Species-habitat Relationships Research Articles

The pitfalls of ignoring behaviour when quantifying habitat selection

AbstractAimHabitat selection is a behavioural mechanism by which animals attempt to maximize their inclusive fitness while balancing competing demands, such as finding food and rearing offspring while avoiding predation, in a heterogeneous and changing environment. Different habitat characteristics may be associated with each of these demands, implying that habitat selection varies depending on the behavioural motivations of the animal. Here, we investigate behaviour‐specific habitat selection in African elephants and discuss its implications for distribution modelling and conservation.LocationNorthern Botswana, Africa, case study.MethodsWe use Bayesian state‐space models to characterize location time series data of elephants into two behavioural states (encamped and exploratory). We then develop habitat selection models for each behavioural state and contrast them to models based on data pooled among behaviours.ResultsSpatial predictions of habitat use were often markedly different among the models. Behaviour‐specific and pooled habitat selection models differed in model structure, the magnitude of model coefficients and the form of the selection curve (linear or quadratic). Selection was typically strongest in the behaviour‐specific models, although this varied according to behavioural state and habitat covariate.Main conclusionsIgnoring behavioural states often had important consequences for quantifying habitat selection. Quantifying selection irrespective of behaviour (among all behaviours) can obscure important species–habitat relationships, thereby risking weak or incorrect inferences. Behaviour‐specific habitat selection provides greater insight into the process of habitat selection and can improve predictive habitat selection estimates. As some behaviours are more relevant to specific conservation objectives than others, focusing on behaviour‐specific selection could improve how habitats are prioritized for conservation or management.

A Multi-Scale Evaluation of the Effects of Forest Harvesting for Woody Biofuels on Mammalian Communities in a Northern Hardwood Forest

Research addressing the implications of forest harvesting for mammals has focused on different categories of silvicultural prescriptions. However, the effects of these prescriptions on forest structure can vary considerably, and categories of prescriptions rarely incorporate the market for which timber is being harvested. The latter information is important given the recent shift from conventional round-wood harvesting to wholetree removal for biofuels production, and corresponding reductions in post-harvest woody biomass left on-site. Our goal was to assess the effects of forest harvesting for biofuels on mammal species. Objectives included 1) evaluating how structural components influenced mammals, and 2) assessing the role of scale on species-habitat relationships. We sampled mammals in a 97-ha area of hardwood forest in the Adirondack Mountains in New York that had been partially harvested for biofuels in 2010. We used Sherman traps and track plates to assess the distribution of mammals. We captured 6 species of mammals in Sherman traps and identified 8 species using track plates. Mammalian species varied in their sensitivity to changes in habitat characteristics associated with biofuels harvest (coarse woody debris and slash). Our study reveals a complex suite of factors driving the response of mammals to variation in forest structure as a result of biofuels production. The harvesting practices used in the focal region are unlikely to lead to dramatic changes in the abundance and distribution of individual species of small mammals, but may influence the occurrence of common species including deer-mice and voles.

Setting Favourable Habitat Reference Values for breeding birds: general principles and examples for passerine birds

SummarySetting Favourable Reference Values (FRVs) can assist the definition of the conservation status of a species. FRVs may consider population, habitat, and range. FRVs can indicate a range of values for different parameters, which should allow the long-term persistence of a species/population. We propose a method for the definition of reference values for the habitat (FRV-H or HRV) of breeding bird species. HRV should cover habitat extent and quality, both required to ensure long-term persistence. Extent HRV should express a measure of suitable area, whereas quality HRV could be defined as the range of values for habitat variables known to affect habitat quality. To define an extent HRV, we built species distribution models (SDMs) and set extent HRV as the extent of potentially suitable habitat under a conservative approach. Quality HRV should refer to environmental determinants/correlates of occurrence and breeding success, and should be defined by the identification of the habitat factors affecting occurrence and reproduction. When habitat selection is adaptive, habitat suitability may approximate habitat quality, being correlated with breeding success. In that case, fine-scaled habitat/distribution models may be used to identify determinants/correlates of reproductive output, and such species-habitat relationships may help define quality HRV. We show examples using the Red-backed Shrike Lanius collurio as a model. The use of habitat selection models, which can be made spatially explicit generating distribution models, may assist the definition of both extension and quality HRVs. Species-habitat models can allow the individuation of factors and relative values affecting species occurrence/reproduction (quality HRV), and the definition of the spatial distribution and quantity of potentially suitable habitat (extent HRV). Our approach is one of the possible ones, aiming at finding a “suitable” trade-off between affordable data and scientific precision. HRVs should be used together with population and range FRVs to assess the status of a species/population.

