如何调试艾伦赫赛GL240016路调音台_百度知道PMCID: PMC4235582Behavioral responses of wolves to roads: scale-dependent ambivalence,
bDepartment of Ecology, Grimsö Wildlife Research Station, Swedish University of Agricultural Science, SE-73091 Riddarhyttan, SwedenCorresponding author.Address correspondence to B. Zimmermann. E-mail: .
Handling editor: Shinichi Nakagawa
This article has been
other articles in PMC.Throughout their recent recovery in several industrialized countries, large carnivores have had to cope with a changed landscape dominated by human infrastructure. Population growth depends on the ability of individuals to adapt to these changes by making use of new habitat features and at the same time to avoid increased risks of mortality associated with human infrastructure. We analyzed the summer movements of 19 GPS-collared resident wolves (Canis lupus L.) from 14 territories in Scandinavia in relation to roads. We used resource and step selection functions, including &#x field-checked GPS-positions and 315 kill sites. Wolves displayed ambivalent responses to roads depending on the spatial scale, road type, time of day, behavioral state, and reproductive status. At the site scale (approximately 0.1 km2), they selected for roads when traveling, nearly doubling their travel speed. Breeding wolves moved the fastest. At the patch scale (10 km2), house density rather than road density was a significant negative predictor of wolf patch selection. At the home range scale (approximately 1000 km2), breeding wolves increased gravel road use with increasing road availability, although at a lower rate than expected. Wolves have adapted to use roads for ease of travel, but at the same time developed a cryptic behavior to avoid human encounters. This behavioral plasticity may have been important in allowing the successful recovery of wolf populations in industrialized countries. However, we emphasize the role of roads as a potential cause of increased human-caused mortality.Key words: Canis lupus, functional response, movement, resource selection, road, step selection function, travel speed.Roads are man-made habitat features that are hardly comparable to any natural habitat: they are linear, have an open canopy, a hard surface, and often have parallel open-canopy strips with ground-cover vegetation on both sides. Connected with other roads, they form a network causing fragmentation of natural habitats. Roads are among the most recent of man-made habitat alterations, having spread dramatically during the past century following the development of motor vehicles (). Reviews of ecological effects of roads on wildlife populations highlight the direct mortality caused by collisions with vehicles and the indirect alteration of individual behavior due to habitat loss, fragmentation, and increased human access (; ; ; ). Increased access implies higher human-caused disturbance and predation risk as perceived by wildlife, thereby linking the two main types of human impact, habitat alteration and hunting. Our study aims to explore this link and its consequences for the behavioral ecology of the wolf, a top predator which is currently re-covering in many European countries ().The wolf is a pack-living, highly mobile species that defends large pack territories by frequent scent-marking (). Roads may therefore be a positive new addition to the landscape of wolves. Indeed, roads have been shown to ease travel for wolves (; ; ; ; ; ). Roads can facilitate territorial patrolling and serve as distinct features for scent marking (; ). Roads can lead to increased encounter rates between wolves and their prey and therefore increase kill rates (; ; ). As road sides are typical edge habitats with plant communities in early succession, they may provide minerals and energy-rich food for grazing and browsing prey species of wolves (; ; ) close to shelter habitat.On the other hand, roads have been shown to increase mortality of wolves directly due to traffic accidents and indirectly by increasing access for hunters and poachers (; ; ). Fragmentation decreases the availability of undisturbed habitat () whereas large roads can affect wolf movement () and in some cases act as barriers that limit dispersal and distribution on the population level (), but see also . Several studies have shown that the occurrence of wolf territories is higher in areas with low densities of roads and built-up areas (; ; ; ; ). The use of roads by wolves is likely constrained by the extent of human activity which may be perceived as an indicator of the risk of human-caused mortality. Previous studies have concluded that wolves prefer to use roads and trails with low human use (; ) or during times of low human activity, such as during night as opposed to day or winter as opposed to summer (; ). Preference for forest roads decreases with increasing road density within wolf home ranges (functional response, ). Reproducing wolves and their pups are likely most vulnerable to humans and other large predators during the summer season because movements are centered round a den site and later rendezvous sites (; ; ). Survival of both adults and pups may be severely affected if these sites are detected. As a result, such sites are often located far from sources of human disturbance (; ; ).In summary, roads pose a trade-off for wolves between human-induced negative effects (disturbance and increased mortality) and the positive effects resulting from increased ease of travel, efficient scent-marking, and access to prey. Our study examines ambivalent responses toward roads at different spatial and temporal scales from single wolf steps up to the landscape level for wolves in Scandinavia. After the functional extinction of the Scandinavian wolves in the 1960s (), a couple of immigrant wolves from Finland or Russia founded today’s population on the Scandinavian Peninsula in the early 1980s (). The population increased rapidly after 1990 () and totalled 33 family groups and 27–28 scent-marking pairs of wolves in winter
(). Although listed as critically endangered in Norway () and endangered in Sweden (), the wolf is still subject to both legal and illegal hunting in both countries. Poaching is assumed to account for half of all wolf mortality (). Conflicts regarding depredation of livestock, perceived competition for game species, and the loss of hunting dogs to wolves, all lead to an acceptance of poaching, especially in rural areas with free-ranging livestock and strong hunting traditions (). The dense network of gravel roads created during recent decades for forest exploitation is facilitating the access of poachers to remote areas. Traffic accidents also contribute substantially to wolf mortality in Scandinavia (; ; ) despite low densities of main roads and humans (; ).Habitat selection is a process acting at different spatial scales. According to
classification, animals choose locations at the landscape level (population range, first order), at the home range level within the population range (second order), at the patch level within the home range (third order), and at the site level within a patch (fourth order). Our study focused on the home range, patch, and site levels. We first tested the ease-of-travel hypothesis by predicting that the travel speed of wolves is higher on roads as compared to off roads. Secondly at the site scale (approximately 0.1 km2), we examined how road type, time of day, and wolf behavioral state affect road use and the distance wolves stay from roads (Johnson’s fourth order). We predicted that wolves 1) prefer to use gravel roads
2) use roads more often and are closer to roads durin and 3) use roads mainly while traveling and avoid being close to roads while handling prey and resting. Thirdly, we examined whether the wolves’ selection of patches within the home range (third order, spatial scale of 10 km2) was dependent on road density. We expected wolves to prefer areas of low road density in order to avoid human disturbance, and we expected breeding wolves to show a stronger avoidance than nonbreeders. Finally, at the home range scale (second order, approximately 1000 km2), we tested whether road use by wolves was a function of road density. We predicted a functional response due to human disturbance, that is wolves living in territories with high road densities use roads relatively less frequently than wolves in territories with low road densities. Our study focused on the summer period because 1) all gravel roads are available for wolves and people as opposed to winter when only some parts of gravel roads are snow-ploughed, and 2) breeding wolves may be more constrained during this period when movements are restricted to denning or rendezvous sites.Study areaThis study was carried out within the wolf breeding range in south-central parts of the Scandinavian Peninsula that is Sweden and Norway (; 59–62°N, 10–15°E, approximately 100000 km2). The wolf territories were primarily covered by boreal coniferous forest (mean ± SE: 81.7±1.3% for n = 14 territories) dominated by Scots pine (Pinus sylvestris L.) and Norway spruce (Picea abies L.), with some deciduous