Emerging ecological functions and relationships of Alpine ungulates amid wildlife recovery and anthropogenic influences: A multi-disciplinary approach

Human-induced environmental changes are placing unprecedented pressures on ecosystems worldwide, driving biodiversity losses that threaten ecosystem functioning and resilience. Among the most critical yet understudied consequences of these losses is the disruption or rapid modification of ecological interactions among species. However, investigating such interactions requires disentangling multiple co-occurring and interrelated factors within complex pathways that often cannot be addressed with a single approach. In this PhD dissertation, I explore how emerging ecological interactions and functions shape a large herbivore community in the Alps — including red deer (Cervus elaphus), roe deer (Capreolus capreolus), and Alpine chamois (Rupicapra rupicapra) — amidst the ongoing natural recolonisation by wolves (Canis lupus) and decades of diverse anthropogenic influences, such as hunting and outdoor recreation. To address these complexities, I designed a comprehensive field study using a quasi-experimental approach along a temporal gradient since wolf recolonisation across 350 km² in the Italian Alps. This study combines complementary techniques — including camera-trapping, animal movement tracking, stable isotope analysis, genetic diet metabarcoding, and gut microbiota metataxonomics — each offering unique but interconnected insights into ecological processes. Together, these methods reveal patterns spanning spatial and temporal scales, biological organisation levels (e.g. population, community, ecosystem), ecological pathways (e.g. horizontal and vertical, top-down and bottom-up) and biological mechanisms (e.g. trophic, behavioural). The dissertation opens with an overview of the study context, objectives and thesis structure. Particular attention is then given to field sampling, presenting a synthetic version of the study design (detailed in the Field Protocol, Appendix A) and the descriptive results. The core of the thesis follows in Chapters 1 through 5, each exploring specific ecological dimensions and analytical perspectives. Chapters 1 and 2 focus on vertical functions of ungulates, examining their antipredator behaviours in response to wolves and human hunters, while also considering the influences of vegetation and topography. Chapters 3 and 4 delve into horizontal relationships, analysing environmental, behavioural and trophic mechanisms of herbivore coexistence, considering isotopic niche hypervolumes, multi-species co-occurrence patterns and intraspecific grouping behaviour. Chapter 5 outlines ongoing analyses and preliminary results expanding on these findings. Finally, all these results are synthesised and discussed in an integrated framework. In Chapter 1, I examine diel activity responses in a community of sympatric ungulates along the ongoing wolf recolonisation in the Central-Eastern Alps, considering the pervasive presence of human activity. Camera-trap data from 158 locations across four sites representing a temporal gradient of wolf recolonisation (with reproductive pairs established in 2017, 2019, 2021 and 2025) were used to analyse ungulate activity during two six-month periods (May–October, 2022–2023). In summer, red deer showed higher diurnality in sites with a longer history of wolf presence (a 7% increase over five years, on average) and progressively reduced nocturnality within sites as local wolf establishment advanced (a 5% decrease per year, on average), also heightening activity overlap with humans. This ‘diel shield effect’ disappeared when human hunting occurred. In contrast, roe deer adjusted diel activity only in response to hunting, while chamois responded solely to the spatial distribution of outdoor recreation. These findings highlight that wolf recovery can induce detectable diel activity shifts in large herbivores over relatively short timescales, yet responses depend on species biology and behavioural plasticity. Importantly, human risk and disturbance can offset or override these behavioural adjustments, potentially altering the ecosystemic effects of returning large carnivores. In Chapter 2, I explore the scale-specific mechanisms of emerging antipredator behaviours by examining how proaction and reaction interact to shape the movements of ungulates in response to hunter-related lethal risk. Using GPS-collared red deer and high-resolution hunting data as a case study, I applied movement metric models and integrated step selection functions that incorporated high-resolution canopy cover data (LiDAR) and individual site familiarity (derived from movement-based utilisation distributions), within a spatiotemporally dynamic landscape of risk. Red deer either proactively avoided areas of chronic risk, or they selected canopy cover where and when risk was