Analysis of habitat determinants in contact zones of parapatricEuropean salamanders

AbstractParapatry is a remarkable distributional pattern where the ranges of two species come into contact but only narrowly overlap. Theory predicts and empirical data suggest that parapatric range margins are most likely to form along environmental gradients when there is interspecific competition. Here, we study the ecology of the narrow contact zones of two parapatricEuropean land salamanders,Salamandra salamandraandSalamandra atra. Previous research showed that abiotic conditions determine parapatric range margins of these two species. However, in contrast to other parapatric salamander species and theoretical predictions, there is no evidence for competitive interactions in the twoSalamandraspecies. To study whether competition restricts these species' ranges and to understand local syntopic co‐occurrence within their contact zones, we used site‐occupancy models (1) to assess species–habitat relationships and (2) to test whether there is evidence for competition. We found that the two salamanders show dissimilar species–habitat relationships. The slope of the site positively affected the site‐occupancy probability ofS. salamandra, while none of the habitat characteristics explained the occupancy probability ofS. atra. The local presence of one species had no effect on the occupancy probability of the other, suggesting that there is no effect of competition on local occurrence or that competition does not lead to spatial segregation. To fully understand the mechanisms that determine the parapatric range margins between the salamander species and to unravel the role of interspecific interactions, it is necessary to further study species' functional traits.

Modelling tree height–diameter relationships in multi-species and multi-layered forests: A large observational study from Northeast China

Projection of stand development over time relies on accurate height–diameter functions. In complex forests where trees of various ages, species, sizes, vigor and crown classes, and shade tolerance levels co-exist, tree height predictions are more difficult than in uniform conditions. In such complex situations, individual tree height–diameter relationships and predictions are often further complicated by structure (e.g., relative position of trees) and stand density which sometime are exacerbated by selective harvesting and other disturbances, and changing stand dynamics. Height–diameter (H–D) relationships are significantly modulated by microsite (elevation, aspect and slope), climate, and competition. Traditionally, tree height–diameter relationships are modeled or examined using linear or nonlinear regression models, in which the microsite influences or species–habitat relationships are largely ignored. Studies that have examined H–D relationships for multi-species and multi-layered forests using species–habitat relationships or microsite influences are limited or non-existent.Using an extensive dataset, we evaluated two sets of height–diameter equations for estimating tree heights in multi-species and multi-layered forests in north-eastern China. The first set included three equations for estimating height as a function of individual tree diameter, the second set included nine equations for estimating height as a function of individual tree diameter, relative tree position, and stand density variables. The inclusion of relative position of a tree and stand density into the base height–diameter equations increased the accuracy of prediction for all species. On the average, root mean square values were reduced by 12.2% by including these variables. After comparing and selecting a base H–D equation, we evaluated its statistical attributes among habitat types constructed for forest a large observational study site in north-eastern China.Nonlinear mixed effect (NLMEM) was used to account for the hierarchical data structure and to provide predictive equations. Dummy variables were used to indicate species effect and allow simultaneous estimation of all species. The NLMEM framework along with species effect made tree height predictions compatible for 23 tree species found in multi-layered forests in NE China. Model coefficients for selected equations are documented for future use.

Movement patterns of an arboreal marsupial at the edge of its range: a case study of the koala.