species, of which birch (Betula pubescens Ehrh.) and aspen (Populus tremula L.) were most abundant. Mire was the second most frequent land cover type within the wolf territories (10.5±1.2%), followed by water (4.6±0.8%), agricultural fields (1.6±0.6%), open areas (e.g. mountains, 1.4±0.9%), and built-up areas (0.2%). The density of main roads within territories averaged 0.19±0.02 km/km2, and the maximum distance to main roads ranged from 3.72 to 14.88 km (). A large network of gravel roads has been created due to extensive commercial logging and forest management practices () (). Gravel road densities in the territories were on average 4.6 times higher than main road densities and the maximum distance to gravel roads within territories ranged from 1.25 to 6.09 km (). Human density within the distribution of the Scandinavian wolf population is low, including vast areas with <1 person per km2 (). House densities within the territories averaged 3.0±0.4 per km2.Location of wolf territories included in this study (black outlines), and all main roads (bold purple lines) and gravel roads (thin purple lines) within the study area. The study area comprised most of the wolf breeding range on the Scandinavian Peninsula ...Wolf territories with area (100% MCP of all hourly GPS-positions during the study periods per territory and year), road densities and maximum distances to gravel roads and main roadsMoose (Alces alces L.) are the most important prey of wolves in Scandinavia, with a population density of approximately 1–2 moose/km2 in summer. For Scandinavian wolf packs, moose represent more than 95% of the food biomass in summer (). Other ungulate prey are roe deer (Capreolus capreolus L.), semi-domestic and wild reindeer (Rangifer tarandus L.), red deer (Cervus elaphus L.), and domestic sheep (Ovis aries L.). Smaller prey are also available for the wolf, including beaver (Castor fiber L.), badger (Meles meles L.), mountain and European hares (Lepus timidus L., Lepus europeus Pallas), capercaillie (Tetrao urogallus L.), and black grouse (Lyrurus tetrix L.) ().Study animals and periodAs part of the Scandinavian Wolf Research Project (SKANDULV), the data for this study were collected on 19 adult scent-marking wolves resident in 14 territories between 2002 and 2010 (). The wolves were immobilized from a helicopter following standard procedures (; ; ) and equipped with a GPS neck collar (GPS-Simplex, Web-Direct, or Tellus by Followit, Sweden, or GPS-Plus by Vectronic Aerospace, Germany). The study included 24 study periods of 8–29 days between June 1 and September 29. In nine study periods, both the adult male and female were GPS-collared, resulting in 33 individual data sets (). GPS-collars were programmed for hourly (8 data sets) or half-hourly (25 data sets) positioning intervals. GPS wolf position data for this study originated in a study of summer kill rates (), but we included additional, more recent time periods from two territories. During the study periods, eight of the territories had breeding wolves and in the other five the wolves were nonbreeding. In one territory, the adult wolves bred successfully but lost their pups after the first study period, and we therefore treated them as nonbreeders for the second study period (Koppang, ).GPS positions and cluster definitionOur original data sets consisted of 13188 hourly and 18910 half-hourly GPS positions in total. GPS success that is the percentage of successful positioning attempts per data set, averaged 81% (range 29–99%). We restricted our analysis of travel speed to half-hourly positions only and used positions at hourly intervals for all other analyses. In order to detect prey remains, we created 100 m buffers around all positions (). We dissolved the area of overlapping buffers and defined them as clusters (. , ; ). Clusters included not only consecutive positions but also revisits to the same spot over the entire study period. In this way, we classified 10951 hourly positions as wolf cluster positions. The other 2237 hourly positions were single positions further than 200 m from the next nearest position (). This method of spatially clustering positions at a given buffer radius of 100 m has proven to be the most successful for detecting prey remains in Scandinavia () and has since been applied in all Scandinavian kill rate studies (; , ).Movement analysis of GPS-positioning data of Scandinavian wolves showing how positions were separated into single positions &#x m from the nearest position, and cluster positions (a) and the step category-specific frequency distributions of step-lengths ...All cluster positions and most of the single positions, that is &#x positions in total, were visited in the field a few days after the wolf was present and searched for prey remains with the help of dogs (). This resulted in the detection of 315 wolf kills, of which 250 (79%) were associated with clusters of at least two positions (kill clusters), 39 (12%) were at single positions, and 26 (8%) were further than 100 m from the closest GPS-position. Clusters containing a den or rendezvous site, that is sites where cubs were moved to and placed during the summer (; ), were identified by a star-formation movement pattern of the adult wolves and were not visited in the field until at least 1 week after wolves had left the area. They consisted of a minimum of 44 hourly positions. We categorized all other clusters as bed clusters, associated with resting behavior. Bed clusters consisted of a maximum of 36 hourly positions.Resource mapsAs the wolf population was cross-border, we joined digital maps of Norway and Sweden. Vector data of roads and houses were derived from the Norwegian (scale 1:50000) and Swedish (1:100000) national maps. We categorized roads as either gravel roads (“Enskilda vägar” in Sweden and “Privatvei” in Norway) or main roads (“Allmäna vägar” in Sweden and “Riksvei,” “Fylkesvei” and “Kommunal vei” in Norway). In general, main roads were paved. Due to the large size of the study area (), we were not able to measure human activity on the &#x km of main roads and &#x km of gravel roads included in this study (). The classification of roads is based on the assumption that main roads connect human settlements, implying higher and more regular human disturbance. Gravel roads are mostly logging roads that experience short-term heavy use during logging or thinning operations once every 5–50 years, and moderate to low use during the fall game hunting season. The unified house map included all types of human buildings in both countries.Land cover data were derived from satellite-based maps that had been classified and approved by national authorities in both countries. In Norway, the SatVeg raster data (Source: Norwegian Environment Agency) had a pixel size of 30 m. The Swedish Corine Land Cover map with a pixel size of 25 m was provided by the National Land Survey of Sweden. For this study, we simplified the original habitat classes (24 in Norway and 59 in Sweden) into four land cover categories: Forest, Water, Mire, and Open. The last class included built-up areas, agricultural fields, and other terrestrial, nonforested areas.Wolf travel speedWe calculated wolf travel speed from individual steps with a maximum time length of 35 min and minimum step length of 200 m, resulting in a total of 3188 steps from 25 data sets. We categorized start and end positions of each step as being either on or off of a road, within a threshold of 30 m. The distance of 30 m corresponds to the inaccuracy of GPS positions (; ; ). In this way, we may have misclassified some wolf positions in close vicinity to roads as being on the road. Still, we considered this distance to be close enough for wolves to be aware of the road and to use it or the ditch as a guide through the landscape. For each step, we grouped the class variable Road into one of three states: 1) start AND 2) either start OR and 3) start AND end off road. In addition to this variable, we included the reproductive status of the wolves (Reproduction, Breeding, or Nonbreeding) and a two-class variable Time of day (Day 08:00&#x; Night 20:00&#x) in a linear mixed model (LMM, R 2.13.2). The classes of Time of day were more connected to human activity level than light conditions. During the summer months, nights do not get completely dark in the study area. Data set-ID (i.e. unique individual-study period combination) was nested within territory as a random factor and travel speed was the response variable. We eliminated nonsignificant (P > 0.05) variables and interactions in a stepwise backward procedure. For the final model, we reported the marginal and conditional R
2 for LMMs as proposed by .Wolf movement at the site scale: step selection functions (SSFs)For the site level habitat selection study (fourth order, ), we applied a matched case-control design (; ; ). By contrasting the resources at each animal position with a set of paired random points, we could estimate whether and to what extent certain resources affected animal movement. This design is statistically solved using conditional logistic regression (; ; ).
