Seeking temporal refugia to heat stress: increasing nocturnal activity despite predation risk.

Flexibility in activity timing may enable organisms to quickly adapt to environmental changes. Under global warming, diurnally adapted endotherms may achieve a better energy balance by shifting their activity towards cooler nocturnal hours. However, this shift may expose animals to new or increased environmental challenges (e.g. increased predation risk, reduced foraging efficiency). We analysed a large dataset of activity data from 47 ibex (Capra ibex) in two protected areas, characterized by varying levels of predation risk (presence versus absence of the wolf-Canis lupus). We found that ibex increased nocturnal activity following warmer days and during brighter nights. Despite the considerable sexual dimorphism typical of this species and the consequent different predation-risk perception, males and females demonstrated consistent responses to heat in both predator-present and predator-absent areas. This supports the hypothesis that shifting activity towards nighttime may be a common strategy adopted by diurnal endotherms in response to global warming. As nowadays different pressures are pushing mammals towards nocturnality, our findings emphasize the urgent need to integrate knowledge of temporal behavioural modifications into management and conservation planning.

First evidence of widespread positivity to anticoagulant rodenticides in grey wolves (Canis lupus).

Second-generation Anticoagulant Rodenticides (ARs) can be critical for carnivores, due to their widespread use and impacts. However, although many studies explored the impacts of ARs on small and mesocarnivores, none assessed the extent to which they could contaminate large carnivores in anthropized landscapes. We filled this gap by exploring spatiotemporal trends in grey wolf (Canis lupus) exposure to ARs in central and northern Italy, by subjecting a large sample of dead wolves (n = 186) to the LC-MS/MS method. Most wolves (n = 115/186, 61.8 %) tested positive for ARs (1 compound, n = 36; 2 compounds, n = 47; 3 compounds, n = 16; 4 or more compounds, n = 16). Bromadiolone, brodifacoum and difenacoum, were the most common compounds, with brodifacoum and bromadiolone being the ARs that co-occurred the most (n = 61). Both the probability of testing positive for multiple ARs and the concentration of brodifacoum, and bromadiolone in the liver, systematically increased in wolves that were found at more anthropized sites. Moreover, wolves became more likely to test positive for ARs through time, particularly after 2020. Our results underline that rodent control, based on ARs, increases the risks of unintentional poisoning of non-target wildlife. However, this risk does not only involve small and mesocarnivores, but also large carnivores at the top of the food chain, such as wolves. Therefore, rodent control is adding one further conservation threat to endangered large carnivores in anthropized landscapes of Europe, whose severity could increase over time and be far higher than previously thought. Large-scale monitoring schemes for ARs in European large carnivores should be devised as soon as possible.

Estimating Wolf Population Size and Dynamics by Field Monitoring and Demographic Models: Implications for Management and Conservation.

We estimated the current size and dynamics of the wolf population in Tuscany and investigated the trends and demographic drivers of population changes. Estimates were obtained by two different approaches: (i) mixed-technique field monitoring (from 2014 to 2016) that found the minimum observed pack number and estimated population size, and (ii) an individual-based model (run by Vortex software v. 10.3.8.0) with demographic inputs derived from a local intensive study area and historic data on population size. Field monitoring showed a minimum population size of 558 wolves (SE = 12.005) in 2016, with a density of 2.74 individuals/100 km2. The population model described an increasing trend with an average annual rate of increase λ = 1.075 (SE = 0.014), an estimated population size of about 882 individuals (SE = 9.397) in 2016, and a density of 4.29 wolves/100 km2. Previously published estimates of wolf population were as low as 56.2% compared to our field monitoring estimation and 34.6% in comparison to our model estimation. We conducted sensitivity tests to analyze the key parameters driving population changes based on juvenile and adult mortality rates, female breeding success, and litter size. Mortality rates played a major role in determining intrinsic growth rate changes, with adult mortality accounting for 62.5% of the total variance explained by the four parameters. Juvenile mortality was responsible for 35.8% of the variance, while female breeding success and litter size had weak or negligible effects. We concluded that reliable estimates of population abundance and a deeper understanding of the role of different demographic parameters in determining population dynamics are crucial to define and carry out appropriate conservation and management strategies to address human-wildlife conflicts.

Laterality in roe deer embryos implantation.

