Anthropogenic impacts are reshaping plant biodiversity patterns, yet how community-composition shifts track environmental change at large spatial and temporal scales remains unclear. Here, we quantified trends in community-mean plant ecological indicator values (light, temperature, soil moisture, soil nitrogen, and soil reaction) across European vegetation between 1960 and 2020. We used spatiotemporal interpolation based on 644,524 plots and analyzed 18,345 time series encompassing diverse habitats. We found a clear shift in community composition over the past six decades with a steep increase in nitrogen-demanding species across all main habitat types, accompanied by a moderate increase in shade-tolerant species. Forest communities shifted toward species associated with higher soil pH, while wetland communities showed a decline in moisture-dependent species over time. Conversely, temperature indicator values were largely stable, except for recent thermophilization in alpine habitats. Our results indicate a widespread trend toward denser vegetation driven by eutrophication and changes in management practices.

Predicting which non-native plant species will become established and where is critical for conserving and managing biodiversity. Theory suggests that the mycorrhizal strategy of non-native plants may predict their establishment success. Here we combine a global dataset of 440,788 vegetation plots with data on plant native status and mycorrhizal type to assess mycorrhizal strategy of non-native plants. The mycorrhizal strategy of non-native plants varies strongly across biomes. Across grassland and desert biomes, non-native species are more frequently non-mycorrhizal than native species, whereas in other biomes non-native species are more likely to be mycorrhizal, most commonly arbuscular-mycorrhizal. Disturbance type and intensity are key predictors of mycorrhizal strategy of non-native species, as mycorrhizal species are favoured by landscape modification and non-mycorrhizal species by natural and human-caused disturbance events. Facultatively mycorrhizal species are consistently under-represented among non-native plants compared with natives, suggesting that symbiotic flexibility does not confer an advantage for non-natives as previously expected. Our study shows that non-native mycorrhizal strategy varies across biogeographical contexts and disturbance, highlighting the need for region-specific prevention and management approaches to plant species introductions.

Light regulates ecological processes from organism performance to ecosystem functioning, yet its role in shaping plant diversity across spatial scales remains unresolved. Here, we combine > 650,000 vegetation plots, 14,835 species-specific light indicator values based on Ecological Indicator Values for Europe (EIVE), and Earth observation data to provide the first pan-European assessment of how light regimes near the ground shape plant diversity across scales. Incorporating light-regime metrics beyond macroclimatic and edaphic drivers improved predictions of plant diversity, with relative increases in explained variance of 3% and 27.4–34.5% at the local and landscape scale, respectively. Plant diversity peaked under intermediate-to-high light availability coupled with high spatial heterogeneity. Strikingly, at the landscape level, we identified a pronounced mismatch between the light regimes currently prevailing across Europe and those associated with high plant diversity. Together, these findings reveal light regime as a fundamental axis for plant biodiversity and suggest that widespread shifts towards very low light availability and reduced spatial heterogeneity may decrease plant diversity across Europe. Maintaining non-shaded or increasing spatial heterogeneity in light conditions near the ground could therefore help reduce this mismatch, expanding the extent of light regimes with high potential to support plant diversity.

To address the urgent biodiversity crisis, it is crucial to understand the nature of plant assemblages. The distribution of plant species is shaped not only by their broad environmental requirements but also by micro-environmental conditions, dispersal limitations, and direct and indirect species interactions. While predicting species composition and habitat type is essential for conservation and restoration purposes, it remains challenging. In this study, we propose an approach inspired by advances in large language models to learn the 'syntax' of abundance-ordered plant species sequences in communities. Our method, which captures latent associations between species across diverse ecosystems, can be fine-tuned for diverse tasks. In particular, we show that our methodology is able to outperform other approaches to (1) predict species that might occur in an assemblage given the other listed species, despite being originally missing in the species list (16.53% higher accuracy in retrieving a plant species removed from an assemblage than co-occurrence matrices and 6.56% higher than neural networks), and (2) classify habitat types from species assemblages (5.54% higher accuracy in assigning a habitat type to an assemblage than expert system classifiers and 1.14% higher than tabular deep learning). The proposed application has a vocabulary that covers over 10,000 plant species from Europe and adjacent countries and provides a powerful methodology for improving biodiversity mapping, restoration and conservation biology. As ecologists begin to explore the use of artificial intelligence, such approaches open opportunities for rethinking how we model, monitor and understand nature.