coined the term SSF where the creation of random points or steps is based on step characteristics of the animal path, rather than on the animal positions per se, as proposed for Resource Selection Functions (RSFs) ().For the SSF analyses, we categorized steps between hourly positions with a minimum length of 200 m into traveling (from single to single hourly positions), travel-to-cluster, cluster-to-travel and cluster-to-cluster steps (). We excluded steps to and from den and rendezvous clusters to avoid spatial autocorrelation, resulting in a total of 3154 steps. Frequency distributions of step length and turning angle differed among step categories (), with traveling steps being longer and more linear, and cluster-to-cluster steps being short and having the highest diversity of turning angles. Due to these differences, we used the frequency distributions of turning angles and step lengths to create 10 random steps per real step for each step category separately, using the conditional point sampling tool of Hawth’s tools (). A 10:1 ratio between paired random and real steps has been successfully applied in other earlier studies (; ; ). The 10 random steps together with the real step are called a stratum in the statistical language. We calculated the distances to the closest road and house, and determined the habitat type for the end points of each real and random step of each stratum. If a random step ended in water, we assigned the closest terrestrial habitat type to this end point.The response variable of the step selection models was a binary term with 1 for the used wolf locations, that is the end points of the real steps, and 0 for the end points of the random steps. To match the random steps to the corresponding real step, we applied conditional logistic regression (R 3.0.0) following the approach chosen by . Due to expected autocorrelation within territories, we used generalized estimating equations (GEE) including territory in the cluster term of the coxph-command (R package survival, ) to create robust standard errors. We preferred GEE to general linear mixed models (GLMM) because the sample size varied between individuals, and we were interested in the marginal rather than conditional estimates, that is drawing inferences for the entire Scandinavian wolf population rather than analyzing the differences between the studied individuals ().We estimated SSFs for all steps compiled, and separately for all combinations of time of day and behavioral state, that is at the end point of the real step the wolf was either handling prey (Kill), resting at a bed cluster (Rest), or traveling (single position, Travel) (). We ran an initial set of SSF models that included only the predictor on (<30 m) or off road (>30 m from closest road). Cross-validation of this set of SSFs was not possible because there was only one predictor and it had just two classes (on or off road). We however present the number of observed and random steps ending on roads.With the second set of SSF-models, we tested whether distance to human infrastructure (roads, houses) and land cover type predicted the choice of where wolves ended a step. Here, we included a quadratic term for distance to infrastructure to test for selection of intermediate distances. After checking predictor variables for collinearity with a pairwise Pearson rank correlation, we started with a full model that included all uncorrelated (r < 0.60) predictors and subsequently eliminated nonsignificant (P > 0.05) predictors in a stepwise backward procedure. We cross-validated the final models by excluding one wolf territory at the time, estimating model coefficients for the retained territories, and using these coefficients to predict the SSF for all real and random steps of the excluded territories. The predicted SSFs were ranked within each stratum of 11 paired steps with ranks from 1 to 11. From each stratum, we randomly sampled one random step. We then used a paired Wilcoxon signed rank test to validate whether the ranks of the real steps were higher than the ranks of the random steps. If P > 0.05, model fit was regarded as insufficient and the model was rejected.Wolf habitat selection at the patch scale: RSFsTo test whether the wolves selected patches with low road densities within their home ranges (third order habitat selection, ), we applied RSF-models with a presence/available design (). The home range was defined as the 100% minimum convex polygon (MCP) of all GPS positions in each of the 33 data sets. Used patches were considered to be all single positions and the first position in time of each cluster, in total 3291 positions, buffered with 1.78 km, resulting in a circular patch of 10 km2. As the mean number of used patches per data set was 99.7, we generated 100 random patches of equal dimensions (10 km2) within each home range to describe habitat availability.For each used and random patch, we derived the density of gravel and main roads (km roads/km2), the density of houses (km−2), and the percentage availability of the different land cover types. These variables were the fixed factors in mixed effects logistic regression models (GLMM), and we included the data set-ID nested within territory as a random factor. The response was a binary term of 1 for used and 0 for available patches. We started with a full model including all noncorrelated (Pearson’s r < 0.6) predictor variables in linear and quadratic form and used stepwise backward selection to exclude nonsignificant variables (P > 0.05). We did this modeling procedure separately for breeding and nonbreeding wolves in combination with the three different behaviors (handling kills, resting, and traveling).We used 10-fold cross-validation to validate our models (cf. ; ). For each training set, we extracted the model coefficients of the fixed effects and used them to predict the RSF values of the corresponding validation set. The validation set was then sorted by the RSF and split into 10 equal-sized bins. For each bin, we calculated the relative frequency of used patches. The degree of correlation (Spearman rank correlation) between the rank of the bin and the relative frequency of used patches was used as an indicator of model fit. We repeated this process 100 times for each final model and rejected models with an average Spearman’s r < 0.6.Functional response of road use at the home range scaleWe explored the relationship between road availability and use at the home range scale (second order, ) separately for breeding and nonbreeding wolves as home range use differed strongly with reproductive status. Nonbreeders ranged over areas 2.2 times larger than breeders (average ± SE home range size (100% MCP) of nonbreeders 818±138 km2; breeders 377±39 km2; t = 3.062, P = 0.009).For gravel road availability, we created 30 m buffers along all gravel roads and calculated the proportion of the land area covered by the buffered roads within individual home ranges. Gravel road use was the proportion of hourly positions per data set within 30 m of the closest gravel road. We ran linear regression models of the proportion of wolf positions on gravel roads, with proportion of gravel road area as the main predictor. If wolves used gravel roads in proportion to their density, we would expect a linear relationship with slope = 1 and intercept = 0. In addition we added the following covariates into the full model: Home range size, median Julian date of the study period, and sex of wolves. We eliminated nonsignificant variables (P > 0.05) in a stepwise backward procedure.Wolves traveled faster on than off roadsWhile traveling off road, wolves had an average speed of 2.15 km/h [n = 2500 steps, standard deviation (SD) = 1.54 km/h]. On roads, wolves traveled on average 1.8 times faster at a speed of 3.84 km/h (n = 91 steps, SD = 1.53 km/h). Speed was intermediate if wolves traveled partly on roads (n = 597 steps, mean = 2.96 km/h, SD = 1.41 km/h). The final LMM of travel speed included the variable Road in interaction with Reproduction, and Time of day (). Marginal and conditional R