Female reproductive success is one of the most important life-history traits to be monitored when determining population dynamics in free-ranging ungulates. Several studies have described how phenotypic characteristics of the mother, climatic conditions, population status, and habitat can impact on potential reproductive output in wild ungulates. However, little is known regarding the internal, physiological factors, that may account for differences in implantation rates. The present study investigated the differences in implantation rates and site on the basis of site and number of ovulations through the examination of about 3000 intact uteri collected from pregnant roe deer does (Capreolus capreolus). Although ovulation occurs with the same frequency in the left and right ovary, we revealed a higher frequency of embryos implantation in the left uterine horn in odd litter size, demonstrating that embryos can migrate between the uterine horns. In our study, a greater proportion of reproductive wastage was associated to females with three and four corpora lutea and interestingly, in relation to the site of ovulation, the percentage of corpora lutea that did not correspond to a fetus was higher in the right ovary than in the left one (73.2% vs. 26.8%). Our research described for the first time the absence of laterality in ovulation and the presence of laterality in implantation in roe deer, thus laying the foundations for in-depth studies about the functionality of this uterine side and for comparisons with populations located in other geographical areas to understand whether it is a widespread phenomenon or a local adaptation.

A human-neutral large carnivore? No patterns in the body mass of gray wolves across a gradient of anthropization.

The gray wolf (Canis lupus) expanded its distribution in Europe over the last few decades. To better understand the extent to which wolves could re-occupy their historical range, it is important to test if anthropization can affect their fitness-related traits. After having accounted for ecologically relevant confounders, we assessed how anthropization influenced i) the growth of wolves during their first year of age (n = 53), ii) sexual dimorphism between male and female adult wolves (n = 121), in a sample of individuals that had been found dead in Italy between 1999 and 2021. Wolves in anthropized areas have a smaller overall variation in their body mass, during their first year of age. Because they already have slightly higher body weight at 3-5 months, possibly due to the availability of human-derived food sources. The difference in the body weight of adult females and males slightly increases with anthropization. However, this happens because of an increase in the body mass of males only, possibly due to sex-specific differences in dispersal and/or to "dispersal phenotypes". Anthropization in Italy does not seem to have any clear, nor large, effect on the body mass of wolves. As body mass is in turn linked to important processes, like survival and reproduction, our findings indicates that wolves could potentially re-occupy most of their historical range in Europe, as anthropized landscapes do not seem to constrain such of an important life-history trait. Wolf management could therefore be needed across vast spatial scales and in anthropized areas prone to social conflicts.

Insights on the best release strategy from post-release movements and mortality patterns in an avian scavenger.

Conservation translocations involving vultures rely either on soft- or hard-release strategies. To investigate whether these strategies affect home range stability and survival, we compared the spatial behavior and mortality of 38 Griffon vultures (Gyps fulvus) released in Sardinia. Griffons were released after no acclimatization or after 3 (short) or 15 (long acclimatization) months in an aviary. In the two years that followed their release, griffons without acclimatization did not stabilize their home range size, while those subjected to long acclimatization stabilized it in the second year. Short-acclimatized griffons always had a large home range, soon after their release. The number of individuals that reached sexual maturity was higher (71.4%) in long-acclimatized griffons than in short-acclimatized ones (40%) or in griffons that were hard released (28.6%). Soft release with a long acclimatization period seems to be the most successful method to ensure stable home ranges and the survival of griffon vultures.

Multiple aspects of the maternal reproductive investment in a polytocous species: What do mothers really control?

One of the factors facilitating the expansion and proliferation of wild boar Sus scrofa is the plasticity of its reproductive biology. Nevertheless, the real influence of maternal and environmental factors on number and sex of the offspring is still controversial. While the litter size was shown to be related with the maternal condition, the strength of this relation remains to be understood, together with the possible role played by environmental conditions. Analogously, it is unclear whether wild boar females can adjust their offspring sex. We investigated multiple aspects of wild boar maternal investment by means of a 10-year dataset of female reproductive traits and a set of biologically meaningful environmental variables. The maternal condition slightly affected the litter size but not the offspring sex, and environment did not affect the litter size or the offspring sex. Moreover, mothers did not cope with the higher costs entailed by producing sons by placing them in the most advantageous intrauterine position, nor by allocating less resources on daughters. Our set of results showed that the female reproductive investment is quite rigid in comparison with other aspects of wild boar reproductive biology. Wild boar females seem to adopt a typical r-strategy, producing constantly large litters and allocating resources on both sexes regardless of internal and external conditions. Such strategy may be adaptive to cope with environmental unpredictability and an intense human harvest, contributing to explain the extreme success of wild boar within human-dominated landscapes.

