Forests speak in chemistry
Walk through a forest and you may notice a resinous or citrus-like scent. Some of that aroma comes from monoterpenes, a diverse family of volatile molecules made by plants. These compounds can help plants interact with their environment, but once they enter the air they also react with oxidants and contribute to the chemistry that shapes ozone and airborne particles.
Monoterpenes are often grouped by molecular shape. Cyclic compounds, such as α-pinene, contain carbon rings. Acyclic compounds, including β-ocimene, myrcene and linalool, have open-chain structures. That distinction may sound technical, yet it matters: the two groups can follow different biosynthetic pathways, serve different physiological roles and react at different speeds in the atmosphere. Despite this, widely used emission models generally give them similar temperature responses. We wanted to know whether a real heatwave would prove that assumption wrong.
From a warmed leaf to the air above a forest
We worked at the Heshan National Field Research Station of Forest Ecosystems in Guangdong, South China. In controlled leaf experiments, we warmed mature leaves from typical conditions to 35 and 45 °C, while keeping light, humidity and carbon dioxide controlled. We collected the gases leaving each leaf and identified individual compounds by gas chromatography-mass spectrometry. The experiments covered eight broadleaf species and one pine species that are common or relevant to tropical and subtropical forests in the region.
A leaf chamber, however, cannot reproduce a whole forest. We therefore added two further lines of evidence: repeated standardized leaf measurements during the summer of 2024, and air samples collected from a tower before, during and after a five-day heatwave. Finally, we asked whether a commonly used emission model could reproduce the pattern and compared our observations with previously published measurements from an Amazonian forest.
Figure 1. Figures show the forest canopy, leaf sampling, tower observatory at the Heshan National Field Research Station of Forest Ecosystems in Guangdong, South China. Credit: Dr. Jianqiang Zeng
The heat changed the recipe, not only the amount
Between 25 and 35 °C, monoterpene emissions from the measured leaves rose by about three- to seven-fold. Yet the increase was not evenly shared. The strongest temperature sensitivities were concentrated among acyclic compounds. At the same time, net carbon dioxide assimilation fell by about 13%, showing that the emission increase was not simply a reflection of faster photosynthesis.
The forest air told the same broad story. Before the August 2024 heatwave, acyclic monoterpenes made up about 12% of the measured mixture. During the heatwave they reached 47%, while the midday total monoterpene mixing ratio peaked at 2.3 parts per billion. Published Amazon data showed a remarkably similar shift during an El Niño-related heat event, from 8% to 48% acyclic compounds.
Ambient air is influenced by emissions, chemical reactions and transport, so it is not a direct readout of what every leaf emitted. In fact, many acyclic monoterpenes are shorter-lived than cyclic ones. Their strong enrichment in hot forest air is therefore consistent with, rather than definitive proof of, an underlying shift in emissions.
Figure 2. Heatwave observations reveal enrichment of acyclic monoterpenes in South China and in previously published Amazon data. The default temperature scheme missed the shift, whereas the revised class-specific scheme captured its temperature dependence. Source: Zeng et al., Nature Communications (2026).
Why the atmosphere notices
Chemical composition matters because molecules do not all react at the same speed. During the Heshan heatwave, ambient acyclic monoterpenes increased 7.9-fold relative to pre-heatwave conditions. The calculated reactivity of all measured monoterpenes with ozone increased 6.1-fold, and their calculated reactivity with the hydroxyl radical - often described as an atmospheric cleaning agent - increased 5.9-fold. Acyclic compounds accounted for about 80% of the monoterpene-ozone reactivity and 71% of the monoterpene-OH reactivity during the heatwave.
These numbers describe how rapidly the measured mixture could consume oxidants under assumed conditions. They are not direct measurements of ozone production, aerosol formation or cloud effects. Those outcomes depend on many other ingredients, including nitrogen oxides and the products formed as each compound breaks down. Our result is more focused: a heat-driven change in the chemical recipe can alter atmospheric reactivity even when total monoterpene abundance alone would not reveal the full story.
Figure 2. During the heatwave, the measured mixture became richer in acyclic monoterpenes and its calculated ozone and OH reactivity rose sharply. These are reactivity estimates, not direct measurements of ozone formation. Source: Zeng et al., Nature Communications (2026).
A model needs the ingredients, not just the total
The default model treated cyclic and acyclic monoterpenes as if heat affected them in nearly the same way, so it predicted little change in composition. When we introduced separate temperature-response functions for the two groups, the model reproduced the observed rise in the acyclic fraction. Matching the absolute composition still required adjusting the baseline amount of acyclic emissions, which tells us that temperature sensitivity is better constrained than the starting mixture.
This distinction is important for future modeling. A forest can release the same broad family of compounds yet produce a chemically different atmosphere when the proportions change. Class-resolved temperature responses offer a practical next step toward representing that behavior during extreme heat.
What we still need to learn
Why might plants favor acyclic monoterpenes under heat? Many of these compounds are associated with stress responses and are chemically capable of reacting with reactive oxygen species. Our observations are consistent with a possible protective role, but they do not demonstrate antioxidant action or improved heat tolerance. Other plants store terpenes in specialized structures and may respond with sudden bursts or lingering emissions instead.
Our revised functions are also descriptive summaries of one South China dataset, not universal tropical parameters. The agreement with the Amazon observations is encouraging, but broader tests are needed across forest types, species, canopy positions and disturbance histories. We also need chamber experiments and chemistry-transport models to determine how the compositional shift affects ozone, particles and clouds.
The clearest lesson is already visible. During a heatwave, a tropical forest does not simply speak louder to the atmosphere. It changes its chemical vocabulary. Recognizing that change may help us understand and model the increasingly complex conversation between forests, air chemistry and climate extremes.
Reference
Zeng, J., Wang, X., Lu, Y., Pang, W., Song, W., Zhang, Y., Peng, P. a. & Wang, X. Heatwaves favour acyclic monoterpene emissions in tropical forests. Nat Commun (2026). https://doi.org/10.1038/s41467-026-77087-x