Anthropogenic biodiversity decline threatens the functioning of ecosystems and the many benefits they provide to humanity1. As well as causing species losses in directly affected locations, human influence might also reduce biodiversity in relatively unmodified vegetation if far-reaching anthropogenic effects trigger local extinctions and hinder recolonization. Here we show that local plant diversity is globally negatively related to the level of anthropogenic activity in the surrounding region. Impoverishment of natural vegetation was evident only when we considered community completeness: the proportion of all suitable species in the region that are present at a site. To estimate community completeness, we compared the number of recorded species with the dark diversity-ecologically suitable species that are absent from a site but present in the surrounding region2. In the sampled regions with a minimal human footprint index, an average of 35% of suitable plant species were present locally, compared with less than 20% in highly affected regions. Besides having the potential to uncover overlooked threats to biodiversity, dark diversity also provides guidance for nature conservation. Species in the dark diversity remain regionally present, and their local populations might be restored through measures that improve connectivity between natural vegetation fragments and reduce threats to population persistence.

The Natura 2000 (N2K) network of protected areas is one of the main tools for area-based conservation in the European Union (EU), yet its role in preserving plant biodiversity requires better understanding. We examined data kept in the European Vegetation Archive from over 1.2 million vegetation plots and obtained over 14.2 million plant species occurrences. To test the N2K network's representativeness of plant species gamma diversity, we compared the number and percentage of native and conservation priority species in- and outside the N2K network throughout the EU and for individual countries, biogeographical regions, and combinations thereof. We then determined whether N2K sites hosted more species than sites outside the network with the species-area relationship. Overall, almost 90% of the native vascular plant species occurred at least once in the N2K network. Yet, significant variation exists across countries and biogeographical regions-from 0% of species in the Boreal region of Lithuania, to 98% in the Alpine region of Croatia-indicating that local N2K sites are not equally representative of the regional gamma diversity. Nonetheless, the N2K network contains more species than land outside the network when area is taken into account. The planned expansion of the N2K network, as mandated by the European Biodiversity Strategy for 2030, should prioritize areas with currently underrepresented elements of the EU vascular flora.

Forest ecosystems will play a critical role in achieving policy targets for biodiversity and conservation, such as those set out in the EU Biodiversity strategy for 2030. However, practitioners need to know where forests of high conservation value are to make the best-informed decisions about which forests to prioritize. Here, we combine airborne LiDAR (airborne laser scanning/ALS), optical satellite imagery, and gridded datasets on soil and water availability with machine learning models to predict forests' conservation value across Denmark. We then use change-detection algorithms to identify forests that had been disturbed since the collection of the LiDAR data to produce up-to-date estimates for the year 2020. Our models reached a high predictive capacity (82% accuracy) and suggested that 1982 km(2) (~31%) of Denmark's forests were of potential high conservation value. Our study demonstrates the utility of data fusion approaches to identify forest areas of high value for conservation at fine spatial resolutions (~10-100 m) and nationwide extents. However, uncertainties remain in our approach. Hence, our findings should be used to guide field-based assessments to confirm the in situ conservation value of the forests. Only in combination with such in situ data will approaches like ours enable decision makers to better protect forest biodiversity.

