Chemical Conversations Between Forests and the Sky Affect Air Quality, Clouds, and More

Do forests cool the planet or warm it? It’s a deceptively simple question with no clear answer. The catch is that the same forest can push climate in opposite directions at once. Forests are widely viewed as climate change mitigators—and planting trees as a partial climate solution—because they absorb carbon dioxide. Yet that framing is incomplete. Trees also release reactive gases called biogenic volatile organic compounds (BVOCs) that can influence cloud formation, contribute to air pollution, and extend the atmospheric lifetime of methane, a greenhouse gas that can trap about 80 times as much heat as carbon dioxide. Although largely invisible to the public, BVOC emissions are significant. They account for roughly 90% of the global mass of nonmethane volatile organic compounds entering the atmosphere [Khan et al., 2025]. In natural habitats, forest trees are the dominant sources of BVOCs. They’re responsible for the familiar scent of pine stands and the bluish haze often seen above dense tropical canopies. Agricultural crops and wetlands are major sources in rural regions. In cities, BVOC emissions come primarily from parks, street trees, and remnant woodlands. Accounting for the competing effects of BVOCs has become increasingly important for climate scientists, especially as forest habitats and footprints shift and land use patterns continue to change worldwide [Wang et al., 2024]. Better understanding whether, where, and to what extent these plant emissions affect the atmosphere will allow researchers to improve forecasts of air quality, cloud formation, and future warming. From Leaves to Clouds Whether forests cool or warm the planet depends on how BVOCs behave once they leave the foliage. Cloud-shaping plant chemistry begins with photosynthesis in sunlit leaves. Cloud-shaping plant chemistry begins with photosynthesis in sunlit leaves. From there, the normal steps of plant metabolism synthesize and release BVOCs. Globally, isoprene makes up the largest share of BVOC emissions and is especially emitted from many broad-leaved trees. Additional volatiles, such as monoterpenes and sesquiterpenes, are released in smaller amounts, often by conifers and stressed vegetation. Once airborne, these invisible vapors oxidize rapidly and condense into microscopic specks known as secondary organic aerosols (SOAs). As SOAs drift and collide, they can grow into cloud condensation nuclei, the tiny seeds around which water droplets form. Particles larger than about 1 micrometer, about 1/100 the diameter of human hair, can influence cloud brightness and longevity, as well as Earth’s energy balance. The result is a chain of interactions that begins at the level of individual leaves and extends to regional and global climate systems. These interactions are dynamic, not fixed. Rising temperatures generally amplify BVOC emissions. A 2°C–3°C rise in mean global temperature is expected to increase total BVOC emissions by 30%–45% worldwide. High-latitude, nutrient-limited ecosystems are especially sensitive. Researchers have found, for example, that isoprene emissions from Arctic sedges are exceptionally temperature sensitive, suggesting that warming could produce disproportionately large increases in emissions from these ecosystems. But if temperatures rise beyond the range that plants can tolerate or warming is paired with prolonged drought, then photosynthesis can falter, and BVOC emissions can be suppressed. The system thus comprises a powerful and delicate feedback, in which plants respond to climate while simultaneously helping to regulate it through their effects on the clouds above. A Climate Paradox The paradox is that identical molecules released by living plants can either cool or warm Earth, depending entirely on the conditions in the atmosphere when and where they enter it (Figure 1). Fig. 1. The same compounds emitted by forests can either cool or warm the atmosphere depending on background pollution levels and atmospheric chemistry. BVOCs, biogenic volatile organic compounds; NOx, nitrogen oxides. Credit: Mary Heinrichs/AGU In cleaner air, the balance tilts toward a cooling effect of BVOCs. As SOAs seed clouds, droplet numbers increase, and clouds grow brighter. They reflect more of the Sun’s energy back into space. The longer those droplets persist, the more they extend a thin planetary shade over the landscape below. Recent model estimates suggest that BVOC-driven aerosol effects may have a radiative cooling effect between −2.0 and −0.4 watts per square meter [Bellouin et al., 2020]. This value is smaller than the roughly 2–3 watts per square meter of warming from carbon dioxide currently. Nevertheless, the effect is large enough to shape regional weather and cloud feedbacks. Under polluted skies, however, the same emissions can have the opposite effect. In the presence of nitrogen oxides from tailpipe exhaust, power plant emissions, and other combustion sources, for example, BVOCs help generate ozone, a greenhouse gas found in urban smog. They