Aim Temporal transferability is an important issue when habitat models are used beyond the time frame corresponding to model development, but has not received enough attention, particularly in the context of habitat monitoring. While the combination of remote sensing technology and habitat modelling provides a useful tool for habitat monitoring, the effect of incorporating remotely sensed data on model transferability is unclear. Therefore, our objectives were to assess how different satellite-derived variables affect temporal transferability of habitat models and their usefulness for habitat monitoring.Location Wolong Nature Reserve, Sichuan Province, China.Methods We modelled giant panda habitat with the maximum entropy algorithm using panda presence data collected in two time periods and four different sets of predictor variables representing land surface phenology. Each predictor variable set contained either a time series of smoothed wide dynamic range vegetation index (WDRVI) or 11 phenology metrics, both derived from single-year or multi-year (i.e. 3-year) remotely sensed imagery acquired by the Moderate Resolution Imaging Spectroradiometer (MODIS). We evaluated the ability of models obtained with these four variable sets to predict giant panda habitat within and across time periods by using threshold-independent and threshold-dependent evaluation methods and five indices of temporal transferability. ResultsOur results showed that models developed with the four variable sets were all useful for characterizing and monitoring giant panda habitat. However, the models developed using multi-year data exhibited significantly higher temporal transferability than those developed using single-year data. In addition, models developed with phenology metrics, especially when using multi-year data, exhibited significantly higher temporal transferability than those developed with the time series.Main conclusions The integration of land surface phenology, captured by high temporal resolution remotely sensed imagery, with habitat modelling constitutes a suitable tool for characterizing wildlife habitat and monitoring its temporal dynamics. Using multi-year phenology metrics reduces model complexity, multicollinearity among predictor variables and variability caused by interannual climatic fluctuations, thereby increasing the temporal transferability of models. This study provides useful guidance for habitat monitoring through the integration of remote sensing technology and habitat modelling, which may be useful for the conservation of the giant panda and many other species.

Despite the near universal recognition that roads negatively affect wildlife, the mechanisms that elicit animal responses to roads are often ambiguous or poorly understood. We conducted a multi-year, multi-season study to assess the relative influence of roads on elk (Cervus elaphus) in a human-dominated landscape in South Dakota. We evaluated the effects of habitat covariates including security cover, forage quality, distance to roads (primary, secondary, and tertiary), and visibility from roads at the home range scale. We radio-collared 28 elk (21 adult females and 7 adult males) and calculated seasonal (winter, spring, summer, and autumn) utilization distributions (UDs). We assigned habitat covariates to use percentiles within the UDs (1% increments; from 1 to 98 percentiles) and used spatially explicit mixed linear regression to model the relationship between use percentile and habitat covariates. For each season and sex, we evaluated 15 candidate models and used Akaike's Information Criterion weights (v i ) to identify top-ranking models. We plotted influential coefficients from these models with 95% confidence intervals to examine the magnitude of effects. Our analysis revealed fundamental differences in response to roads, by road type, between sexes, and across seasons. Male elk established home ranges near roads devoid of vehicle traffic in winter, spring, and autumn. In summer, coinciding with peak vehicle traffic levels, male elk reduced their use of habitat that was both visible from and close to primary roads. Female elk subherds similarly responded to primary roads in spring and autumn, during times of year when they were calving and mating, respectively. In spring and summer, female elk subherds selected habitat near roads that were closed to vehicle traffic. Forage quality and security cover were influential in the periphery (>50th use percentile) of elk home ranges, whereas road covariates were more influential towards the core of elk home ranges. This analysis further demonstrates the utility of visibility from road metrics and suggests that the retention of vegetation structures that screen visibility potential from roads could be important components of elk management strategies. ß 2012 The Wildlife Society.

Telemetry data have been widely used to quantify wildlife habitat relationships despite the fact that these data are inherently imprecise. All telemetry data have positional error, and failure to account for that error can lead to incorrect predictions of wildlife resource use. Several techniques have been used to account for positional error in wildlife studies. These techniques have been described in the literature, but their ability to accurately characterize wildlife resource use has never been tested. We evaluated the performance of techniques commonly used for incorporating telemetry error into studies of wildlife resource use. Our evaluation was based on imprecise telemetry data (mean telemetry error = 174 m, SD = 130 m) typical of field‐based studies. We tested 5 techniques in 10 virtual environments and in one real‐world environment for categorical (i.e., habitat types) and continuous (i.e., distances or elevations) rasters. Technique accuracy varied by patch size for the categorical rasters, with higher accuracy as patch size increased. At the smallest patch size (1 ha), the technique that ignores error performed best on categorical data (0.31 and 0.30 accuracy for virtual and real data, respectively); however, as patch size increased the bivariate‐weighted technique performed better (0.56 accuracy at patch sizes >31 ha) and achieved complete accuracy (i.e., 1.00 accuracy) at smaller patch sizes (472 ha and 1,522 ha for virtual and real data, respectively) than any other technique. We quantified the accuracy of the continuous covariates using the mean absolute difference (MAD) in covariate value between true and estimated locations. We found that average MAD varied between 104 m (ignore telemetry error) and 140 m (rescale the covariate data) for our continuous covariate surfaces across virtual and real data sets. Techniques that rescale continuous covariate data or use a zonal mean on values within a telemetry error polygon were significantly less accurate than other techniques. Although the technique that ignored telemetry error performed best on categorical rasters with smaller average patch sizes (i.e., ≤31 ha) and on continuous rasters in our study, accuracy was so low that the utility of using point‐based approaches for quantifying resource use is questionable when telemetry data are imprecise, particularly for small‐patch habitat relationships.