No abstract

Remote camera traps are often used in large-mammal research and monitoring programs because they are cost-effective, allow for repeat surveys, and can be deployed for long time periods. Statistical advancements in calculating population densities from camera-trap data have increased the popularity of camera usage in mammal studies. However, drawbacks to camera traps include their limited sampling area and tendency for animals to notice the devices. In contrast, autonomous recording units (ARUs) record the sounds of animals with a much larger sampling area but are dependent on animals producing detectable vocalizations. In this study, we compared estimates of occupancy and detectability between ARUs and remote cameras for gray wolves (Canis lupus Linnaeus, 1758) in northern Alberta, Canada. We found ARUs to be comparable with cameras in their detectability and occupancy of wolves, despite only operating for 3% of the time that cameras were active. However, combining cameras and ARUs resulted in the highest detection probabilities for wolves. These advances in survey technology and statistical methods provide innovative avenues for large-mammal monitoring that, when combined, can be applied to a broad spectrum of conservation and management questions, provided assumptions for these methods are rigorously tested and met.

Estimating the population abundance of landbirds is a challenging task complicated by the amount, type, and quality of available data. Avian conservationists have relied on population estimates from Partners in Flight (PIF), which primarily uses roadside data from the North American Breeding Bird Survey (BBS). However, the BBS was not designed to estimate population sizes. Therefore, we set out to compare the PIF approach with spatially explicit models incorporating roadside and off-road point-count surveys. We calculated population estimates for 81 landbird species in Bird Conservation Region 6 in Alberta, Canada, using land cover and climate as predictors. We also developed a framework to evaluate how the differences between the detection distance, time-of-day, roadside count, and habitat representation adjustments explain discrepancies between the 2 estimators. We showed that the key assumptions of the PIF population estimator were commonly violated in this region, and that the 2 approaches provided different population estimates for most species. The average differences between estimators were explained by differences in the detection-distance and time-of-day components, but these adjustments left much unexplained variation among species. Differences in the roadside count and habitat representation components explained most of the among-species variation. The variation caused by these factors was large enough to change the population ranking of the species. The roadside count bias needs serious attention when roadside surveys are used to extrapolate over off-road areas. Habitat representation bias is likely prevalent in regions sparsely and non-representatively sampled by roadside surveys, such as the boreal region of North America, and thus population estimates for these regions need to be treated with caution for certain species. Additional sampling and integrated modeling of available data sources can contribute towards more accurate population estimates for conservation in remote areas of North America.

One of the continuing challenges in wildlife ecology and management is the ability to obtain reliable estimates of species’ distributions at large spatial extents. Multi‐scale occupancy models using a cluster sampling design offer the opportunity to increase the resolution of estimates and model processes occurring at multiple spatial scales, increasing the efficiency of large‐scale monitoring and mitigating the tradeoff between extent and grain. However, accounting for spatial correlation among subsamples in a way that allows for the addition of covariates remains an issue. Using tracking transect surveys for carnivores as an example, we describe and evaluate a hierarchical, multi‐scale occupancy model that integrates existing approaches to estimate occupancy at multiple spatial scales simultaneously, and uses a conditional autoregressive (CAR) process to account for spatial correlation in use between subsamples. We evaluated 3 versions of the model under a single‐survey and a multi‐survey sampling design: a non‐spatial model, a model that accounted for spatial correlation in use between transect segments, and a model that also accounted for spatial correlation in the detection process. Simulations showed that accounting for spatial correlation gave better estimates of transect‐level occupancy under both sampling designs, whereas accurate estimates of segment‐level use required a multi‐survey design. When applied to historical snow track data, the differences in estimates among models followed the same pattern found in the simulations. The multi‐survey design was able to detect equivalent declines in segment use with much less survey effort than the single‐survey design. The modeling framework presented here offers researchers and managers a powerful tool for monitoring populations at large spatial extents while being able to detect ecologically important dynamics at finer spatial scales. © 2018 The Wildlife Society.

