Contrasting thermophilization among forests, grasslands and alpine summits
Article metadata
Article Date: 18 March 2026
Article URL: https://www.nature.com/articles/s41586-025-09622-7
Article Image: https://www.nature.com/articles/s41586-025-09622-7/figures/4
Summary
This multi-author Nature study synthesises resurveys of plant communities across thousands of plots in forests, grasslands and alpine summits to compare rates and mechanisms of thermophilization — the community-level shift towards species associated with warmer temperatures. The authors combine large vegetation databases (forestREplot, GRACE, GLORIA), high-resolution climate products (CRU TS, WorldClim, CHELSA) and species records to estimate thermophilization rates, climatic debt and the relative roles of species gains, losses and abundance shifts.
Key Points
- Data and code are publicly available (Figshare DOI: https://doi.org/10.6084/m9.figshare.28368743.v5) and the study uses forestREplot, GRACE and GLORIA for the three ecosystem types.
- Thermophilization occurs across all three ecosystems but the drivers differ: grasslands show strong gains of warmth‑demanding species; forests show both gains and losses; alpine summits show losses and abundance shifts as dominant mechanisms.
- Median rates of growing-season temperature change at plot scale were around 0.22–0.31 °C per decade across ecosystems, but vegetation responses (thermophilization) do not map one-to-one onto warming, producing climatic debt especially in forests.
- Forests exhibit stronger microclimatic buffering and larger climatic mismatches over time, increasing the proportion of thermally mismatched species in understoreys.
- Topography, snow cover and slope aspect modulate thermophilization on alpine summits: rugged terrain, snow dynamics and aspect affect how communities shift.
- Analytical checks (different climate data resolutions, inclusion/exclusion of rare species, and spatial autocorrelation models including a Gaussian Process) show results are robust to methodological choices.
- Mechanistic decomposition confirms that gains, losses and abundance shifts together explain observed community temperature-affinity changes, but their relative importance varies by ecosystem.
- Implications include differential conservation priorities: grasslands may rapidly acquire warm‑adapted species, forests may carry growing climatic debt, and alpine summits face loss-driven community shifts with potential local extinctions.
Content summary
The authors compiled tens of thousands of vegetation records from resurveys across forests (forestREplot), grasslands (GRACE) and alpine summits (GLORIA), linked them to fine-scale climate and snow-cover datasets, and estimated shifts in community thermal affinity using quantiles (5th, median, 95th) of floristic temperature distributions. Bayesian mixed-effects models were used to estimate thermophilization rates and climatic debt, with sensitivity analyses for rare species, climate data resolution and spatial autocorrelation. Key findings are that thermophilization is ubiquitous but spatially and mechanistically heterogeneous: grasslands exhibit thermophilization mainly via species gains; forests show a mix of gains and losses and increasing mismatch with macroclimate; alpine summits are dominated by species losses and abundance changes influenced by snow and topography. The study stresses that microclimate buffering and local terrain create important lags and variation in community responses to warming.
Context and relevance
This paper advances our understanding of biodiversity responses to climate warming by comparing ecosystem types with different microclimatic buffering, dispersal constraints and topographic complexity. It links to broader findings on climatic debt, range shifts and microclimate mediation of warming impacts. For conservationists and land managers, the contrasted mechanisms imply different interventions: monitor and manage colonisation pathways in grasslands, address understorey climatic debt and stand structure in forests, and prioritise refugia and connectivity for alpine flora. Methodologically, the open data and code promote reproducibility and further cross-ecosystem syntheses.
Why should I read this?
Quick version: if you care about how plant communities are actually responding to warming (not just the air temperature numbers), this paper is a solid, data-heavy read. It tells you which ecosystems are changing fast, how they’re changing (new arrivals vs. losers vs. shifting abundances) and why local factors like snow and slope matter. Saves you combing dozens of resurveys — they did the heavy lifting and show where to look next.
Author note
Punchy take: this is a major, well-backed synthesis from a huge international team. The findings are directly relevant to researchers, conservation planners and anyone tracking climate-driven biodiversity change — especially where microclimate and topography modulate responses.