LMM of this model were 7.6% and 15.5%, respectively. Breeding wolves moved on average 1.23 times faster than nonbreeders, with the difference being most pronounced when wolves used roads. Breeders had an average speed of 4.04 and 3.75 km/h on roads at night and day respectively, as compared to 3.15 and 2.86 km/h for nonbreeders. Speed was similar for male and female wolves, and sex was not retained in the final model.Model estimates of the final model for wolf travel speed (km/h) in Scandinavia, based on n = 3188 half-hourly stepsTo what extent did wolves use roads?Of all 3154 hourly steps used in the SSF models, 328 (10.4%) ended on gravel roads and 30 (1.0%) on main roads. Taking into account road availability in the SSF-models, wolves were 3.1 times more likely (eβ, ) to end a step on a gravel road than off road (P < 0.001), but there was no such preference or avoidance of main roads (P = 0.097). The separate SSFs for different times of day and behavior revealed that wolves preferred to use roads of either type for traveling, but not for other behaviors. They were 3.5 and 5.2 times more likely to travel on gravel roads during day and night, respectively (eβ, ), than off roads. Main roads were only preferred for travel during night time, with wolves being 1.9 times more likely to travel on a main road than off road (P = 0.002, ).SSFs of Scandinavian wolves for the different combinations of time of day and behavior, together with number of real and random steps ending on a gravel road or a main road, and a summary of the conditional logistic regression models with robust SE estimation ...Distance to human infrastructureFor each combination of time of day and behavior, the distance to the closest gravel road and/or main road was a significant predictor of the SSF (). However, model validation revealed a poor model fit for SSFs of wolves moving to kill sites (), and we are therefore unable to conclude whether and how wolves chose kill sites in relation to human infrastructure.Coefficients (β) and robust SE of SSFs for wolves in Scandinavia in different behavioral states (handling prey [kill], resting and traveling) during day and night hoursWhile resting during day time, wolves preferred intermediate distances to gravel roads, and they were 1.4 times more likely to bed at distances of 1&#x km from the closest gravel road as compared to directly at the road (). They preferred to have day bed sites far away from main roads () and at intermediate distances to houses, with a peak at 2 km from the closest house (). In addition, they avoided open habitats for day bed sites (). While resting during night time, they also avoided open habitats and preferred areas far from main roads, although this preference pattern was weaker than during the day time (lower odds ratios in
and significance of coefficients in ). Gravel roads did not seem to affect the choice of resting sites during night time.SSFs for Scandinavian wolves during summer, expressed as odds ratios (eβ) in relation to distance to gravel roads (a), main roads (b) and houses (c). Estimates of βs are given in .While traveling during day time, wolves preferred areas far from main roads () and at intermediate distances to houses (), whereas they seemed to be indifferent to gravel roads. They avoided open habitats during day time travel but were indifferent to the habitat type during night time travel (). The strong preference of wolves for using gravel roads for traveling during night time () is reflected by the U-shaped odds ratio in , indicating preference for near and far distances to gravel roads. While traveling during night time, wolves showed a weak preference for intermediate distances to main roads (, ).Patch selection as a function of house densityThe RSF-models developed to predict patch selection for kill sites of breeding and nonbreeding wolves did not have sufficiently good fit to draw conclusions about the potential effects of roads and houses on kill site selection (). Nonbreeding wolves were most likely to travel in patches with high gravel road densities (), and a patch was 1.7 times more likely to be used (eβ) if the gravel road density was increased by 1 km/km2 (). Nonbreeders selected patches with low main road densities to rest in (), and the likelihood of using a patch for resting was reduced by one-third if the main road density increased by 1 km/km2 (). In contrast, road densities did not relate to patch selection of breeding wolves. House densities were negatively correlated with the likelihood of wolves using a patch for resting or traveling, regardless of their reproductive status (, ). This negative relationship was stronger if wolves were resting than if they were traveling. The likelihood of breeding wolves using a patch for resting or traveling increased with the proportion of the patch covered by mire (, ).Coefficients (β) ± SE of RSF models describing the patch selection (patch size = 10 km2) of Scandinavian wolves within their home ranges during summerResource selection functions for patch (10 km2) selection of Scandinavian wolves during summer, expressed as odds ratios (eβ) in relation to density of gravel roads (a), main roads (b), houses (c), and the proportion of the patch covered with ...Functional response to gravel road densityGravel roads with a 30 m buffer covered on average 5±0.5% (2 SE, range 2.9&#x%) of the land area of the home ranges (n = 33). Road use varied highly among nonbreeding wolves (n = 12), with on average 4.4±2.5% (range 0.7&#x%) of hourly positions on gravel roads (). There was no correlation between road availability and road use by nonbreeding adult wolves (P = 0.499, ). However, road use by breeding wolves (n = 21) was less variable (mean = 3.3±0.6%) and positively related to road availability (). The proportion of gravel road area was the only significant variable in the final linear model describing this relationship (Intercept = 0.779±1.035; slope = 0.504±0.201; P = 0.021; R