A Predator on the Doorstep: Kill Site Selection by a Lone Wolf in a Peri-Urban Park in a Mediterranean Area.

The aim of the study was to assess which kill site characteristics were selected by a lone wolf living in a protected Mediterranean coastal area near the city of Pisa, Italy, where both wild and domestic ungulates were available as potential prey. Between 2017 and 2019, we monitored the wolf's predatory behaviour through a combination of camera trapping and active search for kill sites and prey carcasses. The main prey found was the fallow deer (n = 82); only two wild boars and no domestic ungulates were found preyed upon. The features and habitat of kill sites were modelled to test for selection by the wolf. The habitat type of kill site was composed of meadows and pastures (89.3%), woods (7.3%), degraded coastal areas (1.9%), roads and rivers (1.1%), and marshes (0.5%). We calculated their distance from landscape features and ran a binomial generalised linear model to test the influence of such landscape variables. The distance of kill sites from landscape elements was significantly different from random control sites, and a positive selection for fences was found. In fact, the wolf pushed fallow deer towards a fence to constrain them and prevent them from escaping. We also analysed the body condition of predated fallow deer as a percentage of fat content in the bone marrow of the hind legs. Our results revealed the selection of the lone wolf for deer in good body condition. This is a possible outcome of the habitat selection shown by fallow deer in the study area, where fenced open pastures are the richest in trophic resources; therefore, our findings suggest a high efficacy for the lone wolf hunting strategy, but also the adoption of a high risk feeding strategy by deer. This study suggests that a lone predator can take advantage of human infrastructures to maximise its predatory effectiveness.

It is time to mate: population-level plasticity of wild boar reproductive timing and synchrony in a changing environment.

On a population level, individual plasticity in reproductive phenology can provoke either anticipations or delays in the average reproductive timing in response to environmental changes. However, a rigid reliance on photoperiodism can constraint such plastic responses in populations inhabiting temperate latitudes. The regulation of breeding season length may represent a further tool for populations facing changing environments. Nonetheless, this skill was reported only for equatorial, nonphotoperiodic populations. Our goal was to evaluate whether species living in temperate regions and relying on photoperiodism to trigger their reproduction may also be able to regulate breeding season length. During 10 years, we collected 2,500 female reproductive traits of a mammal model species (wild boar Sus scrofa) and applied a novel analytical approach to reproductive patterns in order to observe population-level variations of reproductive timing and synchrony under different weather and resources availability conditions. Under favorable conditions, breeding seasons were anticipated and population synchrony increased (i.e., shorter breeding seasons). Conversely, poor conditions induced delayed and less synchronous (i.e., longer) breeding seasons. The potential to regulate breeding season length depending on environmental conditions may entail a high resilience of the population reproductive patterns against environmental changes, as highlighted by the fact that almost all mature females were reproductive every year.

Main roads and land cover shaped the genetic structure of a Mediterranean island wild boar population.

Patterns of genetic differentiation within and among animal populations might vary due to the simple effect of distance or landscape features hindering gene flow. An assessment of how landscape connectivity affects gene flow can help guide management, especially in fragmented landscapes. Our objective was to analyze population genetic structure and landscape genetics of the native wild boar (Sus scrofa meridionalis) population inhabiting the island of Sardinia (Italy), and test for the existence of Isolation-by-Distance (IBD), Isolation-by-Barrier (IBB), and Isolation-by-Resistance (IBR). A total of 393 Sardinian wild boar samples were analyzed using a set of 16 microsatellite loci. Signals of genetic introgression from introduced non-native wild boars or from domestic pigs were revealed by a Bayesian cluster analysis including 250 reference individuals belonging to European wild populations and domestic breeds. After removal of introgressed individuals, genetic structure in the population was investigated by different statistical approaches, supporting a partition into five discrete subpopulations, corresponding to five geographic areas on the island: north-west (NW), central west (CW), south-west (SW), north-central east (NCE), and south-east (SE). To test the IBD, IBB, and IBR hypotheses, we optimized resistance surfaces using genetic algorithms and linear mixed-effects models with a maximum likelihood population effects parameterization. Landscape genetics analyses revealed that genetic discontinuities between subpopulations can be explained by landscape elements, suggesting that main roads, urban settings, and intensively cultivated areas are hampering gene flow (and thus individual movements) within the Sardinian wild boar population. Our results reveal how human-transformed landscapes can affect genetic connectivity even in a large-sized and highly mobile mammal such as the wild boar, and provide crucial information to manage the spread of pathogens, including the African Swine Fever virus, endemic in Sardinia.