BackgroundConservation strategies derived from research carried out in one part of the range of a widely distributed species and then uniformly applied over multiple regions risk being ineffective due to regional variations in species-habitat relationships. This is particularly true at the edge of the range where information on animal movements and resource selection is often limited. Here, we investigate home range size, movement patterns and resource selection of koalas Phascolarctos cinereus in the semi-arid and arid landscapes of southwest Queensland, Australia. We placed collars with GPS units on 21 koalas in three biogeographic regions. Home range sizes, resource selection and movement patterns were examined across the three regions.ResultsHabitat selectivity was highest at the more arid, western edge of the koala’s range with their occupancy restricted to riparian/drainage line habitats, while the more easterly koalas displayed more variability in habitat use. There was no significant difference between home range sizes of koalas at the western edge of the range compared to the more easterly koalas. Instead, variability in home range size was attributed to spatial variations in habitat quality or the availability of a key resource, with a strong influence of rainfall and the presence of freestanding water on the home range size of koalas. Within a 580 m spatial range, movement patterns of male and female paths showed a tortuous trend, consistent with foraging behavior. Beyond this spatial range, male paths showed a trend to more linear patterns, representing a transition of movement behavior from foraging to breeding and dispersal.ConclusionsThe difference in home range movement patterns and resource use among the different koala populations shows that behavior changes with proximity to the arid edge of the koala’s range. Changes in home range size and resource use near the range edge highlight the importance of further range-edge studies for informing effective koala conservation and management actions, especially when developing species-specific adaptation responses to climate change.Electronic supplementary materialThe online version of this article (doi:10.1186/2051-3933-1-8) contains supplementary material, which is available to authorized users.

Occupancy patterns of large mammals in the Far North of Ontario under imperfect detection and spatial autocorrelation

AbstractAimAn understanding of the factors that influence species distributions in heterogeneous landscapes is important when making decisions regarding conservation. Moreover, occupancy probabilities based on detection data can reveal important species–habitat relationships. Accounting for the spatial autocorrelation of detection data increases the statistical validity of occupancy models, but is not often considered. Using novel occupancy modelling that explicitly incorporates detectability and spatial autocorrelation, we assessed the influence of habitat on occupancy patterns of woodland caribou (Rangifer tarandus caribou), moose (Alces alces) and wolves (Canis lupus) across a broad biogeographical extent where fire is the dominant agent of disturbance.LocationNorthern Ontario, Canada.MethodsWe aerially surveyed 3851 sampling units, each covering 100 km2, for woodland caribou, moose and wolves in February–March in 2009, 2010 and 2011, and visited 1663 units more than once to estimate detectability. We used restricted spatial regression to model occupancy probabilities of each species with respect to habitat factors in two ecozones, accounting for both imperfect detection and lack of independence of sampling units.ResultsCovariates influencing species detection varied among ecozones and species. Caribou occupancy was positively related to bogs and negatively related to disturbed areas, while moose occupancy showed opposite responses to these covariates. Wolf occupancy was related to high prey occupancy. Explicitly accounting for spatial autocorrelation in detection data reduced the chance of type I error in occupancy estimates compared with non‐spatial models.Main conclusionsHabitat relationships and occupancy patterns support the hypothesis that caribou remain spatially segregated from moose to reduce predation risk. The broad scale of analysis indicated changes in species–habitat relationships, suggesting that limiting factors vary across biogeographical gradients. The spatial pattern in caribou occupancy allowed us to identify important areas used by caribou across the region, including the ecotone between fire‐driven boreal forests and peatland complexes. The evidence for significant relationships between caribou and land cover, predators and alternate prey underscores the need for careful planning of development and infrastructure in the area.

Distribution and environmental relationships of three species of wolffish (Anarhichas spp.) in the Gulf of St. Lawrence