2 = 0.25). The slope was significantly lower than 1 () indicating that although breeding wolves increased road use with increasing gravel road availability in their home range, this increase was 50% lower than expected. Sex of wolves, home range size, and Julian date were not related to road use by breeding or nonbreeding wolves.Gravel road use, expressed as % wolf positions on gravel roads, in relation to gravel road availability in the home ranges of Scandinavian wolves, expressed as % of land area covered with gravel roads, for nonbreeding (a) and breeding (b) adult wolves. ...We have demonstrated that the behavioral response of Scan dinavian wolves to roads is a complex multi-factorial process dependent on time of day, road type, behavioral state, reproductive status, and spatial scale. In the discussion below, we break this complexity down by considering each spatial scale from site level to patch level and finally to the home range level. Special emphasis is given to the differences between breeding and nonbreeding wolves, as reproductive status is an important determinant of population viability and has not previously been studied in comparable wolf-road publications.At the site scale, wolves in Scandinavia showed a clear preference for traveling on gravel roads and even on main roads during night hours, in summer. Similar preference patterns of wolves for minor roads and other man-made linear features have been described elsewhere (, ; ; ). Ease of travel is the most plausible reason why wolves in Scandinavia displayed this strong preference. They traveled nearly twice as fast on roads compared to off roads and breeding wolves traveled faster than nonbreeders, especially on roads. When food has to be provided to other pack members at den or rendezvous sites, roads likely serve as a positive medium for traveling in terms of minimizing energy expenditure and maximizing speed of food delivery. Higher travel speeds on linear features have also been recorded for wolves collared with Very High Frequency (VHF) tags in Poland () and GPS-collared wolves on seismic lines in Alberta, Canada () while GPS-collared cougars (Puma concolor L.) () and bison (Bison bison L.) () traveled faster on dirt roads than off road.Territory maintenance by scent-marking, which allows efficient communication toward intruders, is another plausible explanation for the extensive use of roads by wolves. Wolves regularly scent-mark along roads (; ), and have higher scent-marking frequencies on roads than off roads (). An alternative explanation is that prey use roads for travel and roadside or post-logging vegetation for food and minerals (; ), which in turn attracts the wolves. However, an earlier case study in Scandinavia showed that moose with calves, the main prey of wolves in this area, avoided being close to gravel roads during summer (), perhaps as an anti-predator strategy against wolves.Although at the site scale Scandinavian wolves displayed an overall preference for roads during summer, this preference was dependent on the road type, with main roads being less attractive than gravel roads. These findings support results from other studies that showed that wolves decreased use of roads and paths or increased their distance from these features with increasing size of roads or increasing rate of human use (; ; ; ; ; ; ). Time of day influenced the behavioral response of wolves to human infrastructure, likely caused by the diurnal activity pattern of humans. A similar night-bias of road use has been reported for wolves in Europe and Canada (; ; ) and a meta-analysis across wolf studies concluded that nocturnal activity and movement were positively correlated with public road density ().Resource selection is a function of the behavioral state of an animal (). Although the wolves in our study preferred using roads for traveling, they preferred to rest at intermediate distances to gravel roads and far away from main roads. We interpret the observed preference of resting at intermediate distances to gravel roads as a trade-off between the risk of encountering humans and good access to roads for increased travel speed and scent-marking.Minimizing the probability of encountering humans has been identified as an important factor shaping habitat selection of wolves within the home range, and road density has been used as a proxy for this (; ; ). However, our prediction that road densities at the patch level would affect wolf habitat selection negatively was only partly supported. Nonbreeding wolves preferred to rest in patches with low main road densities while contrary to our prediction they were more likely to travel in patches with high gravel road densities. For breeding wolves, we could not detect any relationship between patch selection and road density. More importantly however, patch selection by wolves was negatively related to house densities within the home range, both for breeders and nonbreeders, when resting and traveling.
described a similar relationship between the number of inhabitants of settlements and avoidance of close surroundings by wolves in Poland. At this regional scale, houses may be a more predictable indicator than roads of the probability of encountering humans. Human activity along gravel roads depends on the land use cover and history in the area (; ), whether the road has restricted access for motorized traffic, and whether it connects with major traffic arteries and settlements. These are factors that we did not measure or account for in our study. Another indicator of the probability of encountering humans is the ratio between productive and unproductive lands in a patch. As the proportion of the patch area covered by mire (unproductive land) increases, we expect a decrease in human activity. In our study, preferred patches of the home range of breeding wolves had a relatively high proportion of mire.Finally, we expected gravel road use by wolves to vary with gravel road density across home ranges. This functional response of wolves to roads was only supported for breeding wolves in Scandinavia. Although road use by breeders was positively correlated with gravel road density, the functional response was less than proportional to gravel road availability, that is breeding wolves used gravel roads less frequently than expected as gravel road densities increased. In a study of two wolf packs tracked on snow in Alberta, Canada,
found that the use of roads and trails was negatively coupled with road density. Their study looked at the functional response at the patch scale within the home range. Another study of GPS-collared wolves in Quebec, Canada, found selection of forest roads decreased with increasing forest road density within the home range, while at the between-pack scale, road density was not related to road use (). RSF-models of GPS-data from a wolf study in Alberta revealed a more complex picture of the selection of proximity to human activity (). Although wolf packs in areas of low human activity were indifferent to proximity to humans, wolves from a few packs in areas of high human activity showed a pattern of selection for close proximity to man-made linear features that varied with season and time of day.