Deadwood is a vital part of the habitat for many threatened species of animals, plants and fungi. Thus, presence of deadwood is an important indicator for the probability that a given site harbors threatened species. Nowadays, field work is the most common method for monitoring dead trees. However, it is time consuming, costly and labor-intensive. Therefore, there is a need for an automatic method for mapping and monitoring deadwood. The combination of fine-resolution remote sensing and deep learning techniques have a potential to provide exactly this. Unfortunately, due to the typical location of lying deadwood under the canopy, this is a challenging task as the visibility of the lying trees is limited notably with optical remote sensing techniques. Therefore, laser scanning data seems to be the most appropriate for this purpose as it can penetrate the canopy to some extent and hence gather data from a forest floor. In this work we aim at the development of methods enabling detection of lying deadwood at the national scale in protected forests and we focus on the presence of deadwood in 15-meter-radius circular plots. To achieve this goal, we use Airborne Laser Scanning (ALS) data that is publicly available for the whole Denmark and, as a reference, almost 6000 forestry plots acquired as a part of the Danish national habitats monitoring program. The binary classification into plots that contain deadwood and the ones that do not is performed using SparseCNN deep neural network. In this study we showed that it is possible to detect plots having deadwood with an overall accuracy of around 61%. However, the accuracy of the classifier depends on the volume of the deadwood present in a plot.

Biodiversity change forecasts rely on long-term time series, but such data are often scarce in space and time. Here, we interpolated spatiotemporal changes in species richness using a new method based on machine learning that does not require temporal replication at sites. Using 698,692 one-time sampled vegetation plots, we estimated trends in vascular plant alpha diversity across Europe and validated our approach against 22,852 independent time series. We found an overall near-zero net change in species richness between 1960 and 2020. However, species richness generally declined from 1960 to 1980 and increased from 2000 to 2020 across habitats. Declines were most pronounced in forests, but trends varied across habitats and regions, with overall increases at higher latitudes and elevations, and declines or stable trends elsewhere. Our findings demonstrate how data without temporal replication can be used to reveal context-dependent biodiversity dynamics, underscoring their importance for conservation and management.

Knowledge of local plant community characteristics is imperative for practical nature planning and management, and for understanding plant diversity and distribution drivers. Today, retrieving such data is only possible by fieldwork and is hence costly both in time and money. Here, we used nine bands from multispectral high-to-medium resolution (10–60 m) satellite data (Sentinel-2) and machine learning to predict local vegetation plot characteristics over a broad area (approx. 30,000 km2) in terms of plants’ preferences for soil moisture, soil fertility, and pH, mirroring the levels of the corresponding actual soil factors. These factors are believed to be among the most important for local plant community composition. Our results showed that there are clear links between the Sentinel-2 data and plants’ abiotic soil preferences, and using solely satellite data we achieved predictive powers between 26 and 59%, improving to around 70% when habitat information was included as a predictor. This shows that plants’ abiotic soil preferences can be detected quite well from space, but also that retrieving soil characteristics using satellites is complicated and that perfect detection of soil conditions using remote sensing—if at all possible—needs further methodological and data development.

Plant communities are composed of species that differ both in functional traits and evolutionary histories. As species’ functional traits partly result from their individual evolutionary history, we expect the functional diversity of communities to increase with increasing phylogenetic diversity. This expectation has only been tested at local scales and generally for specific growth forms or specific habitat types, for example, grasslands. Here we compare standardized effect sizes for functional and phylogenetic diversity among 1,781,836 vegetation plots using the global sPlot database. In contrast to expectations, we find functional diversity and phylogenetic diversity to be only weakly and negatively correlated, implying a decoupling between these two facets of diversity. While phylogenetic diversity is higher in forests and reflects recent climatic conditions (1981 to 2010), functional diversity tends to reflect recent and past climatic conditions (21,000 years ago). The independent nature of functional and phylogenetic diversity makes it crucial to consider both aspects of diversity when analysing ecosystem functioning and prioritizing conservation efforts. Functional diversity and phylogenetic diversity are expected to be positively correlated. Here the authors show that the covariation between these metrics in vascular plant communities around the world is often either inconsistent or negative.

