Being in the right place at the right time: wolves spatio-temporal niche in a human-altered environment

Large carnivores are among the most controversial and challenging group of species to manage and conserve, because of the conflict they raise with human interests. As a group, large carnivores exert a strong influence on biological communities via predation and inter-specific competition, as well as by limiting and often regulating the numbers of their prey. At the same time, there is a strong hostility to these species in human history and culture, because of perceptions of their negative impacts on human livelihoods. Human-carnivore conflict mainly arises because carnivores' protein-rich diet and large home ranges draw them into recurrent competition with humans for food and space. As this competition over food and space show no sign of reduction, an intuitive forecast could be that large carnivores will persist only in protected areas or in some remote and uninhabited wilderness areas. The alternative scenario follows a landscape-scale conservation approach that aims at human-carnivore coexistence in a shared environment through conflict mitigation programs. This approach seemed to be realistic in Europe, that is succeeding in maintaining, and to some extent restoring, viable large carnivore populations on a continental scale. This has recently been reported for four large carnivore species, e.g., wolf (Canis lupus), brown bear (Ursus arctos), wolverine (Gulo gulo) and lynx (Lynx lynx), that are persisting in human-dominated landscapes and largely outside protected areas in the European continent (Chapron et al. 2014). Among these species, the gray wolf proved to be the most adaptable to live in proximity to humans, and therefore represents a good model species to investigate the questions on human-carnivore coexistence. In my PhD thesis, I considered a specific case study of coexistence between humans and the gray wolf. I focused on the wolf population of the Abruzzo, Lazio and Molise National Park (PNALM), located in the central Apennines (Italy), and representing one of the few historical strongholds of the Italian wolf population. Wolves in the PNALM therefore constitutes one of the source populations that made possible wolf recolonization throughout the Italian peninsula in the last decades. The PNALM features a landscape that can be considered relatively pristine compared to other parts of Italy it but is still affected by multiple human activities such as livestock grazing, tourism and logging. Despite their protected status and a long co-existence history with humans, wolves are still illegally killed in the park, and show higher levels of stress compared to other protected wolf populations, possibly due to a multitude of direct and indirect anthropogenic effects such as strong presence of sympatric free-ranging dogs. From a conservation perspective, it is important to understand how wolves have adapted their spatio-temporal niche to live and persist in proximity to humans. The general working hypothesis of my PhD research was to find a partitioning in the wolf spatio-temporal niche leading to a complementary presence and activity of wolves with respect to those of humans in the central Apennines. To address my research questions, I used a Global Positioning System (GPS) data bank, consisting of locations collected from 11 wolves (6 females and 5 males) belonging to 5 packs in the PNALM. In Chapter I, I investigated wolf spatial niche by means of home range analysis. Whereas home range behaviour in wolves has been extensively studied throughout their range, increasing attention has been devoted in Europe to the anthropogenic impact on wolf spatial behaviour. However, such studies are still particularly rare in Southern Europe where, due to a relatively higher human density and impact on the landscape, conditions are ideal to investigate spatial and habitat correlates of behavioural responses by wolves. I used state-of-the-art techniques for home range and core areas estimation, i.e., Brownian Bridge Movement Models and time-maximizing functions based on exponential regression curves, respectively. Then, I used Generalized Linear Mixed Models (GLMMs) to assess seasonal, circadian and anthropogenic effects on wolf home range size, and I used a modified version of the Ecological Niche Factor Analysis to describe the composition of wolf core areas compared to the entire home ranges. I revealed a response by wolves to humans at the home range scale, including larger home ranges detected during night and in areas of higher road density. Thus, human influence on wolf spatial niche manifested itself in two ways. First, larger nocturnal home ranges indicate that wolves in the PNALM display spatial avoidance of higher levels of human activity during daylight hours by restricting their daytime movements. Second, habitat fragmentation caused by roads likely force wolves to increase home range size to reach a minimum threshold of undisturbed area to meet their ecological needs, such as the presence of refuge areas for resting and attending pups. In addition, wolf spatial responses to human presence were habitat-mediated, as core areas were preferentially established at higher elevation and in the more forested portions of the home range. Overall, these landscape characteristics reflect safe places, where human accessibility, and related risk of human-caused mortality, are low. In Chapter II, I assessed wolf habitat selection and behavioural responses towards anthropogenic