Recent vegetation shifts in the French Alps with winners outnumbering losers

Abstract

1 INTRODUCTION

Maintaining the integrity and biodiversity of natural ecosystems is a growing global concern, as reflected in initiatives such as the European Green Deal, the EU Biodiversity Strategy for 2030, the Sustainable Development Goals, and the Kunming-Montreal Global Biodiversity Framework (GBF). Protecting terrestrial ecosystems, preserving their biodiversity and restoring degraded ecosystems are essential for sustaining their contributions to people (e.g. carbon sequestration, food supply, timber production and flood protection). Despite these efforts, terrestrial biodiversity has continued to decline substantially over recent decades (IPBES, 2019). Achieving the GBF targets, that is, stabilize biodiversity loss by 2030 and foster the recovery of natural ecosystems in the subsequent two decades, requires a robust framework for identifying declining and expanding species, quantifying range shifts and understanding the associated functional and evolutionary consequences (Cardinale et al., 2018; Dornelas et al., 2019). Regional-scale studies are increasingly recognized as pivotal in advancing progress towards these targets Gonzalez et al. (2023).

Mountain ecosystems are critical sentinels of global change Guisan et al. (2019), making them essential for studying biodiversity dynamics. These ecosystems are characterized by pronounced plant stratification along elevational gradients, historically shaped by a combination of temperature and humidity. Over the past ~10,000 years, this stratification has also been influenced by human activities, including cycles of settlement and land abandonment (Gehrig-Fasel et al., 2007; MacDonald et al., 2000). The stratification of vegetation along these gradients significantly impacts other components of the local biodiversity, influencing both above-ground communities Martinez-Almoyna et al. (2024) and below-ground systems (Calderón-Sanou et al., 2024). Together, these components provide essential services for human well-being in mountainous regions, such as carbon sequestration, timber production and pastures (Bardgett & van der Putten, 2014; Delgado-Baquerizo et al., 2020).

Due to their unique environmental contexts, mountain ecosystems world-wide host exceptionally high levels of plant biodiversity (Rahbek et al., 2019), with many endemic species and a diverse array of life forms. Higher alpine belts, for instance, are dominated by a few plant families that have evolved traits to tolerate low temperatures (Qian et al., 2021), which could make them particularly susceptible to ongoing climate change (Chen et al., 2011; Lenoir et al., 2008). Extreme environmental conditions have historically protected mountain ecosystems from biological invasions. However, growing human appropriation of landscapes and climate change are now driving shifts in plant distributions. While an increase in plant richness on high latitude (Myers-Smith et al., 2011) and high elevation is well-documented (Lamprecht et al., 2018; Pauli et al., 2012; Steinbauer et al., 2018), extinctions at lower elevations are progressing more slowly (Alexander et al., 2018). Despite evidence of upward shifts for certain species, a comprehensive assessment of distribution changes across all species remains lacking.

The detection of species change forms a cornerstone of the detection-attribution framework advocated by GEO-BON Gonzalez et al. (2023). Detection of vegetation change is usually assessed via changes in species coverage or frequency (Klinkovská et al., 2024). Robust estimates require large historical datasets with high spatial resolution. For instance, Jandt et al. (2022) used data from long-term repeated vegetation-plot records to report more losses than gains in the German flora over the last century, which corroborates the findings of Timmermann et al. (2015) in the Danish flora. These analyses also offer a means to assess the current conservation of declining species. Klinkovská et al. (2024) used data on the Czech flora, collected over the past six decades, to determine whether expanding species are invasive or opportunistic colonizers responding to climate change (Eichenberg et al., 2021; Jandt et al., 2022). These insights are vital for strengthening biodiversity protection and restoration, particularly at regional scales where conservation efforts are most effective.

