When millions of geometrid moth caterpillars stripped the birch canopy bare across subarctic Sweden’s Abisko region in 2023, they did something that scientists had long theorized but never measured across a full, naturally occurring outbreak at field scale: they rewrote the chemical signature of the air above the forest. A study published October 5, 2026 in Communications Earth & Environment confirmed for the first time — through three consecutive years of in-situ measurements on 256 trees — that moth defoliation increased the quantity of plant-defense compounds pumped into the atmosphere and fundamentally changed what those compounds were. Because those airborne chemicals are the raw ingredients for particles that seed Arctic clouds, the finding points to a biotic climate feedback loop that current Earth-system models do not include.
What Flew Out of the Birch Trees During an Outbreak
Mountain birch forests in northern Fennoscandia do not just sit there and absorb sunlight. Every tree continuously releases a chemical plume — biogenic volatile organic compounds, or BVOCs — that serves as a forest-wide immune system. When a branch is undamaged, the emission profile is relatively simple: a background mix of monoterpenes (carbon-10 compounds like linalool and β-pinene) and small amounts of more reactive sesquiterpenes (carbon-15 compounds including β-caryophyllene and farnesene).
Attack changes that immediately. When Epirrita autumnata (the autumnal moth) or Operophtera brumata (the winter moth) begin feeding, the birch activates at minimum two parallel defense pathways: jasmonate signaling triggers rapid release of green leaf volatiles — hexenols, hexenals, and their esters — that warn neighboring trees within minutes, while a slower-building upregulation of terpene biosynthesis floods the air with more complex monoterpene and sesquiterpene blends over hours to days. At peak herbivory in the 2023 outbreak, the lead author team — Jolanta Rieksta, Amy Samantha Smart, Cleo L. Davie-Martin, and senior author Riikka Rinnan of the University of Copenhagen’s Center for Volatile Interactions (VOLT) — documented, per the study’s BVOC blend analysis, a markedly more complex compound profile than under undamaged conditions. These are not interchangeable chemicals: sesquiterpenes react with atmospheric oxidants several times faster than monoterpenes, and the proportions of the mixture matter for what kind of aerosol particles form downstream.
Why the Chemistry of the Plume Matters More Than Its Volume
Once BVOCs leave the forest, they enter the troposphere and react with ozone, hydroxyl radicals, and nitrate radicals. Those reactions produce low-volatility oxidation products that nucleate or condense onto existing particles to form secondary organic aerosols, or SOA. In the clean, particle-sparse atmosphere over the subarctic, SOA growth takes on outsized importance: research published in Geophysical Research Letters in 2022 found that SOA formation accounts for over half of cloud condensation nuclei concentrations in remote environments, and that SOA-derived cloud condensation nuclei alters the magnitude of cloud radiative forcing by approximately 0.1 watts per square meter. In the Arctic, which is warming nearly four times faster than the global average, the cooling effect of biogenically seeded clouds is not a footnote — it is one of the few negative feedbacks operating in a system dominated by positive ones.
The composition shift documented by Rieksta et al. matters because not all SOA particles are equally effective cloud seeds. Laboratory work by VanReken and colleagues has shown that limonene SOA activates more readily as cloud condensation nuclei than α-pinene-derived SOA. A moth outbreak that shifts the forest’s BVOC blend toward more reactive sesquiterpenes does not simply increase the quantity of aerosol particles — it may change their chemical composition and therefore their cloud-forming efficiency. Current BVOC emission models such as MEGAN (the Model of Emissions of Gases and Aerosols from Nature) do not capture composition shifts at this level of detail.
Following an Outbreak Year by Year in Abisko
The central methodological achievement of the Rieksta et al. study was temporal depth. The team made enclosure-based measurements across 256 mountain birch (Betula pubescens ssp. czerepanovii) trees during the growing seasons of 2021, 2022, and 2023, tracking the outbreak as it built from background herbivory through its escalating approach and peak in 2023. The last major outbreak in the same landscape had run through 2012–2013, making the three-year measurement window a rare opportunity to capture outbreak progression in a real forest rather than a controlled experimental setting.