AbstractThis study examines the spatial distribution of three species of wolffish in the Gulf of St. Lawrence on the basis of trawl surveys.A standardized method is proposed to assess species–habitat associations for the purpose of management and recovery of endangered marine species.Catch data (presence/absence in trawl sets) and landscape and environmental characteristics of the sea floor were aggregated using a common grid (100 km2 cells), and species–habitat relationships were explored using geospatial tools.Relative occurrence was lower and area of occupancy and concentration were much smaller for the northern wolffish (Anarhichas denticulatus) than for the spotted wolffish (Anarhichas minor), with the striped wolffish (Anarhichas lupus) being most widespread. Significant relationships were observed between values of the local spatial autocorrelation Gi* statistic and habitat descriptors for each of the three species.Hot spots for spotted and striped wolffish occurred in areas where a greater diversity of relief and habitats was found. They were associated with intermediate depths, coarse sediments and rock outcrops, and lower salinities and temperatures than for northern wolffish.Northern wolffish appeared to be associated mainly with the deep water sloped habitat bordering deep channels, whereas spotted and striped wolffish both concentrated most intensively into neighbouring deep water shelf habitats and relatively cold shallow to mid‐depth shelf habitats of the northern Gulf.The RDA analysis indicated a significant relationship between Gi* scores of the three species and environmental variables. The model explained 52% of the variability in the data; northern wolffish showed a distinct relationship compared with the two other species.The conservation of marine species and protection of their habitats pose a major challenge given the limited amount of information typically available on rare species and the scale and complexity of marine processes that affect those species. The broad‐scale approach presented here allows the provision of advice on important habitats on the basis of the best available knowledge. Copyright © 2013 Her Majesty the Queen in Right of Canada

A spatial framework for representing nearshore ecosystems

The shallow, coastal regions of the world’s oceans are highly productive ecosystems providing important habitat for commercial, forage, endangered, and iconic species. Given the diversity of ecosystem services produced or supported by this ecosystem, a better understanding of its structure and function is central to developing an ecosystem-based approach to management. However this region – termed the ‘white strip’ by marine geologists because of the general lack of high-resolution bathymetric data – is dynamic, highly variable, and difficult to access making data collection challenging and expensive. Since substrate is a key indicator of habitat in this important ecosystem, our objective was to create a continuous substrate map from the best available bottom type data. Such data are critical to assessments of species distributions and anthropogenic risk. Using the Strait of Georgia in coastal British Columbia, Canada, as a case study, we demonstrate how such a map can be created from a diversity of sources. Our approach is simple, quantitative, and transparent making it amenable to iterative improvement as data quality and availability improve. We evaluated the ecological performance of our bottom patches using observed shellfish distributions. We found that observations of geoduck clam, an infaunal species, and red urchins, a species preferentially associated with hard bottom, were strongly and significantly associated with our soft and hard patches respectively. Our description of bottom patches also corresponded well with a more traditional, morphological classification of a portion of the study area. To provide subsequent analyses (such as habitat models) with some confidence in the defined bottom type values, we developed a corresponding confidence surface based on the agreement of, and distance between observations. Our continuous map of nearshore bottom patches thus provides a spatial framework to which other types of data, both abiotic (e.g., energy) and biotic, can be attached. As more data are associated with the bottom patches, we anticipate they will become increasingly useful for representing and developing species-habitat relationships, ultimately leading to a comprehensive representation of the nearshore ecosystem.

Changes in Song ThrushTurdus philomelosDensity and Habitat Association in Apple Orchards During the Breeding Season

Within-season changes in habitat and distribution of breeding birds have been poorly investigated but are likely to be important for several species. With this work, we investigated what factors affect the density of a multi-brooded bird species breeding in apple orchards, the song thrush Turdus philomelos, in three different periods of the breeding season in the Trento province, northern Italy. Birds were counted along transects, and analyses carried out using a Generalised Mixed Linear Model (GLMM) approach. A further analysis assessed the effect of conspecific occurrence and habitat variables on the changes in local density. Thrush density changed during the season, increasing from early to late period, and was mainly affected by tree height (positively), bare soil (negatively in third period), shrubland (negatively), wetland (positively), grassland (negatively), percentage of rows irrigated by droplet system (negatively), percentage of orchard covered by anti-hail nets (negatively), urbanised area (negatively in first period, positively in third period). Intra-seasonal variation in density was affected by tree height (positively), anti-hail nets (negatively), use of herbicide during the late period (negatively), difference in bare ground (negatively), and initial thrush density (negatively), the latter suggesting that best habitats are occupied first. Multi-brooded bird species may show varying density and changing species-habitat relationships during the breeding season, and their distribution could change during the period of reproduction. This pattern could apply to a large number of bird species, requires deeper investigation and could have important implications for conservation and management.