Beyond identifying species that lost or gained territory over the past decades, the ability to profile which species are expanding or contracting their ranges but also whether some clades show distinct shifting patterns may provide a deeper understanding of the underlying mechanisms behind observed changes and may allow better prediction of future changes (Lavergne et al., 2010). Trait-based and phylogenetic approaches offer a particularly valuable framework for this. Functional traits—morphological and physiological characteristics of plants that affect their fitness (Violle et al., 2007)—can be used to better understand species' functional strategies and their variation along environmental gradients (Garnier et al., 2015). For instance, traits provide insights into a species' tolerance for temperature extremes, humidity fluctuations or resource conservation strategies (Lavorel & Garnier, 2002). Revealing that winner species exhibit specific combinations of traits, such as those adapted to warmer conditions or frequent disturbances, can help elucidate the mechanisms behind biodiversity changes. For example, Guo et al. (2018) linked colonizing species to ruderal strategies, that is, species with a rapid completion of the life cycle and benefiting from disturbances (Grime, 1977), while Henn et al. (2024) associated winners with acquisitive trait strategies, which are species with faster growth but lower stress tolerance (Reich et al., 1997). However, the relationship between functional traits and species temporal trends may vary depending on life forms (Delalandre et al., 2023). Nevertheless, some traits seem to be key to better understand plant responses to warmer temperature such as plant height (Bjorkman et al., 2018; Maes et al., 2020; Timmermann et al., 2015) or specific leaf area (Guittar et al., 2016; Venn et al., 2011) as those traits reflect both competitive ability and resource capture and retention trade-offs (Reich, 2014). While conceptually it makes sense that functional traits should be able to predict winners and losers under climate change, empirical evidence is mixed. Some studies found support for a relationship between the trends of species and their functional traits (Kühn et al., 2021; Pinho et al., 2025; Soudzilovskaia et al., 2013), while others did not (García Criado et al., 2023). Wiens et al. (2010) suggested that traits shaping species' ecological niches are often phylogenetically conserved, which implies that responses to climate change could be too (Burns & Strauss, 2011). While bird population declines have shown strong phylogenetic signals (Davis et al., 2010; Lavergne et al., 2013), evidence for plants remains sparse. Building models that predict winner and loser species from their functional traits and phylogeny can become important tools for management to anticipate future changes.

In this paper, we address the challenge of identifying, quantifying, characterizing winning, stable and losing plant species over the last 30 years across the French Alps. To achieve this, we leveraged an extensive semi-structured plant dataset encompassing approximately 4250 species sampled over the past 30 years by expert botanists. This dataset represents about 60% of the total plant diversity in France and captures the entire plant diversity of the French Alps. To correct for temporal and spatial sampling biases, we employed the Frescalo method (Hill, 2012) to generate unbiased time series at the species level, while accounting for associated uncertainties. Using a Bayesian framework, we then quantified temporal trends and then classified species as winners, losers or stable. We expected more species to gain distribution area than to lose area (Finderup Nielsen et al., 2019; Jandt et al., 2022), given that the climate becomes overall more favourable for plants in the French Alps (i.e. warmer temperatures) and grazing pressure decreases in some areas (i.e. land abandonment). Additionally, we examined the distribution of these species across families to identify those that are disproportionately losing area and, potentially, species in the long term. Using a molecular species-level phylogeny of the European Alps, we further tested whether loser and winner species show specific phylogenetic patterns, which could jeopardize or benefit entire clades. After identifying taxonomically and phylogenetically area-gaining and losing species, we characterized the biogeographic status of these species. We expected species with high IUCN conservation status to be over-represented in the group of species losing distribution area, and invasive species to be over-represented in the group of winner species. We also tested Grime's strategies (Grime, 1977), with the expectation that ruderal species colonizing newly disturbed areas would be more likely to be winner species (Klinkovská et al., 2024). Finally, we employed a trait-based modelling approach based on machine learning to assess whether the observed species trends can be predicted based on functional traits. We hypothesized that species exhibiting rapid growth and higher tolerance to elevated temperatures (thermophilous species; De Frenne et al., 2015) would show increasing trends. Through this integrative framework, our study provides critical insights into temporal plant dynamics, offering a robust foundation for understanding and predicting biodiversity changes in mountain ecosystems.