AimsThe accurate classification of habitats is essential for effective biodiversity conservation. The goal of this study was to harness the potential of deep learning to advance habitat identification in Europe. We aimed to develop and evaluate models capable of assigning vegetation-plot records to the habitats of the European Nature Information System (EUNIS), a widely used reference framework for European habitat types.LocationThe framework was designed for use in Europe and adjacent areas (e.g., Anatolia, Caucasus).MethodsWe leveraged deep-learning techniques, such as transformers (i.e., models with attention components able to learn contextual relations between categorical and numerical features) that we trained using spatial k-fold cross-validation (CV) on vegetation plots sourced from the European Vegetation Archive (EVA), to show that they have great potential for classifying vegetation-plot records. We tested different network architectures, feature encodings, hyperparameter tuning and noise addition strategies to identify the optimal model. We used an independent test set from the National Plant Monitoring Scheme (NPMS) to evaluate its performance and compare its results against the traditional expert systems.ResultsExploration of the use of deep learning applied to species composition and plot-location criteria for habitat classification led to the development of a framework containing a wide range of models. Our selected algorithm, applied to European habitat types, significantly improved habitat classification accuracy, achieving a more than twofold improvement compared to the previous state-of-the-art (SOTA) method on an external data set, clearly outperforming expert systems. The framework is shared and maintained through a GitHub repository.ConclusionsOur results demonstrate the potential benefits of the adoption of deep learning for improving the accuracy of vegetation classification. They highlight the importance of incorporating advanced technologies into habitat monitoring. These algorithms have shown to be better suited for habitat type prediction than expert systems. They push the accuracy score on a database containing hundreds of thousands of standardized presence/absence European surveys to 88.74%, as assessed by expert judgment. Finally, our results showcase that species dominance is a strong marker of ecosystems and that the exact cover abundance of the flora is not required to train neural networks with predictive performances. The framework we developed can be used by researchers and practitioners to accurately classify habitats. Hdm-framework, by leveraging vascular plant species composition coupled with environmental data, enhances vegetation classification, clearly outperforming the traditional expert systems on habitat type identification. The deep-learning models, trained on European flora, offer unparalleled accuracy, efficiency and robustness across diverse ecosystems. This advancement marks a crucial step toward effective biodiversity conservation planning, monitoring and assessment.image

Aims: We introduce ReSurveyEurope - a new data source of resurveyed vegetation plots in Europe, compiled by a collaborative network of vegetation scientists. We describe the scope of this initiative, provide an overview of currently available data, governance, data contribution rules, and accessibility. In addition, we outline further steps, including potential research questions. Results: ReSurveyEurope includes resurveyed vegetation plots from all habitats. Version 1.0 of ReSurveyEurope contains 283,135 observations (i.e., individual surveys of each plot) from 79,190 plots sampled in 449 independent resurvey projects. Of these, 62,139 (78%) are permanent plots, that is, marked in situ, or located with GPS, which allow for high spatial accuracy in resurvey. The remaining 17,051 (22%) plots are from studies in which plots from the initial survey could not be exactly relocated. Four data sets, which together account for 28,470 (36%) plots, provide only presence/absence information on plant species, while the remaining 50,720 (64%) plots contain abundance information (e.g., percentage cover or cover-abundance classes such as variants of the Braun-Blanquet scale). The oldest plots were sampled in 1911 in the Swiss Alps, while most plots were sampled between 1950 and 2020. Conclusions: ReSurveyEurope is a new resource to address a wide range of research questions on fine-scale changes in European vegetation. The initiative is devoted to an inclusive and transparent governance and data usage approach, based on slightly adapted rules of the well-established European Vegetation Archive (EVA). ReSurvey:Europe data are ready for use, and proposals for analyses of the data set can be submitted at any time to the coordinators. Still, further data contributions are highly welcome.