Modeling ecological minimum requirements for distribution of greater sage‐grouse leks: implications for population connectivity across their western range, U.S.A

Greater sage-grouse Centrocercus urophasianus (Bonaparte) currently occupy approximately half of their historical distribution across western North America. Sage-grouse are a candidate for endangered species listing due to habitat and population fragmentation coupled with inadequate regulation to control development in critical areas. Conservation planning would benefit from accurate maps delineating required habitats and movement corridors. However, developing a species distribution model that incorporates the diversity of habitats used by sage-grouse across their widespread distribution has statistical and logistical challenges. We first identified the ecological minimums limiting sage-grouse, mapped similarity to the multivariate set of minimums, and delineated connectivity across a 920,000 km2 region. We partitioned a Mahalanobis D2 model of habitat use into k separate additive components each representing independent combinations of species–habitat relationships to identify the ecological minimums required by sage-grouse. We constructed the model from abiotic, land cover, and anthropogenic variables measured at leks (breeding) and surrounding areas within 5 km. We evaluated model partitions using a random subset of leks and historic locations and selected D2 (k = 10) for mapping a habitat similarity index (HSI). Finally, we delineated connectivity by converting the mapped HSI to a resistance surface. Sage-grouse required sagebrush-dominated landscapes containing minimal levels of human land use. Sage-grouse used relatively arid regions characterized by shallow slopes, even terrain, and low amounts of forest, grassland, and agriculture in the surrounding landscape. Most populations were interconnected although several outlying populations were isolated because of distance or lack of habitat corridors for exchange. Land management agencies currently are revising land-use plans and designating critical habitat to conserve sage-grouse and avoid endangered species listing. Our results identifying attributes important for delineating habitats or modeling connectivity will facilitate conservation and management of landscapes important for supporting current and future sage-grouse populations.

Scale dependence in habitat selection: the case of the endangered brown bear (Ursus arctos) in the Cantabrian Range (NW Spain)

Animals select habitat resources at multiple spatial scales. Thus, explicit attention to scale dependency in species–habitat relationships is critical to understand the habitat suitability patterns as perceived by organisms in complex landscapes. Identification of the scales at which particular environmental variables influence habitat selection may be as important as the selection of variables themselves. In this study, we combined bivariate scaling and Maximum entropy (Maxent) modeling to investigate multiscale habitat selection of endangered brown bear (Ursus arctos) populations in northwest Spain. Bivariate scaling showed that the strength of apparent habitat relationships was highly sensitive to the scale at which predictor variables are evaluated. Maxent models on the optimal scale for each variable suggested that landscape composition together with human disturbances was dominant drivers of bear habitat selection, while habitat configuration and edge effects were substantially less influential. We found that explicitly optimizing the scale of habitat suitability models considerably improved single-scale modeling in terms of model performance and spatial prediction. We found that patterns of brown bear habitat suitability represent the cumulative influence of habitat selection across a broad range of scales, from local resources within habitat patches to the landscape composition at broader spatial scales.

Habitat association of Little Grebe (Tachybaptus ruficollis) at Kallar Kahar Lake, Pakistan

Understanding of species–habitat relationship is fundamental to effective conservation planning and management. We studied the association of Little Grebe abundance with habitat parameters at Kallar Kahar Lake, District Chakwal, Punjab, Pakistan, from September 2010 to July 2011. Bird population density data, vegetation and physico-chemical parameters were collected along three strips from different areas of the lake based on habitat variability. The association of Little Grebe abundance with physico-chemical parameters of water was determined using simple linear regression. The physico-chemical parameters differed significantly (P<0.05) among the three strips. Water temperature and turbidity were negatively related while pH and depth were positively related with grebe population. We concluded that the species more frequently occurred in shallow water habitats with preponderance of reed vegetation, slightly high alkalinity, low water temperature and low turbidity. The reed vegetation such as Phragmites and Typha provided shelter and nesting sites while open water had abundant planktons and crustaceans, algae such as Spirogyra spp. and submerged vegetation such as Chara spp. provided foraging sites. Threats to the species included eutrophication, fluctuation in the water level and littering by visitors which need to be addressed through appropriate management interventions.