Estimating distribution and abundance of species depends on the probability at which individuals are detected. Butterflies are of conservation interest worldwide, but data collected with Pollard walks—the standard for national monitoring schemes—are often analyzed assuming that changes in detectability are negligible within recommended sampling criteria. The implications of this practice remain poorly understood. Here, we evaluated the effects of sampling conditions on butterfly counts from Pollard walks using the Arctic fritillary, a common but cryptic butterfly in boreal forests of Alberta, Canada. We used an open population binomial N‐mixture model to disentangle the effects of habitat suitability and phenology on abundance of Arctic fritillaries, and its detectability by sampling different conditions of temperature, wind, cloud cover, and hour of the day. Detectability varied by one order of magnitude within the criteria recommended for Pollard walks (P varying between 0.04 and 0.45), and simulations show how sampling in suboptimal conditions increases substantially the risk of false‐absence records (e.g., false‐absences are twice as likely than true‐presences when sampling 10 Arctic fritillaries at P = 0.04). Our results suggest that the risk of false‐absences is highest for species that are poorly detectable, low in abundance, and with short flight periods. Analysis with open population binomial N‐mixture models could improve estimates of abundance and distribution for rare species of conservation interest, while providing a powerful method for assessing butterfly phenology, abundance, and behavior using counts from Pollard walks, but require more intensive sampling than conventional monitoring schemes.

In response to the decline of northern bobwhite (Colinus virginianus; hereafter, bobwhite) in eastern Oklahoma, USA, a cost-share incentive program for private landowners was initiated to restore early successional habitat. Our objectives were to determine whether the program had an effect on bobwhite occupancy in the restoration areas and evaluate how local-and landscape-level habitat characteristics affect occupancy in both restoration and control areas. We surveyed 14 sample units that received treatment between 2009 and 2011, and 17 sample units that were controls. We used single-season occupancy models, with year as a dummy variable, to test for an effect of restoration treatment and habitat variables on occupancy. We found no significant treatment effect. Model selection showed that occupancy was best explained by the combination of overstory canopy cover and habitat area at both the local and landscape scales. Moran's I revealed positive spatial autocorrelation in the 1,000-3,000-m distance band, indicating that the likelihood of bobwhite occupancy increased with proximity to other populations. We show that creating !20 ha of habitat within 1-3 km of existing bobwhite populations increases the chance of restoration being successful. Published 2013. This article is a U.S. Government work and is in the public domain in the USA.

Aim Most large‐scale species distribution models assume spatially constant habitat selection throughout a species' geographic range. However, there is evidence this assumption may not be valid for a number of boreal bird species, which could lead to biased predictions of density and distribution in range‐wide models. Our goal was to test for and quantify differential habitat selection (DHS) in songbirds among regions of the Canadian boreal forest. Location Northern Alberta, western Ontario and southern Quebec, Canada. Methods We used hierarchical analysis of covariance models with region‐specific parameter estimates to test for differential selection of forest attributes among three regions for six boreal bird species. We used the results of these models to quantify intraspecific niche overlap between regions and compared posterior predictive accuracy to models that did not account for DHS. Results We found a generally large standardized effect size (median effect size = 1.674) of region on selection of specific habitat variables for all six species, although there was high variability among species, variables and regional comparisons. The proportion of niche overlap between regions was generally low (mean overlap = 0.309 for all pairwise comparisons), with no spatial pattern to the overlap. Models accounting for DHS had significantly higher posterior predictive accuracy according to the Watanabe–Akaike information criterion. Main Conclusions We found strong evidence for DHS among regions for six boreal songbird species in individual habitat attributes and overall niche space. The higher predictive accuracy of our DHS models suggests that failure to account for spatial variability in habitat selection can lead to biased estimates of density and spatial distribution. Models that did not account for DHS overestimated density relative to DHS models. We conclude that large‐scale species distribution models should account for regional variation in habitat selection in order to obtain accurate estimates of population size and distribution.

No abstract