Predictive logistic regression models are commonly used to make informed decisions related to wildlife management and conservation, such as predicting favourable wildlife habitat for land conservation objectives and predicting vital rates for use in population models. Frequently, models are developed for use in the same population from which sample data were obtained, and thus, they are intended for internal use within the same population. Before predictions from logistic regression models are used to make management decisions, predictive ability should be validated. We describe a process for conducting an internal model validation, and we illustrate the process of internal validation using logistic regression models for predicting the number of successfully breeding wolf packs in six areas in the US northern Rocky Mountains. We start by defining the major components of accuracy for binary predictions as calibration and discrimination, and we describe methods for quantifying the calibration and discrimination abilities of a logistic regression model. We also describe methods for correcting problems of calibration and future predictive accuracy in a logistic regression model. We then show how bootstrap simulations can be used to obtain unbiased estimates of prediction accuracy when models are calibrated and evaluated within the same population from which they were developed. We also show how bootstrapping can be used to assess coverage rates and recalibrate the endpoints of confidence intervals for predictions from a logistic regression model, to achieve nominal coverage rates. Using the data on successfully breeding wolf packs in the northern Rocky Mountains, we validate that predictions from a model developed with data specific to each of six analysis areas are better calibrated to each population than a global model developed using all data simultaneously. We then use shrinkage of model coefficients to improve calibration and future predictive accuracy for the area-specific model, and recalibrate confidence interval endpoints to provide better coverage properties. Following this validation, managers can be confident that logistic regression predictions will be reliable in this situation, and thus that management decisions will be based on accurate predictions.

The southeastern portion of the Edwards Plateau of Texas, historically a stronghold of Rio Grande wild turkeys Meleagris gallopavo intermedia, has seen the numbers of turkeys declining since the 1970s. Because hen survival is a key parameter affecting turkey population dynamics, we monitored radio-tagged Rio Grande wild turkey hens on the Edwards Plateau during 2001–2007 to compare survival during the breeding season in areas where turkey abundance has declined vs areas with stable populations. Breeding season survival was best predicted by a model that allowed variation during different periods of the reproductive season (initial nesting during 25 March – 21 April, renesting during 22 April – 12 June and post-nesting during 13 June – 2 July) and differences between stable and declining regions. Model-averaged estimates of breeding season survival were higher in the stable region (0.88; 95% CI: 0.78–0.94) than in the declining region (0.67; 95% CI: 0.55–0.78). Data collected on nests and hens during the intensive monitoring part of the study conducted during 2005–2007 indicated that breeding season survival was negatively affected by the amount of reproductive effort each season, i.e. the number of days a hen spent incubating a nest. Breeding season (25 March – 2 July) survival (0.86) was higher for non-nesting hens than for hens that nested (0.68; average of 15 days spent nesting). Our results suggest that differences in productivity between stable and declining Rio Grande wild turkey populations in the Edwards Plateau of Texas were associated with differences in breeding season survival due to variable reproductive effort, which, when combined with precipitation-dependent boom-bust dynamics common to galliforms in this region, could limit long-term population productivity and maintenance.

Ruffed grouse Bonasa umbellus populations in North America have declined as forests have matured and the extent of early successional forest habitat required by the species has diminished. When wildlife species decline because of habitat loss, determining where to focus habitat management efforts is difficult because both the wildlife population and the required habitat(s) are usually limited in distribution. We adopted a relatively new ecological modeling method, partitioned Mahalanobis D2, which allowed us to predict the distribution of potential ruffed grouse habitat across a landscape of management concern where high quality habitat was uncommon. We used presence data derived from radio-telemetry locations, and GIS habitat data from publicly available sources to create competing partitioned Mahalanobis D2 models. The competing models identified important habitat variables and predicted ruffed grouse habitat distribution at 1-ha and 25-ha scales in southwestern Rhode Island, USA. The 1- and 25-ha models produced comparable overall classification accuracy (83.1% and 81.4%, respectively) but differed substantially in the area of predicted habitat (4,475.5 ha and 10,133.8 ha, respectively). We selected the more conservative 1-ha model as the ‘best’ model, and expanded it to a larger landscape extent. Once expanded, the model predicted 11,463 ha (15.5% of total land area) of potential ruffed grouse habitat for a 735-km² landscape in southwestern Rhode Island. This model identified those areas with varying proximities to the following features as likely to contain ruffed grouse habitat: early successional forests, river and stream corridors, mixed conifer forests, conifer forests, shrub wetlands and deciduous forests. Early successional forests were the most consistent component of habitat used by grouse, despite the fact that this habitat type was uncommon in our study area (< 1% of total land area). Our model can be used to identify areas of existing ruffed grouse habitat for management focus.