Assessing Regional and Interspecific Variation in Threshold Responses of Forest Breeding Birds through Broad Scale Analyses

BackgroundIdentifying persistence and extinction thresholds in species-habitat relationships is a major focal point of ecological research and conservation. However, one major concern regarding the incorporation of threshold analyses in conservation is the lack of knowledge on the generality and transferability of results across species and regions. We present a multi-region, multi-species approach of modeling threshold responses, which we use to investigate whether threshold effects are similar across species and regions.Methodology/Principal FindingsWe modeled local persistence and extinction dynamics of 25 forest-associated breeding birds based on detection/non-detection data, which were derived from repeated breeding bird atlases for the state of Vermont. We did not find threshold responses to be particularly well-supported, with 9 species supporting extinction thresholds and 5 supporting persistence thresholds. This contrasts with a previous study based on breeding bird atlas data from adjacent New York State, which showed that most species support persistence and extinction threshold models (15 and 22 of 25 study species respectively). In addition, species that supported a threshold model in both states had associated average threshold estimates of 61.41% (SE = 6.11, persistence) and 66.45% (SE = 9.15, extinction) in New York, compared to 51.08% (SE = 10.60, persistence) and 73.67% (SE = 5.70, extinction) in Vermont. Across species, thresholds were found at 19.45–87.96% forest cover for persistence and 50.82–91.02% for extinction dynamics.Conclusions/SignificanceThrough an approach that allows for broad-scale comparisons of threshold responses, we show that species vary in their threshold responses with regard to habitat amount, and that differences between even nearby regions can be pronounced. We present both ecological and methodological factors that may contribute to the different model results, but propose that regardless of the reasons behind these differences, our results merit a warning that threshold values cannot simply be transferred across regions or interpreted as clear-cut targets for ecosystem management and conservation.

Modeling the distribution of wild rabbits (Oryctolagus cuniculus) on a Mediterranean island

AbstractThe European rabbit (Oryctolagus cuniculus) is a widely distributed mammal with an often contradictory ecological role, imposing the need for population management. Sound management requires an in‐depth understanding of the complex species–habitat relationships. In this study, CART analysis was employed to identify the most important environmental and anthropogenic factors affecting the spatial distribution of wild rabbit on Lemnos Island, northeastern Aegean Sea in Greece. On Lemnos, this species is considered an agricultural pest due to its overabundance resulting from the long‐term absence of viral diseases, limited predation pressure, and lack of effective management. The study was carried out during the summer of 2008 by surveying rabbit densities in 181 2 × 2‐km squares. Seven environmental and 14 anthropogenic variables, measured at two spatial scales, were used as explanatory variables. Soil hardness was the most influential variable, dividing the island into two distinct areas, namely the rabbit‐poor areas with hard rocky soils and the rabbit‐rich areas where soft soils prevail. In the former, the presence of a sharp relief can lead to complete absence of the species, while a combination of gentle relief, low altitudes, and low presence of arable land can lead to moderate rabbit density. In the latter, human‐caused disturbance can reduce the number of rabbits, while a high density of ecotones and streams and a high presence of riparian vegetation can increase population densities to its highest levels observed. Our findings can formulate a scientific basis for the development of an effective management strategy for its population control.

A cost-effective and informative method of GPS tracking wildlife

In wildlife research, our ability to GPS track sufficient numbers of individuals is always limited by cost, which restricts inference of species–habitat relationships. Here, we describe the modification and use of a relatively new and inexpensive off-the-shelf GPS device, to provide detailed and accurate information on the movement patterns of individuals (mountain brushtail possums, Trichosurus cunninghami), including how movement varies through time, and how individuals interact with each other. Our results demonstrated that this technology has enormous potential to contribute to an improved understanding of the movement patterns and habitat preferences of wildlife at a fraction of the cost of traditional GPS technology.