Dentition and body condition: tooth wear as a correlate of weight loss in roe deer.

BACKGROUND: In many mammalian species, once the permanent teeth have erupted, the only change to dentition is a gradual loss of tooth surface/height through wear. The crown of the teeth cannot be repaired once worn. When dental crown tissue has been depleted due to wear, the animal is expected to have a suboptimal body condition. We evaluated the role of tooth wear in causing a reduction of physical condition in adult roe deer females (Capreolus capreolus). RESULTS: The progressive wearing of the lower cheek teeth was assessed in a Northern Apennines (Italy) population with a new scoring scheme based on objectively described tooth characteristics (morphotypes) being either present or absent. Eviscerated body mass and mandible length, which is a good proxy for body size in roe deer, were related to the tooth wear score by the use of linear regressions. The sum of wear scores for molariform teeth correlated most strongly with body condition (i.e., eviscerated body mass/mandible length), showing the importance of the entire chewing surface for acquiring energy by food comminution, chewing, and digestion. In comparison with individuals of comparable size experiencing minor tooth wear, the body mass of those with the most advanced stage of tooth wear was decreased by 33.7%. This method was compared to the height and the hypsodonty index of the first molar, the most commonly used indices of tooth wear. The sum of molariform wear scoring scheme resulted in a more suitable index to describe the variation in body condition of roe deer. CONCLUSIONS: Describing tooth wear patterns in hunted populations and monitoring at which tooth wear level (and therefore dental morphotype) an animal is no longer able to sustain its physical condition (i.e. when it begins to lose body mass) can be a useful tool for improving the management of the most widespread and abundant deer species in Europe. At the same time, such an approach can clarify the role of tooth wear as a proximate cause of senescence in ungulates.

Anthropogenic and Environmental Factors Determining Local Favourable Conditions for Wolves during the Cold Season.

Winter resources are crucial for wildlife, and, at a local scale, some anthropogenic and environmental factors could affect their availability. In the case of wolves, it is known that vocalisations in response to unfamiliar howls are issued to defend their territory and the important resources within it. Then, we studied the characteristics of winter response sites (WRS) during the cold season, aiming to assess their eventual ability to provide insights into the distribution of valuable resources within their territories. Within this scope, we planned a wolf-howling survey following a standardised approach. The study covered an Apennine (Central Italy) area of 500 km2. A hexagonal mesh was imposed on the area, in order to determine the values of different variables at the local scale. A logistic LASSO regression was performed. WRS were positively related to the presence of thermal refuges (odds = 114.485), to patch richness (odds = 1.153), wild boar drive hunting areas (odds = 1.015), and time elapsed since the last hunt (odds = 1.019). Among negative factors, stray dogs reply considerably affects wolves' responsiveness (odds = 0.207), where odds are the exponentiated coefficients estimated by the logistic lasso regression. These results suggest that WRS are related to anthropogenic and environmental factors favouring the predation process.

Capital-income breeding in wild boar: a comparison between two sexes.

Organisms differ in the strategy adopted to fuel reproduction by using resources either previously acquired and stored in body reserves (capital breeding) or, conversely, acquired during their reproductive activity (income breeding). The choice of one or the other strategy is related to several internal and external factors which are counteractive in wild boar. Based on a large dataset of culled wild boar, we investigated individual body weight variability throughout the period of 1st September-31st January, which included the main part of the mating season, among different sex and age classes to determine their position along the capital-income breeding continuum. Though food resources were abundant during the rut, adult males lost body weight suggesting they adopted a predominantly capital breeding strategy, likely owing to the high intra-sexual competition entailed by the peculiar mating system of the species. On the contrary, subadult males seemed to behave as income breeders, likely enhancing the reproductive flexibility of wild boar populations. During the rut, females stored reserves, thus suggesting that they substantially relied on them to cover future reproductive costs.

Ungulate management in European national parks: Why a more integrated European policy is needed.

