We cannot measure O3 exposure and O3-damage in every tropical forest, or in the forests of 2100. Models let us scale up from our experiments and field measurements to whole continents and future climates, and ask what O3 means for the global carbon cycle. Our modelling links three pieces together.
1. How much O3 forests are exposed to
We use platforms such as the UK Earth System Model (UKESM1), which simulates atmospheric chemistry alongside climate, to estimate O3 concentrations across the tropics today and under future emission and climate scenarios. The model includes O3-forming emissions from vehicles and industry, from fires (based on satellite-observed burned area), and from the forest itself, whose emissions of isoprene and other volatile organic compounds (VOCs) are simulated interactively. It produces hourly O3 concentrations in the lowest layer of the atmosphere, roughly 20 m above the ground.
When tested against measurements at Amazon and other tropical sites, UKESM1 captures how O3 varies from hour to hour and season to season, but tends to overestimate absolute concentrations (Brown et al. 2024). Where this matters, we analyse changes relative to the average rather than absolute values.
2. How much damage that O3 does
The hourly O3 fields from atmospheric chemistry models are used alongside climate simulations to drive land-surface models like JULES (Joint UK Land Environment Simulator). JULES simulates photosynthesis and how open the stomata are from moment to moment, responding to light, temperature, humidity, soil moisture and CO2. This lets us calculate the O3 flux into leaves: the dose that actually matters for damage, rather than the concentration in the air.
Above a threshold flux (typically 1 nmol O3 per m² of leaf per second), each additional unit of O3 taken up reduces photosynthesis, and stomatal conductance falls in step. The strength of that reduction, the O3 susceptibility, comes directly from our chamber experiments: we often use the average dose–response of the tropical tree species we have tested, and repeat the simulations using the most and least sensitive responses to bracket the uncertainty.
3. What it means for the carbon cycle
Small reductions in photosynthesis, repeated every day across millions of square kilometres of forest, add up. By running models with and without O3 damage, we calculate how much carbon forests fail to take up each year because of O3 pollution, and how this changes with fire, drought and rising CO2.
What we have found
Ozone is already reducing tropical forest growth. Combining O3 susceptibility measured in tropical tree species with global modelling, we estimate that ground-level O3 reduces new growth in tropical forests by 5.1% a year on average, and by 10.9% in tropical Asia. Since 2000, this has left around 290 million tonnes of carbon a year uncaptured, a cumulative reduction of about 17% in carbon removal by tropical forests this century. Cheesman et al. 2024, Nature Geoscience
Fire-driven O3 adds a hidden carbon cost. Fires release the nitrogen oxides that drive O3 formation, and across the Amazon fire activity is the strongest driver of year-to-year swings in O3. Simulating the Amazon from 1997 to 2024, we found that O3 damage from fire-derived pollution reduces forest carbon uptake by an amount equal to roughly a quarter of the carbon released directly by the fires themselves (24%, with a range of 18–35% depending on species susceptibility). This indirect pathway is missing from most Amazon carbon budgets to date.
Drought can shield forests, but not always. During drought, trees close their stomata to save water, which also keeps O3 out. Rising CO2 and a warming, drier atmosphere have similarly reduced stomatal opening over recent decades, lowering O3 damage. But this protection has limits. In the southern Amazon, and especially along the arc of deforestation, drought-year fires drive O3 uptake up during the dry season. In 2024, record fire activity during an extreme drought produced substantial O3 damage despite widespread stomatal closure. Brown et al. 2026, Communications Earth & Environment
Climate change will add to the problem
A warmer climate favours O3 formation, a so-called climate penalty. Using fully coupled climate model projections, we found that by 2100 climate change alone could raise O3 over South America and Africa by up to 15 ppb in regions with high biomass burning. Brown et al. 2022, Atmospheric Chemistry and Physics.
Testing the models
Model predictions are only as good as the data used to test them, and in the tropics those data are scarce. We check simulated O3 against surface measurements and satellite retrievals, check simulated pollution from fires against satellite observations of nitrogen dioxide, and check simulated forest productivity against satellite-derived estimates. We have also shown that global reanalysis products cannot reproduce the seasonal and daily cycles of O3 measured in the Congo Basin (Vieira et al. 2023).
But there are almost no long-term O3 records from inside tropical forests. This is why we are building capacity for O3 monitoring at tropical forest research sites: multi-year measurements that provide the ground truth needed to test and improve the models.