Greater sage-grouse Centrocercus urophasianus populations in North Dakota declined approximately 67% between 1965 and 2003, and the species is listed as a Priority Level 1 Species of Special Concern by the North Dakota Game and Fish Department. The habitat and ecology of the species at the eastern edge of its historical range is largely unknown. We investigated nest site selection by greater sage-grouse and nest survival in North Dakota during 2005 – 2006. Sage-grouse selected nest sites in sagebrush Artemisia spp. with more total vegetative cover, greater sagebrush density, and greater 1-m visual obstruction from the nest than at random sites. Height of grass and shrub (sagebrush) at nest sites were shorter than at random sites, because areas where sagebrush was common were sites in low seral condition or dense clay or clay-pan soils with low productivity. Constant survival estimates of incubated nests were 33% in 2005 and 30% in 2006. Variables that described the resource selection function for nests were not those that modeled nest survival. Nest survival was positively influenced by percentage of shrub (sagebrush) cover and grass height. Daily nest survival decreased substantially when percentage of shrub cover declined below about 9% and when grass heights were less than about 16 cm. Daily nest survival rates decreased with increased daily precipitation.

The role of food in limiting or regulating populations of mammalian herbivores remains a central question in ecology with great relevance to wildlife and livestock management. Supplementary feeding studies have been widely used to assess the potentially limiting role of food availability, and supplementary feeding is also a common management technique. In both contexts there is an assumption that all individuals in the target population have access to food. There are, however, questions as to whether supplementary feed reaches the target population and how benefits are translated into individual and population level effects. We describe and use a technique using Passive Induced Transponder (PIT) tags to monitor individual use of supplementary feed in wild mountain hare Lepus timidus populations and test the assumption that supplementary feed reaches the target population. Over the course of one winter only 50% of the target hare population used supplementary feed and there was considerable individual variation in the time spent feeding among those individuals that fed. Neither age, sex nor an index of body condition were significant in explaining which individuals visited feeding stations or how long individuals spent feeding. The method and results described here suggest that, at least for the mountain hare, the central assumption that all target individuals have access to and use supplementary feed is invalid. Great care is thus needed in designing and interpreting the results of supplementary feeding studies or management programmes that include supplementary feeding.

We present an individual-based model of the Norwegian wolf Canis lupus population, which is used to evaluate the effectiveness of current and potential management policies in fulfilling the Norwegian Government's stated aim of maintaining three breeding packs within a designated wolf zone. The model estimates the ‘functional extinction rate’ of the population, defined as the proportion of years in which breeding wolf packs are absent. Under the current conditions according to estimates from Scandinavia, with observed values of natural survival rates (0.903) and unauthorised mortality (0.203) and allowing for immigration from Sweden, the model predicts that the probability of functional extinction is as low as 0.07. This output variable is highly sensitive to the demographic parameters, and if alternative estimates of natural survival rates (0.73 for cubs and 0.83 for adults) reported for wolves elsewhere and higher rates of unauthorised mortality (0.4) are utilised, the functional extinction rate is 0.67 ± 0.15, and the population is dependent on maintenance by immigration from Sweden. The main determinants of the functional extinction rate are the unauthorised mortality rate and the immigration rate from Sweden. The Scandinavian population as a whole shows a rapid non-linear increase in probability of extinction at unauthorised mortality rates >0.10. Varying levels of the current management interventions (increasing the size of the wolf zone and target number of packs) are ineffective; only when the unauthorised mortality rate falls below 0.30 is a self-sustaining population in Norway able to establish. An adaptive harvest policy with culls targeted only at dispersing animals, or taking place only when the population exceeds a threshold level, could be sustainable if the unauthorised mortality rate is reduced. The fact that the Norwegian Government has been explicit about its management strategy and objectives has allowed us to test the ability of this strategy to meet the objectives, and we have shown that it is dependant upon maintaining the current circumstances alongside a high adult survival rate to be able to do so. Given the potentially critical role of the Swedish population in sustaining Norway's wolves, there is a strong case for joint management of the Scandinavian population. These insights are likely to be relevant for the management of other species living across geopolitical boundaries.