features by means of Habitat Suitability Models (HSMs). I used Step Selection Functions (SFFs) and Resource Selection Functions (RSFs) to analyze wolf habitat selection patterns along moving trajectories and non-moving clusters, respectively. I focused my analysis particularly on the behavioural responses of wolves to roads, as these anthropogenic features are well known to affect wolf populations in ambivalent ways, posing a trade-off between direct and indirect negative effects (e.g., increased mortality and disturbance), and the positive effects resulting from increased ease of travel, efficient scent marking and access to prey. I found that wolf responses to anthropogenic features mainly varied according to a seasonal effect. In particular, I found that wolves strongly avoided roads during summer, when human disturbance is relatively high in the PNALM due to tourism and livestock raising at high-elevation pastures. On the other hand, wolves were found to select roads during winter, when human activity is low and snow cover may represent an important factor increasing energy expenditure for locomotion; in these conditions, wolves are well known to use roads for ease of travel. In line with previous research, I observed that the spatial segregation adopted by wolves towards anthropogenic features enables them not only to coexist with humans, but also to take advantage of man-made habitat modifications in presence of low human interference. To integrate the analysis of wolf spatial niche with its feeding behaviour, in Chapter III I used data collected from scat samples to describe wolf diet in the PNALM, were livestock represents one of the main economic activities for local people. By using log-linear models, the main goal of wolf diet analysis was to asses to what extent wolves rely on livestock compared to wild ungulates in this human-modified environment. Answering this question may also shed some light on the ecological role played by wolves in the biotic community of the PNALM, which is amongst the richest in the Apennines. I found that, despite a high availability of wild ungulates, wolves consistently consume livestock in this area. This observation is linked to the husbandry methods currently used in the PNALM, that often involve free-ranging herds with little to none supervision. These conditions not only heighten the conflicts with farmers and the subsequent risk of retaliatory killing of wolves due to livestock losses but may also alter wolf ecology and behaviour and interfere with ecological processes. Minimizing livestock accessibility should then be achieved not only to mitigate social conflicts over wolf conservation, but also to restore wolf ecological role as apex predator and, consequently, the integrity of biotic community dynamics. In Chapter IV, I investigated the temporal dimension of the wolf niche by analyzing their activity patterns. Wolf activity patterns in central Apennines have already been investigated in the 80s, when wolves were found to mainly concentrate their activity during night hours. However, at that time wild prey communities were virtually absent in the area, and wolves were forced to mainly rely on anthropogenic food sources such as dumps in proximity of settlements. Nowadays, the ecological context of the area has changed, as wild ungulates became widely available for wolves following roe deer (Capreouls capreolus) and red deer (Cervus elaphus) reintroductions, and wild boar (Sus scrofa) introduction in the 70s. My aim was therefore to assess wolf activity patterns in these renewed ecological conditions. I expressed activity as the average distance travelled by wolves between consecutive GPS locations, and I used these distances to provide a description of wolf activity during the 24-h day. In this way, I first assessed if the renewed presence of wild ungulate populations in the central Apennines produced some detectable changes in wolf temporal niche. Secondly, I used distances travelled by wolves as response variable in Bayesian models to assess the effect of anthropogenic (i.e., roads and settlements) and ecological (i.e., season and social status) factors on wolf activity patterns. I found that wolves showed bimodal activity patterns, with activity peaks around dawn and dusk, that likely reflect wolf selection of crepuscular light for hunting wild ungulates. Therefore, wolf temporal niche in the PNALM currently seems to be shaped by the interaction between human presence and hunting strategies, rather than being mainly determined by the temporal avoidance of human activities. Concurrently, I found no effect of anthropogenic features such as road and settlement densities on wolf activity, whereas I found that this was affected by both season and wolf social association. Overall, my PhD research show that the spatial arrangement, habitat use, activity, and movements of wolves in the human-modified landscape of central Apennines are highly integrated to make the best functional compromise between the necessity to achieve all fundamental biological requirements such as prey and cover, and the need to avoid any direct form of disturbance by humans. Nevertheless, potential limits of my research are represented by the small number of GPS-collared wolves, as well as by the lack of accurate data on prey abundance and distribution in our study area. Moreover, similar studies are needed from less productive and more human-altered ecosystems to fully appreciate the spatio-temporal flexibility of wolves in highly populated countries such as Italy.