The phenomenon the team documents is called overcompensation, and it is defined precisely: annual productivity after a late-spring frost exceeds the productivity expected in a comparable year without frost. Drawing on eddy-covariance carbon flux measurements from tower networks such as FLUXNET2015, AmeriFlux and ICOS, alongside satellite-derived estimates of gross primary productivity and solar-induced chlorophyll fluorescence across the Northern Hemisphere, the researchers found that overcompensation occurred in roughly 21.5 to 41.0 percent of the frost events they analysed, depending on the dataset and frost threshold used. In those cases, the post-frost regrowth was large enough not only to cancel out the initial loss but to push the year’s total carbon uptake above the non-frost baseline.

The technical machinery behind this conclusion is considerable. Eddy-covariance towers measure the net exchange of carbon dioxide between ecosystems and the atmosphere at half-hourly resolution, allowing researchers to partition gross primary productivity — the total amount of carbon plants fix through photosynthesis — from ecosystem respiration. The team used these ground-based records to track productivity anomalies before and after frost events, fitting smoothed anomaly curves with generalized additive models to quantify the initial dip, the subsequent regrowth and the final annual balance. They then repeated the analysis across the hemisphere using multiple independent satellite products, including the contiguous solar-induced chlorophyll fluorescence dataset known as CSIF, FluxSat-GPP, GLASS-GPP and GOSIF-GPP, testing their results against frost temperature thresholds of minus one, minus three and minus five degrees Celsius to ensure the pattern was not an artefact of how a frost is defined.

One of the most striking results is geographic. Overcompensation was strongest at high latitudes, a pattern that held across the range from 30 degrees north to 70 degrees north in both the fluorescence-based and GPP-based satellite records. The authors attribute this latitudinal gradient to two reinforcing mechanisms: frost events at high latitudes tend to cause less damage to vegetation, and the regrowth that follows tends to be greater. Cold-adapted northern ecosystems are, in a sense, built for frost. Many boreal and Arctic plants possess deep freezing tolerance, an evolutionary legacy of climates where sub-zero temperatures during the growing season are routine rather than exceptional. When a frost does strike, these plants lose less photosynthetic machinery, and when the growing season resumes, the remaining canopy — combined with a longer effective period of favourable conditions — can drive photosynthesis to levels that exceed the undamaged baseline.

The satellite and tower observations alone could not settle whether overcompensation is a genuine ecological response or a statistical quirk of particular years. To address this, the team turned to manipulative field experiments conducted between 2018 and 2024, in which researchers imposed controlled spring frosts on grassland plots and measured the consequences for gross primary productivity over subsequent months. These experiments confirmed the phenomenon directly: frosted grassland plots showed the characteristic pattern of initial productivity loss followed by regrowth that, in a meaningful share of cases, carried annual productivity beyond that of unfrosted controls. The experiments also revealed two factors that strengthen the compensatory response — warmer ambient temperatures and higher plant species diversity.

Both of those experimental findings carry implications well beyond grassland ecology. Warmer temperatures accelerate regrowth after damage by extending the window of metabolically favourable conditions and speeding the rebuilding of leaf area, so as the climate warms, the recovery phase following a frost is likely to become more powerful. Species diversity, meanwhile, is thought to buffer ecosystems against disturbance through complementary resource use and the presence of species with varying frost sensitivities: when a frost knocks back the most frost-vulnerable species, less-affected neighbours with different phenologies and canopy positions can expand into the freed-up space, sustaining ecosystem-level photosynthesis. This is consistent with a broader body of biodiversity–ecosystem functioning research showing that diverse communities tend to be more productive and more resilient, and it suggests that conserving species richness may be one lever for maintaining carbon uptake in an era of increasing climate extremes.

Perhaps the most consequential finding is temporal. Over the past two decades, the study found that productivity overcompensation has intensified, driven by two converging trends: warming has decreased the losses that frosts inflict, while simultaneously increasing the regrowth that follows. This does not mean frost has ceased to matter. Other research has documented severe and economically devastating frost events, such as the widespread damage to European vineyards and orchards in 2017, and some analyses suggest that as phenology shifts earlier in a warming climate, the risk of frost exposure for certain species and regions — European beech at higher elevations, for example — may actually be increasing. But at the hemispheric scale and across the datasets examined here, the balance of loss and regrowth has been tipping toward compensation, with warming acting on both sides of the ledger.

To peer into the future, the team used projections from the Norwegian Earth System Model (NorESM2-MM) under CMIP6 climate scenarios, examining how vegetation loss, regrowth and overcompensation are expected to evolve under four different futures. The projections indicate that overcompensatory responses will become increasingly common as the century progresses, because continued warming further erodes frost damage while extending and intensifying the regrowth period. If that projection holds, the standard practice of treating climate extremes as unambiguous negative perturbations to the terrestrial carbon sink will need revision — at least for this particular class of extreme.

The study’s authors are careful about what these results do and do not imply. Overcompensation occurred in a minority of frost events; in most cases, frosts still reduced annual productivity relative to the non-frost baseline, and the magnitude of the initial damage varies enormously with the timing, intensity and duration of the freezing temperatures, as well as with the developmental stage of the vegetation when the frost strikes. Species that have already broken bud and expanded tender foliage are far more vulnerable than those still dormant, and the interaction of frost with subsequent stresses such as summer drought can compound damage in ways that a single-season analysis may not capture. Overcompensation is a real and measurable pattern, not a universal law.

Even so, the implications for Earth system science are substantial. The land biosphere currently absorbs a substantial fraction of human carbon emissions, and the models used to project that sink’s future must correctly represent how ecosystems respond to climate extremes. If a significant fraction of frost events trigger compensatory productivity gains rather than net losses, then models that treat frost purely as a productivity penalty will systematically misestimate the carbon balance of affected regions and years. The authors argue that incorporating vegetation overcompensation into Earth system models will be essential for accurately assessing and predicting the impacts of climate extremes on terrestrial carbon cycling under global climate change. In the meantime, the study offers a vivid reminder that ecosystems are not passive victims of weather: scratched by a late frost, the green machinery of the Northern Hemisphere frequently grows back harder — and sometimes grows back bigger.

Subject of Research: Overcompensation of vegetation productivity following late-spring frost events across the Northern Hemisphere

Article Title: Overcompensation of vegetation productivity after late-spring frost in the Northern Hemisphere

Article References: Qiu, H., Han, J., Gu, H., Peñuelas, J., Wan, S., & Chen, L. (2026). Overcompensation of vegetation productivity after late-spring frost in the Northern Hemisphere. Nature Plants. https://doi.org/10.1038/s41477-026-02407-4

Image Credits: AI Generated

DOI: 10.1038/s41477-026-02407-4

Keywords: late-spring frost, overcompensation, gross primary productivity, eddy covariance, solar-induced chlorophyll fluorescence, carbon cycle, phenology, biodiversity, climate change, Earth system models, grasslands, Northern Hemisphere