Réunion and now spilling to Mauritius and Mayotte, has a cyclone-season signature worth interrogating. Each is, in principle, a generalised-synthetic-control attribution paper waiting to be written.

The next few years of climate-and-disease literature will, if Harris and Mordecai's methodology holds, look quite different from the previous few. The line researchers and journalists have been forced to use — "climate change is making this kind of outbreak more likely" — will be progressively replaced by a more usefully specific sentence: "this outbreak produced this many additional cases because the climate did this specific thing."

That is a better sentence for public-health budgets, for insurance models, for municipal infrastructure investment, and for the editorial register of a serious mosquito publication. Mosticare will be tracking the next papers in this series carefully.

What we know

• 60% of dengue cases in the most-affected northwestern Peruvian districts during a roughly three-month window in 2023 were attributed to climate-amplified rainfall from Cyclone Yaku and a coastal El Niño (Harris et al., One Earth, 20 March 2026).
• 22,014 attributable cases is the headline figure, derived using generalised synthetic control against unaffected comparator districts.
• Extreme precipitation of the kind seen in March 2023 is now 31% more likely than in a pre-industrial baseline; the combined precipitation-plus-temperature configuration has nearly tripled in probability (Harris et al., 2026).
• Peru's 2023 dengue outbreak killed more than 380 people in six months and produced case totals more than 10× the historical regional average (Inside Climate News, 20 March 2026).
• The study is described by its senior author as the first to precisely quantify climate change's contribution to a specific dengue outbreak via a specific storm (Mordecai, in Stanford Woods Institute coverage, March 2026).

Sources cited

1. Harris M, Mordecai E et al., Extreme precipitation, exacerbated by anthropogenic climate change, drove Peru's record-breaking 2023 dengue outbreak, One Earth (20 March 2026) — https://www.cell.com/one-earth/fulltext/S2590-3322(26)00020-5
2. Inside Climate News, A New Study Links a Record-Breaking Tropical Disease Outbreak in Peru to Climate-Driven Extreme Weather, 20 March 2026 — https://insideclimatenews.org/news/20032026/peru-dengue-fever-record-outbreak-climate-driven-weather/
3. Phys.org, Warmer, wetter cyclone weather drove Peru dengue outbreak, March 2026 — https://phys.org/news/2026-03-warmer-wetter-cyclone-weather-peru.html
4. Stanford Global Health, New publication shows climate change exacerbated Peru's 2023 dengue outbreak — https://globalhealth.stanford.edu/news/new-publication-shows-climate-change-exacerbated-perus-2023-dengue-outbreak.html/
5. Stanford Woods Institute for the Environment, Climate change is fueling disease outbreaks — https://woods.stanford.edu/news/climate-change-fueling-disease-outbreaks