The results unfolded in a pattern that surprised even researchers familiar with the laboratory literature. Early-season emissions during the 2023 peak outbreak year were measurably elevated above emissions at the actual moment of maximum caterpillar feeding — a counterintuitive result that the authors attributed to a “source depletion” mechanism: as defoliation consumed leaf area, the physical surface from which BVOCs could be emitted shrank, capping the output even while the per-leaf chemical response intensified. This constraint had not been documented before at natural outbreak scale.
After the peak defoliation passed and the trees flushed a second set of new leaves, some BVOC types remained elevated above baseline during that regenerative flush — extending the atmospheric chemical impact of the outbreak beyond its most visible phase.
When the Forest Floor Went in the Opposite Direction
Published alongside the main canopy study was a companion paper in the same journal examining the forest floor plants beneath the defoliated birch. The findings ran counter to intuition. As the birch canopy vanished, more sunlight reached the understory and soil temperatures rose while moisture declined. Under these changed microclimate conditions, the dominant ground-cover plant Empetrum nigrum (crowberry) experienced a sharp suppression of total BVOC emissions — in some measurements, down by 72 percent — even as the chemical blend shifted toward stress-induced compounds.
The divergence between canopy and forest floor illustrates why scaling from a single measurement to a landscape prediction is difficult. An outbreak simultaneously drives the birch canopy to emit more chemically complex BVOCs while driving the understory to emit far fewer total compounds. The net atmospheric signal depends on the balance between those two competing dynamics — a balance that varies with understory species composition, microclimate severity, and outbreak intensity in ways no current model fully represents.
Why Earth System Models Are Missing This Feedback
The broader climate-science significance of the new Rieksta et al. dataset is most visible when it is fed into regional atmospheric models. A 2023 presentation by collaborators using MEGAN-modelled BVOC data coupled to the WRF-CMAQ regional atmospheric chemistry system found that herbivory-induced emissions tripled monthly terpenes over the Fennoscandian birch forest during peak outbreak months, with measurable consequences for downward shortwave radiation and cloud radiative forcing. That pipeline — from satellite-detected defoliation through field-measured BVOC responses to regional atmospheric chemistry outcomes — had not previously been assembled in a single modeling framework.
It points to a feedback loop that runs like this: Arctic warming reduces overwinter egg mortality for geometrid moths, enabling larger and more frequent outbreaks; those outbreaks drive higher BVOC emissions and more reactive compound blends; those BVOCs form more SOA, which seeds more cloud condensation nuclei over a region with already-sparse particle concentrations; the added clouds reflect more solar radiation; and the resulting cooling partially moderates the warming that accelerated the outbreaks in the first place. The loop is real, physically grounded, and unrepresented in the Earth-system models that policymakers rely on for Arctic climate projections.
Outbreaks Are Getting Bigger and More Frequent
The timing of the new study matters because the geometrid moth–birch forest system is not stable. Moth outbreaks in northern Fennoscandia have been documented since at least the mid-19th century, recurring roughly on 9-to-10-year cycles driven by population dynamics. What has changed under warming conditions is the envelope of that system. Warmer winters have reduced overwinter egg mortality, allowing both Epirrita autumnata and Operophtera brumata to survive at higher densities and at higher altitudes and latitudes than historically possible. Novel geometrid species previously unable to survive subarctic winters have expanded their outbreak ranges northward, and because different species reach population peaks on slightly offset timescales, multi-species outbreaks tend to run longer.
The documented scale of past damage illustrates the magnitude of the phenomenon. During a single outbreak period from 2002 to 2009, roughly 10,000 square kilometers (3,861 square miles) of mountain birch forest across northern Norway, Sweden, and Finland was severely defoliated. In the most severely affected areas, tree mortality exceeded 90 percent a decade after the outbreak — with cascading effects on reindeer browsing, soil carbon, and vegetation structure. The Abisko region experienced its previous major outbreak in 2012–2013; the three-year dataset ends at the 2023 peak.