1. Primary objectives of national parks usually include both, the protection of natural processes and species conservation. When these objectives conflict, as occurs because of the cascading effects of large mammals (i.e., ungulates and large carnivores) on lower trophic levels, park managers have to decide upon the appropriate management while considering various local circumstances. 2. To analyse if ungulate management strategies are in accordance with the objectives defined for protected areas, we assessed the current status of ungulate management across European national parks using the naturalness concept and identified the variables that influence the management. 3. We collected data on ungulate management from 209 European national parks in 29 countries by means of a large-scale questionnaire survey. Ungulate management in the parks was compared by creating two naturalness scores. The first score reflects ungulate and large carnivore species compositions, and the second evaluates human intervention on ungulate populations. We then tested whether the two naturalness score categories are influenced by the management objectives, park size, years since establishment, percentage of government-owned land, and human impact on the environment (human influence index) using two generalized additive mixed models. 4. In 67.9% of the national parks, wildlife is regulated by culling (40.2%) or hunting (10.5%) or both (17.2%). Artificial feeding occurred in 81.3% of the national parks and only 28.5% of the national parks had a non-intervention zone covering at least 75% of the area. Furthermore, ungulate management differed greatly among the different countries, likely because of differences in hunting traditions and cultural and political backgrounds. Ungulate management was also influenced by park size, human impact on the landscape, and national park objectives, but after removing these variables from the full model the reduced models only showed a small change in the deviance explained. In areas with higher anthropogenic pressure, wildlife diversity tended to be lower and a higher number of domesticated species tended to be present. Human intervention (culling and artificial feeding) was lower in smaller national parks and when park objectives followed those set by the International Union for the Conservation of Nature (IUCN). 5. Our study shows that many European national parks do not fulfil the aims of protected area management as set by IUCN guidelines. In contrast to the USA and Canada, Europe currently has no common ungulate management policy within national parks. This lack of a common policy together with differences in species composition, hunting traditions, and cultural or political context has led to differences in ungulate management among European countries. To fulfil the aims and objectives of national parks and to develop ungulate management strategies further, we highlight the importance of creating a more integrated European ungulate management policy to meet the aims of national parks.

A standardized approach to empirically define reliable assignment thresholds and appropriate management categories in deeply introgressed populations.

Anthropogenic hybridization is recognized as a major threat to the long-term survival of natural populations. While identifying F1 hybrids might be simple, the detection of older admixed individuals is far from trivial and it is still debated whether they should be targets of management. Examples of anthropogenic hybridization have been described between wolves and domestic dogs, with numerous cases detected in the Italian wolf population. After selecting appropriate wild and domestic reference populations, we used empirical and simulated 39-autosomal microsatellite genotypes, Bayesian assignment and performance analyses to develop a workflow to detect different levels of wolf x dog admixture. Membership proportions to the wild cluster (qiw) and performance indexes identified two q-thresholds which allowed to efficiently classify the analysed genotypes into three assignment classes: pure (with no or negligible domestic ancestry), older admixed (with a marginal domestic ancestry) and recent admixed (with a clearly detectable domestic ancestry) animals. Based on their potential to spread domestic variants, such classes were used to define three corresponding management categories: operational pure, introgressed and operational hybrid individuals. Our multiple-criteria approach can help wildlife managers and decision makers in more efficiently targeting the available resources for the long-term conservation of species threatened by anthropogenic hybridization.

Effects of forest management and roe deer impact on a mountain forest development in the Italian Apennines: A modelling approach using LANDIS-II.

Forest development is a complex phenomenon which, for the number of actors involved and the response time expressed by forests, is difficult to understand and explore. Forests in Italy, as in several areas of Europe, are experiencing intensive management and recently, an increasing impact by ungulates. The effects on forest development of these two disturbances combined are difficult to predict, and consequently to be properly managed. We used a forest landscape change model, LANDIS-II, to simulate forest development as driven by forestry practices and roe deer impact for 200 years in a mountain forest of the Italian Apennines. We found that each disturbance alters forest tree species richness, forest type abundance and distribution, and forest structure. When considered combined, the two disturbances show additive behavior, enhancing or moderating each other's effects. Forest management has a negative effect on tree species richness. We expected roe deer to have a negative effect on harvest yields, but this result was significant only for two of seven harvesting treatments. On the other hand, roe deer presence had a positive effect on tree species richness. All the simulation scenarios returned some extent of forest loss. The amount of the forest loss is lowest in the scenario without disturbances, and greatest when both disturbances are considered. However, the two disturbances combined, with the magnitude modelled in our simulations, have relatively low effects on the forest dynamics we analyzed in our study area. LANDIS-II was an effective approach for simulating combined management and ungulate driven trends of forest development, and to help understand the dynamics that lay behind it.