can also disrupt the chemistry that normally removes methane from the atmosphere, allowing it to remain aloft longer and trap more heat. These warming pathways partially offset the cooling produced by biogenic aerosols, although current evidence indicates that the global cooling effect remains larger overall, with substantial regional variation and uncertainty among warming effects. Temperature, sunlight, drought, and pollution thus determine whether forests act as air conditioners or furnaces with respect to warming. The Air People Breathe The consequences of plant vapors are not confined to clouds and climate. Biogenic volatile organic compounds (BVOCs) can also contribute to pollution. The consequences of plant vapors are not confined to clouds and climate. BVOCs themselves can also contribute to pollution that exacerbates respiratory health conditions and leads to increased hospital emergency room visits and premature deaths. Photochemical reactions in the air can intensify quickly above traffic corridors and rooftops, especially under summer sunlight. The ground-level ozone and SOAs (an important component of fine particulate matter) formed through reactions of BVOCs with nitrogen oxides are both harmful forms of air pollution. Ozone inflames airways in our lungs, worsens asthma, and is linked to higher mortality. Fine particles small enough to penetrate deep into the lungs are associated with heart disease, stroke, and premature death. Haze hangs over Los Angeles, as seen from the Hollywood Hills. Credit: David Iliff/Wikimedia Commons, CC BY-SA 3.0 The risk is greatest where dense populations and human-caused emissions overlap with strong biogenic emissions, a scenario that characterizes places like Los Angeles, Atlanta, and Beijing. Even with stricter vehicle tailpipe standards and other emissions controls, ozone pollution persists in many cities around the world as heat, sunlight, and mixed emissions from anthropogenic and natural sources continue to interact. A Gray Area of Greening Cities Urban vegetation is one of the most visible tools that cities have for confronting another public health threat: extreme heat. Trees reduce heat exposure by shade- and evapotranspiration-driven cooling. They can also help filter pollutants from the air and are associated with benefits for mental health and well-being. But not all trees are chemically equivalent. A recent meta-analysis covering 357 plant species found that broad-leaved trees such as oaks produce markedly different blends of BVOCs than pines and other conifers [Bao et al., 2023]. For example, 78% of broad-leaved trees emit isoprene, compared with only 48% of conifers, whereas 93% of conifers emit monoterpenes. These emissions are also highly sensitive to climate. Another meta-analysis found that warming increased isoprene emissions globally by an average of 107% and monoterpene emissions by 60%, whereas drought reduced them by 27% and 33%, respectively [Chen et al., 2026]. Thus, the trees planted to cool tomorrow’s cities may behave differently under tomorrow’s climate conditions than they do today. Reducing nitrogen oxide pollution decreases ground-level ozone formation, yet expanding urban tree cover may increase emissions of reactive BVOCs that offset these decreases. This uncertainty creates a challenge for policymakers. Reducing nitrogen oxide pollution decreases ground-level ozone formation, yet expanding urban tree cover may increase emissions of reactive BVOCs that offset these decreases depending on species choice and local conditions. A pioneering 2024 survey by the California Department of Transportation evaluated more than 260 commonly used roadside plants for their potential to contribute to ozone and fine particulate formation. The report identified willows (Salix spp.), American sweet gum (Liquidambar styraciflua), sycamores (Platanus spp.), gum trees (Eucalyptus spp.), and several oak species (Quercus spp.) among the highest BVOC emitters. In contrast, all pines (Pinus spp.) and other conifers were categorized as comparatively low emitters. The survey’s findings underscore that urban greening strategies are not chemically neutral. Species selection can shape local atmospheric composition and influence whether vegetation improves or inadvertently worsens air quality. Questions of urban forestry, once framed mainly around shade and aesthetics, are increasingly considering how trees affect atmospheric chemistry. Challenges for Atmospheric Modeling Climate modelers must address three main challenges to translate the chemistry and behavior of BVOCs into realistic simulations of the atmosphere. The first is accurately estimating the magnitudes of BVOC emissions themselves. Some analyses report uncertainties in modeled emissions estimates approaching or exceeding 100% [Wang et al., 2021]. Even for a single compound, such as isoprene, emissions can vary substantially depending on ecosystem composition. In some regions, updated land cover data reduced estimated emissions modestly, whereas in others they more than doubled them relative to estimates based on