The European Alps, a biodiversity hotspot, are facing threats from climate change and human land use alterations. Alpine vegetation is naturally structured by elevational belts, determined by temperature, moisture, and topography. However, human influence on these landscapes spans millennia, interacting with natural processes to shape biodiversity. Climate change projections predict upward vegetation displacement, reducing alpine species habitats, while land abandonment has led to increased forest cover and the loss of meadow species. Lake sediments offer a valuable record of vegetation changes over long timescales, providing crucial insights for anticipating future responses to global change. Traditional palaeoecological methods, such as pollen analysis, have limitations in taxonomic resolution and representation of certain plant groups (e.g., forbs and graminoids). Sedimentary ancient DNA (sedaDNA) offers a higher resolution approach to studying past vegetation dynamics. This study uses a multi-proxy approach to investigate the drivers of alpine plant diversity in the European Alps over the Holocene, focusing on the interplay between climate change and human activities.

Previous research has highlighted the impact of climate change and human activities on alpine vegetation. Studies have shown that climate change is causing an upward shift in vegetation zones, threatening high-altitude species. Land-use changes, particularly land abandonment in recent decades, have resulted in increased forest cover and decreased meadow species richness. Pollen analysis has been used extensively to reconstruct past vegetation, but its limitations in resolving some taxa, especially forbs, has motivated the use of sedaDNA. Earlier studies using sedaDNA have shown promise in revealing human impacts on vegetation, but many have relied on incomplete regional or global databases, limiting taxonomic resolution. The development of comprehensive regional databases, such as PhyloAlps in this study, greatly improves the accuracy and detail of past vegetation reconstructions from sedaDNA.

This study utilizes a multi-proxy approach to reconstruct Holocene vegetation at Lake Sulsseewli in the Swiss Alps. The methods include: 1. **Sediment Core Analysis:** A composite sediment core was constructed, providing a long sedimentary sequence spanning the Holocene. Lithology, organic matter content (LOI), and radiocarbon dating were used to establish the core chronology. 2. **Temperature Reconstruction:** Chironomid analysis was used to reconstruct past summer temperatures using a well-established transfer function based on modern chironomid distributions. 3. **Precipitation Reconstruction:** A downscaled precipitation dataset (CHELSA-TraCE21k) was used to estimate past precipitation patterns. 4. **Pollen and Spore Analysis:** Standard palynological techniques were applied to identify pollen and spores, providing information on vegetation composition and changes through time. Microcharcoal analysis was also conducted to assess fire activity. 5. **sedaDNA Analysis:** DNA was extracted from sediment samples, amplified using primers targeting the *trn*L P6 loop (plants) and a section of the mammalian mitochondrial 16S locus. The PhyloAlps database, a comprehensive Alpine plant DNA reference library, was used for taxonomic assignments. A relative abundance index (RAI) was developed to account for both read abundance and PCR replicability. Mammalian sedaDNA was analyzed similarly to assess changes in animal populations. 6. **Statistical Analyses:** Redundancy analysis (RDA) was used to investigate the relationships between plant community composition and environmental variables (climate and human activity). Generalized additive models (GAMs) explored the non-linear relationships between plant richness and key drivers. A structural equation model (SEM) examined the direct and indirect effects of different factors on plant richness. Constrained incremental sum of squares (CONISS) analysis was used to define vegetation zones.

The study generated a sedaDNA record of exceptional richness (366 plant taxa) from Lake Sulsseewli. Key findings include: 1. **Early Holocene (11-6 ka):** Vegetation was primarily driven by climate, with cold-adapted alpine species dominating during cooler periods and a subsequent upward migration of subalpine taxa during warming. Low human impact was observed. 2. **Mid-Late Holocene (6 ka onwards):** Increasing evidence of human impact is observed. Human activity (grazing and potentially burning) is associated with the expansion of indicators for human activity (e.g., coprophilous fungal spores, pastoral plants). Land-use shifted to agropastoralism in the Middle Ages with substantial deforestation. 3. **Plant Richness:** The highest plant richness occurred during the mid-late Holocene, coinciding with human impact. The RDA analysis revealed that both climate and human activity explain a significant portion of the variance in plant composition and richness. GAMs and SEMs confirmed the significance of precipitation and human-related factors (e.g., sheep, goat, charcoal) in shaping plant richness. A hump-shaped relationship was observed between sheep RAI and plant richness, suggesting that intermediate levels of grazing promote diversity. 4. **Elevational Shifts:** The study shows an increased coexistence of plant species from different elevational belts, likely due to human-induced habitat changes. This increased species richness did not correspond to a similar increase in pollen diversity, highlighting the limitations of pollen analysis in capturing the full extent of herbaceous plant diversity.

The findings support the hypothesis that the high plant diversity in the European Alps is a product of the long history of human land use. The study shows the significant roles that both climate and human activity have played in shaping alpine vegetation dynamics, particularly during the mid-late Holocene when human impact intensified. The hump-shaped relationship between grazing intensity and plant richness is consistent with the intermediate disturbance hypothesis. The detailed sedaDNA record reveals a hidden diversity previously undetected by pollen analysis, emphasizing the value of sedaDNA in reconstructing past ecosystems. While the study focused on plant diversity, future research should explore the effects of different grazing intensities and land-use practices on other components of alpine biodiversity (e.g., mammals, insects).

This study provides the most detailed palaeoecological reconstruction of Holocene alpine vegetation to date, revealing the complex interplay between climate and human activities in shaping alpine plant diversity. The high plant richness observed in the Mid-Late Holocene is strongly linked to human land use, specifically low-intensity agropastoralism. Maintaining moderate levels of human management may be crucial for preserving the unique plant diversity of alpine ecosystems in the face of climate change. Future research could investigate the impacts of varying grazing intensities and land-use practices on biodiversity across different elevational belts and moisture regimes.

The study focuses on a single lake, limiting the generalizability of the findings to the wider Alpine region. While the PhyloAlps database greatly improved taxonomic resolution, potential limitations remain due to incomplete species representation or haplotype sharing. Dating uncertainties, particularly in the earliest Holocene, can influence the interpretation of some results. The interpretation of human impacts is primarily inferred from indirect indicators, and direct evidence of human activity might be necessary for stronger conclusions.