This preprint applied random forest models to 652,897 Arizona mosquito trap records from 2013 to 2023. Predictors included weather averaged over preceding weeks. Averaged across 44 models, it predicted whether trapped Culex pools tested positive for West Nile virus with 98.9% accuracy. Modelled virus activity explained 65.3% of weekly variation in county-level human cases, which followed one week later. Projections to 2100 showed significant rises in weeks of virus activity over 0.6% to 39.4% of Arizona, by scenario. The projections include only direct climatic effects, leaving out urbanisation, water management and mosquito adaptation.
Why it is interesting: Mosquito abundance and virus occurrence did not fall at Arizona's extreme summer heat, suggesting thermal refugia. Drought followed by monsoon rain, rather than temperature alone, was the main driver of activity.
Cite this item
BibTeX
Free sign-up. A digest of the new papers twice a week, grouped by domain.
Other drivers in Tomorrow's world
-
Innovation in treatment
Treatments that did not exist before — drugs, operations, devices — and whether they reach the people who would benefit.
None this week Earlier → -
Changing behaviour at scale
Whether new technology can change what people do at the scale of a population: apps, wearables, automated messages, digital programmes and platforms, and how far their effects last and reach.
None this week Earlier → -
Medicine at a distance
What wearables, remote monitoring and care at home add to how health is measured and delivered, and whether it shows in outcomes.
1 this week -
A changing climate
How air pollution, heat and cold affect deaths and survival, how the attribution is made, and how the balance is changing.
4 this week -
Slowing ageing itself
Whether anything yet slows ageing itself, rather than treating one disease at a time, and how we would know if it did.
2 this week