attributed this functional response at the pack scale to the constraints of sharing habitats with humans. Wolf avoidance of areas with high road or house densities, either directly at the patch scale or relatively at the home range scale may result in a trophic cascade. Lowered predation pressure aggregates prey, leading to increased browsing pressure in areas of high human activity (; ).Wolves in Scandinavia have adapted to use roads for traveling, scent-marking, and territorial patrolling, but they have also developed cryptic behavioral responses to roads, likely driven by the increased risks associated with human presence. The high behavioral plasticity which allows such ambivalent responses of wolves toward infrastructure is a key factor in the recent wolf recovery in industrialized countries, many of which have higher densities of roads and humans than in Scandinavia. There are other success stories of species that have adaptively responded to man-made habitat alterations while still avoiding increased rates of human-caused mortality, for example urban wildlife or raptors feeding on vehicle-killed prey (; ). However, there are many more examples of species that have maladaptive responses, causing decreased individual fitness with negative effects on population growth and distribution (; ). Migration and connectivity in Scandinavian wild reindeer have been interrupted by the barrier effect of linear features (). In North America, the closure or removal of logging roads and other measures to decrease human access have been proposed to restore grizzly and black bear (Ursus arctos L., Ursus americanus Pallas) habitat by limiting the mortality risk of hunting and poaching (; ). Despite the seemingly well-functioning adaptation of wolves to man-made habitat alterations, we should be aware that roads may interact with human attitudes, enabling increased human-caused mortality of wolves. The acceptance of wolf poaching is relatively high in rural Scandinavia (), and the accessibility of wolf territories by gravel roads is crucial for poachers to increase their encounter rate with wolves, especially during the breeding period. The existing network of gravel roads is likely to be an important factor governing the vulnerability of wolves to human caused mortality and may negatively affect the resilience of the relatively small Scandinavian wolf population which currently suffers from inbreeding depression (; ).This SKANDULV study was supported by the Norwegian Research Council, the Norwegian Directorate of Nature Management, the Norwegian Institute for Nature Research, the County Governor of Hedmark, the Swedish Environmental Protection Agency, World Wildlife Fund for Nature (Sweden), Swedish University of Agricultural Sciences, the Swedish Association for Hunting and Wildlife Management, the Swedish Carnivore Association, several county municipalities, and Hedmark University College.We would like to thank all involved in the field work of the summer predation studies. C. Milleret kindly helped to join resource maps across country borders. Wolf capture, handling, and collaring was approved by the Swedish Committee of Animal Welfare (C 281/6, C 266/99), the Norwegian Animal Research Authority, and the Norwegian Directorate of Nature Management. We thank J. Milner for language check, and L. Simmons, H. P. Andreassen, and 2 anonymous reviewers for their comments on a previous version of this manuscript.
Alexander SM, Waters NM.
The effects of highway transportation corridors on wildlife: a case study of Banff National Park. Transport Res C-Emer. 8:307&#x.
Barja I, de Miguel FJ, Bárcena F.
The importance of crossroads in faecal marking behaviour of the wolves (Canis lupus). Naturwissenschaften. 91:489&#x.
Benitez-Lopez A, Alkemade R, Verweij PA.
The impacts of roads and other infrastructure on mammal and bird populations: a meta-analysis. Biol Conserv. 143:1307&#x.
Bensch S, Andrén H, Hansson B, Pedersen HC, Sand H, Sejberg D, Wabakken P, Akesson M, Liberg O.
Selection for heterozygosity gives hope to a wild population of inbred wolves. PLoS One. 1:e72. []
Hawth’s analysis tools for ArcGIS version 3.27. Available from: .
Beyer HL, Haydon DT, Morales JM, Frair JL, Hebblewhite M, Mitchell M, Matthiopoulos J.
The interpretation of habitat preference metrics under use–availability designs. Philos T R Soc B. 365:2245&#x. []
Beyer HL, Merrill EH, Varley N, Boyce MS.
Willow on Yellowstone’s northern range: evidence for a trophic cascade?
Ecol Appl. 17:1563&#x.
Blanco JC, Cortés Y, Virgós E.
Wolf response to two kinds of barriers in an agricultural habitat in Spain. Can J Zool. 83:312&#x.
Bowman JL, Kochanny CO, Demarais S, Leopold BD.
Evaluation of a GPS collar for white-tailed deer. Wildl Soc Bull. 28:141&#x.
Boyce MS, Mao JS, Merrill EH, Fortin D, Turner MG, Fryxell J, Turchin P.
Scale and heterogeneity in habitat selection by elk in Yellowstone National Park. Ecoscience. 10:421&#x.
Boyce MS, Vernier PR, Nielsen SE, Schmiegelow FK.
Evaluating resource selection functions. Ecol Model. 157:281&#x.
Bruggeman JE, Garrott RA, White PJ, Watson FG, Wallen R.
Covariates affecting spatial variability in bison travel behavior in Yellowstone National Park. Ecol Appl. 17:1411&#x.
Capitani C, Mattioli L, Avanzinelli E, Gazzola A, Lamberti P, Mauri L, Scandura M, Viviani A, Apollonio M.
Selection of rendezvous sites and reuse of pup raising areas among wolves Canis lupus of north-eastern Apennines, Italy. Acta Theriol. 51:395&#x.
Cargnelutti B, Coulon A, Hewison AJM, Goulard M, Angibault JM, Morellet N.
Testing Global Positioning System performance for wildlife monitoring using mobile collars and known reference points. J Wildl Manage. 71:1380&#x.
Ciucci P, Boitani L, Francisci F, Andreoli G.
Home range, activity and movements of a wolf pack in central Italy. J Zool. 243:803&#x.
Coffin AW.
From roadkill to road ecology: a review of the ecological effects of roads. J Transp Geogr. 15:396&#x.
Coulon A, Morellet N, Goulard M, Cargnelutti B, Angibault J-M, Hewison AM.
Inferring the effects of landscape structure on roe deer (Capreolus capreolus) movements using a step selection function. Landscape Ecol. 23:603&#x.
Craiu RV, Duchesne T, Fortin D.
Inference methods for the conditional logistic regression model with longitudinal data. Biometrical J. 50:97&#x.
Dickson BG, Jenness JS, Beier P.
Influence of vegetation, topography, and roads on cougar movement in southern California. J Wildl Manage. 69:264&#x.
Eriksen A, Wabakken P, Zimmermann B, Andreassen HP, Arnemo JM, Gundersen H, Liberg O, Linnell J, Milner JM, Pedersen HC, et al.
Activity patterns of predator and prey: a simultaneous study of GPS-collared wolves and moose. Anim Behav. 81:423&#x.
Eriksen A, Wabakken P, Zimmermann B, Andreassen HP, Arnemo JM, Gundersen H, Milner JM, Liberg O, Linnell J, Pedersen HC, et al.
Encounter frequencies between GPS-collared wolves (Canis lupus) and moose (Alces alces) in a Scandinavian wolf territory. Ecol Res. 24:547&#x.
Fahrig L, Rytwinski T.
Effects of roads on animal abundance: an empirical review and synthesis. Ecol Soc. 14:21 (online).
Forman RTT, Alexander LE.
Roads and their major ecological effects. Annu Rev Ecol Syst. 29:207&#x.
Fortin D, Beyer HL, Boyce MS, Smith DW, Duchesne T, Mao JS.