conventional land cover maps [Opacka et al., 2021]. BVOCs and their oxidation products can travel far beyond forested regions, influencing atmospheric processes well downwind of their sources. The second challenge is that BVOC oxidation in the atmosphere remains incompletely understood. A major limitation for models has been the difficulty of representing the thousands of intermediate reactions involved in this chemistry. Recent work shows that incorporating more detailed oxidation mechanisms into current models can significantly alter predictions of SOA formation and its climatic effects, narrowing some of the models’ largest uncertainties [Zha et al., 2023]. The third challenge is one of scaling. Observations from high-altitude monitoring stations have shown that air masses containing BVOCs and their oxidation products can travel far beyond forested regions, influencing atmospheric processes well downwind of their sources [Okamoto and Tanimoto, 2016]. To capture these effects, models must account for processes occurring in individual leaves and then accurately simulate the transport and chemical evolution of BVOCs throughout the atmosphere [Guenther et al., 2012]. Although these difficulties limit our ability to resolve BVOC-driven climate feedbacks, available models have clearly revised the standard view of forests as passive carbon sinks. Tracking the Chemical Pulse of Forests New tools are narrowing gaps between living ecosystems and the climate models built to represent them. For decades, scientists estimated BVOC emissions by combining vegetation maps with average emission rates for different plant types and equations describing how emissions change with temperature and sunlight. In these models, forests were represented largely as static patches on a grid, with broad vegetation categories standing in for the diversity and responsiveness of real ecosystems. That simplified picture is now giving way to a far more observational, data-rich view of Earth’s vegetation. The Model of Emissions of Gases and Aerosols from Nature (MEGAN) is one example of how that transition is occurring. MEGAN estimates BVOC emissions by combining information about terrestrial vegetation with data on environmental conditions that influence emissions. Its latest generation, MEGANv3.2, improves upon earlier versions by estimating emission factors from measurements of individual plant species rather than relying on fixed categories of plant functional types. Using MEGANv3.2 together with time-varying satellite observations of vegetation and meteorological reanalysis data, Wang et al. [2024] estimated that average annual global BVOC emissions from 2001 through 2020 amounted to about 835 million metric tons, including 348 million tons of isoprene and 185 million tons of monoterpenes. The atmosphere responds not simply to how much vegetation is present but to what grows where and how vegetation is influenced by its changing environment. The new model also illustrates why representing vegetation accurately matters. Although global isoprene emissions declined slightly during the study period, regional patterns differed sharply under varying influences from changes in vegetation, temperature, soil moisture, and atmospheric carbon dioxide [Wang et al., 2024]. In Europe, for example, higher temperatures helped increase emissions, whereas higher soil moisture was more important for increasing emissions in East and South Asia. Earlier generations of MEGAN had shown the importance of vegetation type on BVOC emissions. Guenther et al. [2012] found that tropical forests, which cover less than a fifth of Earth’s land surface, account for roughly 80% of the isoprene and terpenes emitted by vegetation and about half of other plant-emitted volatiles, whereas other forested ecosystems covering a similar area contribute only about 10% of total emissions. Such early estimates helped establish a fundamental principle that newer models are now able to examine in greater detail: The atmosphere responds not simply to how much vegetation is present but to what grows where and how vegetation is influenced by its changing environment. Sensing Biogenic Volatiles from Orbit Satellites are also driving much of the shift toward better data and improved understanding. Instruments such as the Tropospheric Monitoring Instrument (TROPOMI) aboard the European Space Agency’s Sentinel-5P satellite can detect atmospheric formaldehyde, a compound formed when isoprene oxidizes in air. Because isoprene is the dominant BVOC emitted by many forests, formaldehyde serves as a large-scale chemical fingerprint of plant emissions. From hundreds of kilometers overhead, satellite instruments like TROPOMI and the Cross-track Infrared Sounder now allow a comparison of modeled emissions with real atmospheric chemistry across continents rather than at only a scattering of field sites. Measurements of atmospheric formaldehyde reveal regions where vegetation releases large quantities of reactive gases. Because formaldehyde forms during the oxidation of plant-emitted compounds such as isoprene, these