predictably high. However, when risk was encountered anyway, canopy cover was no longer selected, and instead only modulated a reactive response along a remain-to-leave continuum. This reaction was even more evident when the environment was unfamiliar, underlining the importance of memory in such reaction patterns. I describe how proaction and reaction fuse in an antipredator sequence of interconnected movement decisions in a large herbivore, and discuss how this result may help disentangle the ecological consequences of behavioural responses to predation. Moving toward horizontal relationships within the mammal community, Chapter 3 examines the niche-based mechanisms enabling coexistence among sympatric ungulates and the functional community structure, using stable isotope ratios (δ13C, δ15N, δ34S, δ18O, δ2H) in hair. Understanding these mechanisms and structure is essential to anticipate how ecological communities may respond to ongoing environmental change and shifting population dynamics. I applied a Bayesian framework to quantify species-specific n-dimensional niche hypervolumes, and estimate niche region and overlaps, alongside univariate and multivariate analyses to detect niche segregation among sympatric ungulates. Red deer exhibited the broadest niche, consistent with their generalist ecology and large spatial range. Results indicated distinct niche segregation among red deer, roe deer, and chamois, with mean pairwise isotopic niche overlap under 40%. Three primary axes of differentiation emerged: water sourcing (δ18O, δ2H), diet quality (δ15N), and habitat openness (δ13C). Specifically, chamois appeared to derive more water from plant intake rather than direct drinking, consume a higher-quality diet than Cervids, and rely more heavily on open habitats for resource use compared to red deer. Additional isotopic differences between red deer and roe deer may stem from fine-scale abiotic conditions like microclimate and topography. Together, these patterns suggest functional differentiation critical for promoting coexistence in this ungulate community. In Chapter 4, I investigate patterns of co-occurrence among sympatric ungulates, as well as intraspecific grouping behaviour, to highlight behavioural dimensions of coexistence under rapid change. Using camera-trap data, I tested how environmental, predator, and anthropogenic factors shape ungulate co-occurrence and grouping during summer. Overall, red deer, roe deer and chamois neither actively avoided nor preferentially associated with one another. However, with advancing wolf recolonisation, roe deer increasingly occupied sites where red deer were absent, indicating potential unidirectional avoidance. Moreover, red deer formed larger groups as time since wolf establishment increased, particularly in open habitats — a likely response to perceived predation risk. Anthropogenic pressures did not significantly influence interspecific co-occurrence but modulated intraspecific grouping behaviour through species- and habitat-specific responses: red deer formed larger groups near human settlements and under higher recreational activity in forests, whereas chamois grouped more in forested areas close to settlements but less where recreational use was high. Notably, both chamois and red deer individually preferred steep terrain, yet avoided both conspecifics and heterospecifics under such conditions, revealing shared environmental constraints influencing both inter- and intraspecific interactions. Chapter 5 outlines ongoing genetic and isotopic analyses aimed at uncovering mechanisms behind observed niche segregation. These include studies of dietary composition through DNA metabarcoding, the role of gut microbiota in digestion and host-environment interactions, nutrient pathways, spatial foraging strategies, and isotopic calibration via a controlled experiment. Overall, this dissertation advances our understanding of the structure of ecological networks involving Alpine large mammals and how these networks reorganise under the combined pressures of carnivore recolonisation and pervasive human influence. By integrating behavioural, spatial, and trophic data across horizontal and vertical ecological processes, my findings show that simultaneous ecological differentiation and overlap support coexistence in this community. At the same time, ungulate species adjust their behaviours, and intra- and interspecific associations in response to emerging risks, reshaping the network structure through behavioural dynamics. This multi-disciplinary approach underscores that ecosystems cannot be understood through simplified, single-species perspectives, highlighting the crucial role of both predators and ungulates as active agents shaping ecological networks. Such insights contribute to a deeper understanding of species coexistence and future ecosystem dynamics in increasingly human-dominated landscapes.