Wolves influence elk movements: behavior shapes a trophic cascade in Yellowstone National Park. Ecology. 86:1320&#x.
Fortin D, Fortin ME, Beyer HL, Duchesne T, Courant S, Dancose K.
Group-size-mediated habitat selection and group fusion-fission dynamics of bison under predation risk. Ecology. 90:2480&#x.
Francis RA, Chadwick MA.
What makes a species synurbic?
Appl Geogr. 32:514&#x.
Gangaas KE, Kaltenborn BP, Andreassen HP.
Geo-spatial aspects of acceptance of illegal hunting of large carnivores in Scandinavia. PLoS One. 8:e68849. []
Gärdenfors U.
Rödlistade arter i Sverige 2010 – The 2010 Red List of species in Sweden. Uppsala (Sweden): ArtDatabanken, SLU;
Gurarie E, Suutarinen J, Kojola I, Ovaskainen O.
Summer movements, predation and habitat use of wolves in human modified boreal forests. Oecologia. 165:891&#x.
Hebblewhite M, Merrill E.
Modelling wildlife-human relationships for social species with mixed-effects resource selection models. J Appl Ecol. 45:834&#x.
Hebblewhite M, Merrill EH, McDonald TL.
Spatial decomposition of predation risk using resource selection functions: an example in a wolf-elk predator-prey system. Oikos. 111:101&#x.
Hebblewhite M, Munro RH, Merrill EH.
Trophic consequences of postfire logging in a wolf-ungulate system. For Ecol Manage. 257:1053&#x.
Hebblewhite M, White CA, Nietvelt CG, McKenzie JA, Hurd TE, Fryxell JM, Bayley SE, Paquet PC.
Human activity mediates a trophic cascade caused by wolves. Ecology. 86:2135&#x.
Houle M, Fortin D, Dussault C, Courtois R, Ouellet J-P.
Cumulative effects of forestry on habitat use by gray wolf (Canis lupus) in the boreal forest. Landscape Ecol. 25:419&#x.
Huston MA.
The three phases of land-use change: implications for biodiversity. Ecol Appl. 15:1864&#x.
James ARC.
Effects of industrial development on the predator-prey relationship between wolves and caribou in northeastern Alberta. Alberta (Canada): University of Alberta.
James ARC, Stuart-Smith AK.
Distribution of Caribou and Wolves in Relation to Linear Corridors. J Wildl Manage. 64:154&#x.
Jędrzejewski W, Jędrzejewska B, Zawadzka B, Borowik T, Nowak S, Mysłajek RW.
Habitat suitability model for Polish wolves based on long-term national census. Anim Conserv. 11:377&#x.
Jędrzejewski W, Schmidt K, Theuerkauf J, Jędrzejewska B, Okarma H.
Daily movements and territory use by radiocollared wolves (Canis lupus) in Bialowieza Primeval Forest in Poland. Can J Zool. 79:1993&#x.
Johnson DH.
The comparison of usage and availability measurements for evaluating resource preference. Ecology. 61:65–71.
Kaartinen S, Kojola I, Colpaert A.
Finnish wolves avoid roads and settlements. Ann Zool Fenn. 42:523&#x.
Kålås J, Viken Å, Henriksen S, Skjelseth S.
The 2010 Norwegian red list for species. Trondheim (Norway): Norwegian Biodiversity Information Centre.
Karlsson J, Broseth H, Sand H, Andren H.
Predicting occurrence of wolf territories in Scandinavia. J Zool. 272:276&#x.
Koper N, Manseau M.
Generalized estimating equations and generalized linear mixed-effects models for modelling resource selection. J Appl Ecol. 46:590&#x.
Kreeger TJ, Arnemo JM.
Handbook of wildlife chemical immobilization. 4th ed. Sybille (Wyoming): Terry J. Kreeger.
Kunkel KE, Pletscher DH.
Habitat factors affecting vulnerability of moose to predation by wolves in southeastern British Columbia. Can J Zool. 78:150&#x.
Laurian C, Dussault C, Ouellet J-P, Courtois R, Poulin M, Breton L.
Behavior of moose relative to a road network. J Wildl Manage. 72:1550&#x.
Liberg O, Andrén H, Pedersen HC, Sand H, Sejberg D, Wabakken P, Kesson M, Bensch S.
Severe inbreeding depression in a wild wolf (Canis lupus) population. Biol Lett. 1:17–20. []
Liberg O, Chapron G, Wabakken P, Pedersen HC, Hobbs NT, Sand H.
Shoot, shovel and shut up: cryptic poaching slows restoration of a large carnivore in Europe. Proc Biol Sci. 279:910&#x. []
Linnell JDC, Promberger C, Boitani L, Swenson JE, Breitenmoser U, Andersen R.
The linkage between conservation strategies for large carnivores and biodiversity: the view from the “half-full” forests of Europe. In: Ray JC, Redford KH, Steneck RS, Berger J, editors. , editors. Large carnivores and the conservation of biodiversity. Washington (DC): Island P
p. 381&#x.
Manly BFJ, McDonald L, Thomas DL, McDonald TL, Erickson WP.
Resource selection by animals: statistical design and analysis for field studies. Dordrecht: Kluwer.
Mech LD, Boitani L.
Wolf social ecology. In: Mech LD, Boitani L, editors. , editors. Wolves: behavior, ecology and conservation. Chicago: The University of Chicago P
p. 1–34.
Mech LD, Fritts SH, Radde GL, Paul WJ.
Wolf distribution and road density in Minnesota. Wildl Soc Bull. 16:85–87.
Mladenoff DJ, Sickley TA, Haight RG, Wydeven AP.
A regional lanscape analysis and prediction of favorable gray wolf habitat in the Northern Great Lakes Region. Conserv Biol. 9:279&#x.
Morner T, Eriksson H, Bröjer C, Nilsson K, Uhlhorn H, Agren E, af Segerstad CH, Jansson DS, Gavier-Widén D.
Diseases and mortality in free-ranging brown bear (Ursus arctos), gray wolf (Canis lupus), and wolverine (Gulo gulo) in Sweden. J Wildl Dis. 41:298&#x.
Muhly TB, Semeniuk C, Massolo A, Hickman L, Musiani M.
Human activity helps prey win the predator-prey space race. PLoS One. 6:e17050. []
The wolves of Mount McKinley. Washington: U.S. G.P.O.
Musiani M, Okarma H, Jędrzejewski W.