Indicators of habitat condition are essential for tracking conservation progress, but measuring biotic, abiotic and landscape characteristics at fine resolution over large spatial extents remains challenging. In this viewpoint article, we provide a comprehensive synthesis of the challenges and solutions for consistently measuring and monitoring habitat condition with remote sensing using airborne Light Detection and Ranging (LiDAR) and affordable Unmanned Aerial Vehicles (UAVs) over multiple sites and transnational or continental extents. Key challenges include variability in sensor characteristics and survey designs, non-transparent pre-processing workflows, heterogeneous and complex data, issues with the robustness of metrics and indices, limited model generalizability and transferability across sites, and difficulties in handling big data, such as managing large volumes and utilizing parallel or distributed computing. We suggest that a collaborative cloud virtual research environment (VRE) for habitat condition research and monitoring could provide solutions, including tools for data discovery, access, and data standardization, as well as geospatial processing workflows for airborne LiDAR and UAV data. A VRE would also improve data management, metadata standardization, workflow reproducibility, and transferability of structure-from-motion algorithms and machine learning models such as random forests and convolutional neural networks. Along with best practices for data collection and adopting FAIR (findability, accessibility, interoperability, reusability) principles and open science practices, a VRE could enable more consistent and transparent data processing and metric retrieval, e.g., for Natura 2000 habitats. Ultimately, these improvements would support the development of more reliable habitat condition indicators, helping prevent habitat degradation and promoting the sustainable use of natural resources.

Grazing at near-natural stocking rates is increasingly rare, whereas abandonment and overgrazing is common, despite both leading to loss of threatened species. Here, we evaluate a biodiversity-promotion strategy of a beef-producing company involving livestock grazing for conservation. Using field surveys, a national biodiversity map, and farmer interviews, we evaluate the conservation potential of farmers committing to a conservation grazing scheme. Most farmers practiced summer grazing, and-contrary to their pledge-primarily grazed areas of low conservation value, and with stocking rates far above carrying capacity. We conclude that a conservation grazing beef brand will only have real value for threatened species if committed to, and held accountable to, a scientifically informed grazing scheme. Our results point to rural traditions, legal regulations, and economic incentives as factors obstructing of a beneficial grazing practice. Hence, the results of this study call for a critical and scientifically informed approach to future land-sharing policies.

The relationship between phylogenetic diversity (PD) and functional diversity ( FD) is important for understanding the mechanisms of community assembly. The traditional view assumes a coupled (positively correlated) relationship between these two diversity measures, suggesting that competitive exclusion and environmental filtering are important drivers of both phylogenetic and functional structure of communities. In contrast, there is evidence that communities might deviate from this pattern, exhibiting either phylogenetic overdispersion connected with trait convergence (decoupled PD) or functional overdispersion connected with phylogenetic clustering (decoupled FD). In this study, we examined the relationship between PD and FD within vascular-plant communities in European grasslands, focusing on decoupled PD-FD patterns. We hypothesized that the decoupled patterns are connected with past or current environmental changes and are rarer in comparison with the coupled PD-FD pattern, reflecting long-term relatively stable environments. We used 81,484 plots (communities) of European dry, mesic, wet and alpine grasslands, containing 4,119 angiosperm species, and data on six functional traits relevant for different plant functions and habitats (plant height, leaf area, specific leaf area, leaf nitrogen content, seed mass and lateral spreading distance). Functional diversity was evaluated in two ways - as a single combined measure and as variability in each trait separately. We found various PD-FD patterns across different habitats, traits and regions, with the coupled pattern widespread but not universal. In many communities, we detected the tendency towards decoupled PD, likely caused by environmental filtering of phylogenetically diverse species pools. This was most pronounced in dry grasslands, and also in wet and alpine grasslands when FD based on plant height, leaf area or seed masswas considered. In contrast, the tendency towards decoupled FD was detected only in mesic and wet grasslands for leaf nitrogen content and lateral spreading distance, possibly due to competitive interactions among species interplaying with land-use history. Decoupled PD is relatively common in European grasslands, especially in mountainous and hilly areas of central and southern Europe and in parts of western Europe with a mild climate. This likely results from refugial effects that have preserved many distinct phylogenetic lineages, but their species are functionally similar due to environmental filters that affect the assembly of present-day grassland communities. We demonstrate that PD and FD may reflect different aspects of community structure and assembly mechanisms, and suggest that the phenomenon of decoupled PD and FD deserves more systematic study.