Who Funded and Where to Read More
The Rieksta et al. study involved researchers from the University of Copenhagen (lead institution), NTNU University Museum in Norway, Stiftelsen NILU (Norway), Aarhus University (Denmark), Lund University (Sweden), the Institute of Environmental Assessment and Water Research in Barcelona, and Sichuan University in China. Funding for the study came from six international bodies — the Independent Research Fund Denmark, the European Research Council, the VILLUM Foundation, the Danish National Research Foundation’s VOLT Center for Volatile Interactions, the National Natural Science Foundation of China, and the Novo Nordisk Foundation — reflecting the international character of the research collaboration.
Professor Riikka Rinnan, who leads VOLT — the world’s first research center dedicated to understanding how volatile organic compounds from all organisms shape the atmosphere — received a DKK 60 million grant (approximately $8.7 million at current exchange rates; exchange rate as of October 6, 2026; conversions are approximate) from the Danish National Research Foundation to run the center for six years following its June 2023 opening. The goal, in Rinnan’s words at the center’s launch, is to get the numbers right so they can be included in the broader climate equation and produce more accurate climate models.
The full paper is available open-access via Creative Commons Attribution 4.0 license at Communications Earth & Environment. The companion understory paper is also open-access at the same journal.
Frequently Asked Questions
How do moth caterpillars actually change the chemistry of the atmosphere?
When geometrid moth larvae chew through birch leaves, the trees activate chemical defense pathways that release reactive organic compounds — monoterpenes and sesquiterpenes — into the air at higher rates than undamaged trees produce. Those compounds travel into the troposphere and react with ozone and other oxidants to form microscopic solid and liquid particles called secondary organic aerosols. In the clean, particle-sparse Arctic atmosphere, those particles provide surfaces around which cloud droplets can form, potentially affecting cloudiness and how much sunlight reaches the surface. The Rieksta et al. study showed that the specific mix of compounds released during an outbreak is more chemically complex than the background mix — and that complexity affects how efficiently the resulting particles seed clouds.
Are these moth outbreaks getting worse because of climate change?
Yes, according to established research. Warmer winters reduce the number of geometrid moth eggs that freeze to death, allowing larger populations to reach outbreak densities the following year. Species of moth that were historically confined to maritime coastal areas have expanded northward and into continental subarctic zones where they previously could not survive. Because multiple moth species now peak in different years, outbreak events in some areas effectively run longer than the historical 9-to-10-year single-species cycle. Researchers studying the system in northern Norway documented that relatively small increases in defoliation pressure from range expansion could push forests past recovery thresholds, driving permanent transitions from birch woodland to open shrub tundra.
Why do current climate models not account for this feedback?
Standard BVOC emission models, including MEGAN — the Model of Emissions of Gases and Aerosols from Nature widely used in Earth-system modeling — were developed primarily to represent baseline plant emissions driven by temperature and light. They treat insect disturbance as a minor perturbation if they include it at all, and they do not represent the shift in chemical composition that an outbreak drives. The multi-year Abisko dataset, which tracks emission rates and compound blends across a full outbreak progression in a real forest, provides the kind of empirical grounding needed to build those parameters into next-generation models. Until they are, climate projections for the Arctic will be missing a feedback pathway that regional atmospheric modeling suggests is detectable.
Do the caterpillars have any effect on the plants growing beneath the birch trees?
Yes, but in the opposite direction from the trees themselves. When moth defoliation stripped the birch canopy, more sunlight reached the forest floor, raising temperatures and reducing soil moisture. Under those changed conditions, the dominant understory plant Empetrum nigrum (crowberry) suppressed its total BVOC output by up to 72 percent even while the chemical composition of what it did emit shifted toward stress-induced compounds. So at the landscape level, a moth outbreak simultaneously drives the dominant canopy tree to emit more and more complex volatiles while driving the ground-level vegetation to emit far less — a divergence that makes predicting the net atmospheric effect considerably more complicated than simply scaling up the canopy response.