Knowing what an animal is doing where and when is crucial for understanding habitat use as well as for detecting deviations from the norm, e.g. the animal's responses to disturbances or predators. While an animal's position can quite easily be assigned using VHF- or GPS-telemetry, determining its behaviour from a distance is still limited. A new generation of GPS-collars, equipped with a dual-axis acceleration sensor allows insights into the animal activity by continuously (5-minute intervals) delivering x- and y-values on a scale from 0 to 255. However, until now it has not been possible to tell which activity values can be attributed to which kind of behaviour. Therefore, the overall aim of our study was to find a method to distinguish different behavioural categories from these activity values. We used four captive red deer Cervus elaphus (one male and three females) and equipped them with GPS-collars while simultaneously observing their behaviour. Values for different behavioural categories were compared statistically using ANOVA with ‘individual’ as random effect and Tukey's follow-up test. Threshold values between the categories were determined by recursive partitioning and were assured by 5,000 bootstraps. While the difference between feeding and slow locomotion was significant in the x- but not in the y-values, each of these two categories differed significantly from resting and fast locomotion. Specific thresholds were established between the three categories resting, feeding with slow locomotion and fast locomotion. Subsequent comparison of the behaviour determined by these threshold values with observed behaviour revealed a high percentage of correctly assigned behaviour (93%). Taken together, this preliminary study demonstrates the potential of dual-axis acceleration sensors in GPS-collars for estimating the activity of wild-living red deer. However, further observations of activity on more individuals of each age and sex class should be performed to take into account inter-individual variability and to improve the predictive power of the threshold values.

Global positioning system (GPS) based telemetry studies are becoming more popular in large carnivore research. Recent advancements include detection and prediction of kill sites from GPS collar data. Thus far, the majority of models to detect kill sites focus on the patterns generated by a single focal individual. The prediction of kill sites helps increase sample sizes for diet studies of carnivores, especially when continuous observation methods cannot be employed. We propose and report on the feasibility of using the spatial association of multiple individuals from a social carnivore group to locate kill sites, using female lions Panthera leo in the Kruger National Park, South Africa, as an example. Our feasibility study suggests that lionesses cluster in space while at a GPS cluster with a kill. Clustering appeared most strongly in the first two hours of a kill, whereafter a more random association between individuals in space is observed. Additionally we found no difference in the initial spatial clustering pattern for kills of different sizes. When clusters are not associated with a kill (i.e. resting), female lions exhibit the random spatial association similar to the later hours found at kill sites. We feel that based on the initial results, association of social carnivores in space in combination with current spatio-temporal patterns of focal individuals can be used to improve kill-site models, but further research and larger sample sizes are required to validate our findings.

Indirect methods such as faecal counts have been widely used for assessing the abundance and habitat preferences of many mammal species, although their reliability has been long debated. We tested the validity of this method for the Eurasian badger Meles meles in a low-density population in northern Italy by comparing results obtained from radio-tracking and latrine distribution. The pattern of habitat use extrapolated from each method was strikingly similar, with badgers selecting for patches of woodland and avoiding agricultural and urban areas. Latrines were mainly sited in the centre of activity of all individuals along man-made linear features. Although our data need to be validated over a wider range of social groups, evidence suggests that latrines might be used to broadly infer habitat preferences of badgers at the landscape level in low-density areas.

We conducted a controlled feeding trial using two captive river otters Lontra canadensis to determine how prey size may introduce bias into frequency of occurrence analysis using otter scats. Otters were fed specific prey across a range of sizes. We then collected all scats deposited by the otters to determine how many defecation events occurred to remove the prey item from the digestive system. We found a strong, positive relationship between prey item size and the number of scats required to excrete the item. We then examined how the results, of an actual river otter feeding habits study using frequency of occurrence analysis of scats, could be biased towards an over-representation of larger prey items by using a correction factor for prey item size developed from our feeding trials. Frequency of occurrence suggested a strong preference for mid-range size of prey items and a strong avoidance of smaller prey items. Our corrected results indicated that otters exhibit little preferential feeding based on prey item size in the Missouri Ozarks. Our results suggest that bias associated with frequency of occurrence analyses may severely limit the robustness of inferences that can be made from such analyses.

Precise knowledge of how game species react to different hunting practices is a prerequisite for sound management of intensively hunted populations. We compared behavioural and spatial behaviour of five GPS-collared female red deer Cervus elaphus in Denmark before, during and after exposure to 21 driven hunts (2–5 times each). In 53% of all hunts, deer left their normal home ranges within 24 hours, moving on average 4 km and remaining away for an average of six days. Compared to pre-hunt values, deer moved longer distances per unit time on the day of the hunt and during the following two nights. Diurnal activity (based on motion sensors) did not increase significantly on the hunting day, but was lower than normal the day after the hunt. Nocturnal activity was equal before and after hunts. Deer spent 96% of their time in (safer) forest habitats by day and 43% by night before and after hunts. No induced responses were conditional on distance to the hunters (0–1.5 km), hunt duration (1.3–6.4 hours) or the time elapsed since previous hunts (4 to >30 days). The inclination of deer to flee from areas following hunts might complicate attempts to optimise harvesting policies in landscapes with many landowners within a typical flight range.