Speed and actual distances travelled by radiocollared wolves in Białowieża Primeval Forest (Poland). Acta Theriol. 43:409&#x.
Nakagawa S, Schielzeth H.
A general and simple method for obtaining R2 from generalized linear mixed-effects models. Methods Ecol Evol. 4:133&#x.
Nielsen SE, Stenhouse GB, Boyce MS.
A habitat-based framework for grizzly bear conservation in Alberta. Biol Conserv. 130:217&#x.
Person DK, Russell AL.
Correlates of mortality in an exploited wolf population. J Wildl Manage. 72:1540&#x.
Person DK, Russell AL.
Reproduction and den site selection by wolves in a disturbed landscape. Northwest Sci. 83:211&#x.
Peters RP, Mech LD.
Scent-marking in wolves. Am Sci. 63:628&#x.
Rea RV, Child KN, Spata DP, Macdonald D.
Road and rail side vegetation management implications of habitat use by moose relative to brush cutting season. Environ Manage. 46:101&#x.
Rodgers AR.
Tracking animals with GPS: the first 10 years. In: Sibbald AM, Gordon LJ, editors. , editors. Tracking animals with GPS. Aberdeen: The Macaulay Land Use Research I
p. 1–10.
Rogala JK, Hebblewhite M, Whittington J, White CA, Coleshill J, Musiani M.
Human activity differentially redistributes large mammals in the Canadian Rockies National Parks. Ecol Soc. 16:16 (online).
Sand H, Vucetich JA, Zimmermann B, Wabakken P, Wikenros C, Pedersen HC, Peterson RO, Liberg O.
Assessing the influence of prey–predator ratio, prey age structure and packs size on wolf kill rates. Oikos. 121:1454&#x.
Sand H, Wabakken P, Zimmermann B, Johansson O, Pedersen HC, Liberg O.
Summer kill rates and predation pattern in a wolf-moose system: can we rely on winter estimates?
Oecologia. 156:53–64.
Sand H, Zimmermann B, Wabakken P, Andren H, Pedersen HC.
Using GPS technology and GIS cluster analyses to estimate kill rates in wolf-ungulate ecosystems. Wildl Soc Bull. 33:914&#x.
Schmidt K, Jędrzejewski W, Theuerkauf J, Kowalczyk R, Okarma H, Jędrzejewska B.
Reproductive behaviour of wild-living wolves in Bialowieza Primeval Forest (Poland). J Ethol. 26:69–78.
Switalski TA, Nelson CR.
Efficacy of road removal for restoring wildlife habitat: black bear in the Northern Rocky Mountains, USA. Biol Conserv. 144:2666&#x.
Therneau T.
A package for survival analysis in S. R package version 2.37 Available from: .
Theuerkauf J.
What drives wolves: fear or hunger? Humans, diet, climate and wolf activity patterns. Ethology. 115:649&#x.
Theuerkauf J, Gula R, Pirga B, Tsunoda H, Eggermann J, Brzezowska B, Rouys S, Radler S.
Human impact on wolf activity in the Bieszczady Mountains, SE Poland. Ann Zool Fenn. 44:225&#x.
Theuerkauf J, Jedrzejewski W, Schmidt K, Gula R.
Spatiotemporal segregation of wolves from humans in the Bialowieza Forest (Poland). J Wildl Manage. 67:706&#x.
Theuerkauf J, Rouys S, Jedrzejewski W.
Selection of den, rendezvous, and resting sites by wolves in the Bialowieza Forest, Poland. Can J Zool. 81:163&#x.
Relationship between road densities and wolf habitat suitability in Wisconsin. Am Midl Nat. 113:404&#x.
Thurber JM, Peterson RO, Drummer TD, Thomasma SA.
Gray wolf response to refuge boundaries and roads in Alaska. Wildl Soc Bull. 22:61–68.
Tsunoda H, Gula R, Theuerkauf J, Rouys S, Radler S, Pirga B, Eggermann J, Brzezowska B.
How does parental role influence the activity and movements of breeding wolves?
J Ethol. 27:185&#x.
Tuomainen U, Candolin U.
Behavioural responses to human-induced environmental change. Biol Rev. 86:640&#x.
Vilà C, Sundqvist AK, Flagstad Ø, Seddon J, Björnerfeldt S, Kojola I, Casulli A, Sand H, Wabakken P, Ellegren H.
Rescue of a severely bottlenecked wolf (Canis lupus) population by a single immigrant. Proc Biol Sci. 270:91–97. []
Vistnes I, Nellemann C, Jordhoy P, Strand O.
Effects of infrastructure on migration and range use of wild reindeer. J Wildl Manage. 68:101&#x.
Wabakken P, Sand H, Liberg O, Bjärvall A.
The recovery, distribution, and population dynamics of wolves on the Scandinavian peninsula, 1978&#x. Can J Zool. 79:710–-&#x.
Wabakken P, Svensson L, Kojola I, Maartmann E, Strømseth TH, Flagstad Ø, Åkesson M, Zetterberg A.
Ulv i Skandinavia og Finland: Sluttrapport for bestandsovervåking av ulv vinteren 2011&#x. Elverum: Høgskolen i Hedmark.
Whittington J.
Movement of wolves (Canis lupus) in response to human development in Jasper National Park. Alberta (Canada): University of Alberta.
Whittington J, Hebblewhite M, DeCesare NJ, Neufeld L, Bradley M, Wilmshurst J, Musiani M.
Caribou encounters with wolves increase near roads and trails: a time-to-event approach. J Appl Ecol. 48:1535&#x.
Whittington J, St Clair CC, Mercer G.
Path tortuosity and the permeability of roads and trails to wolf movement. Ecol Soc. 9:4 (online).
Whittington J, St Clair CC, Mercer G.
Spatial responses of wolves to roads and trails in mountain valleys. Ecol Appl. 15:543&#x.
Zimmermann B, Wabakken P, Sand H, Pedersen HC, Liberg O.
Wolf movement patterns: a key to estimation of kill rate?
J Wildl Manage. 71:1177&#x.
Zub K, Theuerkauf J, Jędrzejewski W, Jędrzejewska B, Schmidt K, Kowalczyk R.
Wolf pack territory marking in the Białowieża Primeval Forest (Poland). Behaviour. 140:635&#x.Articles from Behavioral Ecology are provided here courtesy of Oxford University Press
)- subscription required
Show all items
Show all items
Show all items
Show all items
Show all items
Show all items
Show all items
Show all items
Show all items
Show all items
Show all items
Show all items
Show all items
Show all items
Show all items
Show all items
CitePeer Related Articles