A key aspect of nature conservation is knowledge of which aspects of nature to conserve or restore to favor the characteristic diversity of plants in a given area. Here, we used a large plant dataset with >40 000 plots combined with airborne laser scanning (lidar) data to reveal the local characteristics of habitats having a high plant dark diversity-that is, absence of suitable species-at national extent (>43 000 km(2)). Such habitats have potential for reaching high realized diversity levels and hence are important in a conservation context. We calculated 10 different lidar based metrics (both terrain and vegetation structure) and combined these with seven different field-based measures (soil chemistry and species indicators). We then used Integrated Nested Laplace Approximation for modelling plant dark diversity across 33 North European habitat types (open landscapes and forests) selected by the European communities to be important. In open habitat types high-dark-diversity habitats had relatively low pH, high nitrogen content, tall homogenous vegetation, and overall relatively homogenous terrains (high terrain openness) although with a rather high degree of local microtopographical variations. High-dark-diversity habitats in forests had relatively tall vegetation, few natural-forest indicators, low potential solar radiation input and a low cover of small woody plants. Our results highlight important vegetation, terrain- and soil-related factors that managers and policymakers should be aware of in conservation and restoration projects to ensure a natural plant diversity, for example low nutrient loads, natural microtopography and possibly also open forests with old-growth elements such as dead wood and rot attacks.

The class Montio-Cardaminetea includes vegetation of springs with constant water flow. These habitats, which function as islands for highly specialized and sensitive biota, are endangered by ongoing landscape and climatic changes. Although a harmonized classification into vegetation units is necessary for effective habitat conservation, there is currently no synthetic classification of the class Montio-Cardaminetea. Here a large set of vegetation-plot records was obtained from national and private databases. The aim was to validate the EuroVegChecklist classification scheme, search for additional ecologically meaningful vegetation types and develop an automatic system for classifying new plots from Europe. We formally defined the cores of eight of the ten EuroVegChecklist alliances: Caricion remotae, Cratoneurion commutati, Lycopodo europaei-Cratoneurion commutati, Epilobio nutantis-Montion, KoenigioMicrojuncion, Mniobryo-Epilobion hornemanii, Philonotidion seriatae (Cardamino-Montion) and Swertio perennis-Anisothecion squarrosi, which were also reproduced by unsupervised classifications. Both unsupervised and semi-supervised classifications further suggested two alliances not previously recognized in the EuroVegChecklist: Anthelion julaceae (liverwort dominated subalpine to alpine springs in highly oceanic regions in Britain) and Cratoneuro filicini-Calthion laetae (mesotrophic herb-rich subalpine and alpine springs). The unsupervised classifications mainly reflected the base saturation gradient, distinguishing between calcareous and non-calcareous springs. Therefore, it is suggested the order Montio-Cardaminetalia, which is broadly delimited in the EuroVegChecklist, be divided into two separate orders and the following three orders should be distinguished within the class Montio-Cardaminetea: CardaminoChrysosplenietalia (non-calcareous forest springs; Caricion remotae), Cardamino-Cratoneuretalia (calcareous springs; Cratoneurion commutati, Lycopodo europaei-Cratoneurion commutati) and Montio-Cardaminetalia (non-calcareous non-forest springs; all other alliances).