Assessing precision and requirements of three methods to estimate roe deer density.

Roe deer (Capreolus capreolus) is the most abundant cervid in Europe and, as such, has a considerable impact over several human activities. Accurate roe deer population size estimates are useful to ensure their proper management. We tested 3 methods for estimating roe deer abundance (drive counts, pellet-group counts, and camera trapping) during two consecutive years (2012 and 2013) in the Apennines (Italy) in order to assess their precision and applicability. During the study period, population density estimates were: drive counts 21.89±12.74 roe deer/km2 and pellet-group counts 18.74±2.31 roe deer/km2 in 2012; drive counts 19.32±11.12 roe deer/km2 and camera trapping 29.05±7.48 roe deer/km2 in 2013. Precision of the density estimates differed widely among the 3 methods, with coefficients of variation ranging from 12% (pellet-group counts) to 58% (drive counts). Drive counts represented the most demanding method on account of the higher number of operators involved. Pellet-group counts yielded the most precise results and required a smaller number of operators, though the sampling effort was considerable. When compared to the other two methods, camera trapping resulted in an intermediate level of precision and required the lowest sampling effort. We also discussed field protocols of each method, considering that volunteers, rather than technicians, will more likely be appointed for these tasks in the near future. For this reason, we strongly suggest that for each method managers of population density monitoring projects take into account ease of use as well as the quality of the results obtained and the resources required.

Ecological conditions experienced by offspring during pregnancy and early post-natal life determine mandible size in roe deer.

Population dynamics studies and harvesting strategies often take advantage of body size measurements. Selected elements of the skeletal system such as mandibles, are often used as retrospective indices to describe body size. The variation in mandibular measurements reflects the variation in the ecological context and hence the variation in animal performance. We investigated the length of the anterior and posterior sections of the mandible in relation to the conditions experienced by juveniles of 8-10 months of age during prenatal and early postnatal life and we evaluated these parameters as ecological indicators of juvenile condition as well as female reproductive condition in a roe deer population living in the southern part of the species range. We analyzed a sample of over 24,000 mandibles of roe deer shot in 22 hunting districts in the Arezzo province (Tuscany, Central Italy) from 2005 to 2015 per age class. Mandible total length in juveniles is equal to 90% of total length in adults. In this stage of life the growing of the mandible's anterior section is already completed while that of the posterior section is still ongoing. Environmental conditions conveyed by forest productivity, agricultural land use, local population density and climate strongly affected the growth of the anterior and posterior sections of the mandibles. Conditions experienced both by pregnant females and offspring played an important role in shaping the length of the anterior section, while the size of the posterior section was found to be related to the conditions experienced by offspring. Temporal changes of the length of the anterior section are a particularly suitable index of growth constraints. Anterior section length in fact differs according to more or less advantageous conditions recorded not only in the year of birth, but also in the previous year. Similarly, the sexual size dimorphism of the anterior section of the roe deer mandible can be used to describe the quality of females above two years of age, as well as habitat value. Hence the anterior section length of the mandible and its sexual size dimorphism are indexes that can provide cues of population performance, because they capture the system's complexities, while remain simple enough to be easily and routinely used in the majority of European countries where roe deer hunting period extends from early autumn to late spring.

Forecasting the response to global warming in a heat-sensitive species.

Avoiding hyperthermia entails considerable metabolic costs for endotherms. Such costs increase in warm conditions, when endotherms may trade food intake for cooler areas to avoid heat stress and maximize their energy balance. The need to reduce heat stress may involve the adoption of tactics affecting space use and foraging behaviour, which are important to understand and predict the effects of climate change and inform conservation. We used resource selection models to examine the behavioural response to heat stress in the Alpine ibex (Capra ibex), a cold-adapted endotherm particularly prone to overheating. Ibex avoided heat stress by selecting the space based on the maximum daily temperature rather than moving hourly to 'surf the heat wave', which minimised movement costs but prevented optimal foraging. By integrating these findings with new climate forecasts, we predict that rising temperatures will force mountain ungulates to move upward and overcrowd thermal refugia with reduced carrying capacity. Our approach helps in identifying priority areas for the conservation of mountain species.