observations provide a global view of the chemical exchanges linking forests, clouds, air quality, and climate. These data, collected by the Sentinel-5P satellite, reflect averaged formaldehyde concentrations across the globe from 1 March to 31 May 2026. Credit: Sentinel-5P Product Algorithm Laboratory; contains modified Copernicus Sentinel-5P/TROPOMI data Satellite observations have already revealed that even widely used emission inventories can contain substantial regional biases. For example, a prior version of MEGAN was found to overestimate total African isoprene emissions by 22%—and emissions in the central African rainforest by 43%—while underestimating emissions from southern deciduous forests by 21% [Marais et al., 2012]. Satellite observations have also shown that some mismatches between observations and models arise not from incorrect emissions estimates, but from incomplete understanding of atmospheric oxidation chemistry, particularly that of isoprene [Wells et al., 2020]. Other satellites observe plants more directly. Missions such as NASA’s Orbiting Carbon Observatories (OCO-2 and OCO-3) measure solar-induced fluorescence emitted by chlorophyll during photosynthesis. Because the amount and composition of BVOC emissions can change with plant productivity and metabolic activity, these data help to determine when ecosystems are active and how emissions respond to drought, heat waves, and seasonal change. Continuing to Reveal Ecosystem Influences The behaviors of clouds, aerosols, and tropical ecosystems still pose major challenges for atmospheric modeling, as does resolving the detailed chemistry of BVOCs. Thousands of reactions happen in air and aerosol droplets, and no satellite or global model can track them all in full. Yet the atmosphere above forests is no longer being inferred and approximated from temperature curves or simplified assumptions. Innovative models and observing technologies are linking forest emissions to the beat of photosynthesis itself and revealing how ecosystems continuously influence climate, clouds, and air quality through dynamic chemical exchanges once treated as background noise. References Bao, X., et al. (2023), A meta-analysis on plant volatile organic compound emissions of different plant species and responses to environmental stress, Environ. Pollut., 318, 120886, https://doi.org/10.1016/j.envpol.2022.120886. Bellouin, N., et al. (2020), Bounding global aerosol radiative forcing of climate change, Rev. Geophys., 58(1), e2019RG000660, https://doi.org/10.1029/2019RG000660. Chen, J., et al. (2026), Global variability in isoprenoid emissions: Divergent responses to warming and drought driven by physiochemical controls, Plant Cell Environ., 49(8), 5,294–5,312, https://doi.org/10.1111/pce.70536. Guenther, A. B., et al. (2012), The Model of Emissions of Gases and Aerosols from Nature version 2.1 (MEGAN2.1): An extended and updated framework for modeling biogenic emissions, Geosci. Model Dev., 5(6), 1,471–1,492, https://doi.org/10.5194/gmd-5-1471-2012. Khan, M. A. H., et al. (2025), Isoprene emissions, oxidation chemistry and environmental impacts, Atmosphere, 16(3), 259, https://doi.org/10.3390/atmos16030259. Marais, E. A., et al. (2012), Isoprene emissions in Africa inferred from OMI observations of formaldehyde columns, Atmos. Chem. Phys., 12(14), 6,219–6,235, https://doi.org/10.5194/acp-12-6219-2012. Okamoto, S., and H. Tanimoto (2016), A review of atmospheric chemistry observations at mountain sites, Prog. Earth Planet. Sci., 3, 34, https://doi.org/10.1186/s40645-016-0109-2. Opacka, B., et al. (2021), Global and regional impacts of land cover changes on isoprene emissions derived from spaceborne data and the MEGAN model, Atmos. Chem. Phys., 21(11), 8,413–8,436, https://doi.org/10.5194/acp-21-8413-2021. Wang, H., et al. (2021), A long-term estimation of biogenic volatile organic compound (BVOC) emission in China from 2001–2016: The roles of land cover change and climate variability, Atmos. Chem. Phys., 21, 4,825–4,848, https://doi.org/10.5194/acp-21-4825-2021. Wang, H., et al. (2024), Regional to global distributions, trends, and drivers of biogenic volatile organic compound emission from 2001 to 2020, Atmos. Chem. Phys., 24(5), 3,309–3,328, https://doi.org/10.5194/acp-24-3309-2024. Wells, K. C., et al. (2020), Satellite isoprene retrievals constrain emissions and atmospheric oxidation, Nature, 585(7824), 225–233, https://doi.org/10.1038/s41586-020-2664-3. Zha, Q., et al. (2023), Oxidized organic molecules in the tropical free troposphere over Amazonia, Natl. Sci. Rev., 11(1), nwad138, https://doi.org/10.1093/nsr/nwad138. Author Information Dennis Clark ([email protected]), Arizona State University (retired), Tempe Citation: Clark, D. (2026), Chemical conversations between forests and the sky affect air quality, clouds, and more, Eos, 107, https://doi.org/10.1029/2026EO260295. Published on 18 September 2026. Text © 2026. The authors. CC BY-NC-ND 3.0Except where otherwise noted, images are subject to copyright. Any reuse without express permission from the copyright owner is prohibited.