Effects of spatial scale and choice of statistical model (linear versus tree-based) on determining species–habitat relationships

The choice of spatial scale and modelling technique used to capture species–habitat relationships needs to be considered when ascertaining environmental determinants of habitat quality for species and communities. Fish densities and environmental data were collected at three Laurentian lakes using underwater surveys by four snorkelers collecting fine spatial data acquired through geographic positioning systems. At both fine (20 m) and broad (100 m) spatial scales, tree-based approaches, which incorporated both linear and nonlinear relationships, explained more variation than their linear counterparts. At the finest spatial scale considered (20 m), local environmental conditions, such as habitat structure and heterogeneity, were important determinants of fish habitat selection. At the broadest spatial scale considered (100 m), fish tended to select habitat based on both local environmental features and riparian development. Moran’s eigenvector maps further revealed that fish–habitat associations were operating at broader spatial scales than the predefined analytical units, which can be partially attributed to the spatial structure of environmental conditions acting at spatial scales greater than 100 m. This study highlights the importance of evaluating statistical approaches at different spatial scales to identify key determinants of habitat quality for species, ultimately to assess the effects of perturbations on ecosystems.

Measuring and selecting scales of effect for landscape predictors in species–habitat models

Wildlife managers often use habitat models to determine species habitat requirements and to identify locations for conservation efforts, uses which depend on accurate specification of species-habitat relationships. Prior study suggests that model performance may be influenced by the way we measure environmental predictors. We hypothesized that species responses to landscape predictors are best represented by landscape composition-based measurements, rather than distance-based measurements. We also hypothesized that models using empirical data to select an appropriate scale of effect for each habitat predictor (multi-scale models) should perform better than models that assume a common scale of effect for all predictors (single-scale models). To test these hypotheses we constructed habitat models for three mammal species, Mephitis mephitis, Mustela erminea, and Procyon lotor, based on surveys conducted in 80 landscapes in southeastern Ontario, Canada. For each species we compared the performance of distance- and composition-based measurements, and we compared the performance of single- and multi-scale models. The composition-based measurement, measured at its empirically determined scale of effect, had greater explanatory power than the distance-based measurement of a given predictor more often than expected by chance, supporting our first hypothesis. Contrary to expectation, multi-scale models did not have better explanatory power or predictive performance relative to single-scale models. We identified and evaluated four potential mechanisms to explain this, and, depending on the species, we found that the best explanation was either that predictors have significant effects at a common scale or that, although the modeled effects were at multiple scales, they were of similar magnitude and direction at the scales modeled in single- and multi-scale models. Our results suggest that habitat modeling based on distance-based measurements could be improved by including composition-based measurements of landscape predictor variables, but that inclusion of predictor-specific scales of effect for composition-based measurements does not necessarily improve performance over models including composition-based measurements at a single scale. Conservation and wildlife management may be simplified when single-scale models perform as well as multi-scale models, as this suggests actions conducted at a single scale may address management objectives as well as actions taken at different scales for different landscape features.

Evaluating the predictive abilities of community occupancy models using AUC while accounting for imperfect detection

The ability to accurately predict patterns of species' occurrences is fundamental to the successful management of animal communities. To determine optimal management strategies, it is essential to understand species-habitat relationships and how species habitat use is related to natural or human-induced environmental changes. Using five years of monitoring data in the Chesapeake and Ohio Canal National Historical Park, Maryland, USA, we developed four multispecies hierarchical models for estimating amphibian wetland use that account for imperfect detection during sampling. The models were designed to determine which factors (wetland habitat characteristics, annual trend effects, spring/summer precipitation, and previous wetland occupancy) were most important for predicting future habitat use. We used the models to make predictions about species occurrences in sampled and unsampled wetlands and evaluated model projections using additional data. Using a Bayesian approach, we calculated a posterior distribution of receiver operating characteristic area under the curve (ROC AUC) values, which allowed us to explicitly quantify the uncertainty in the quality of our predictions and to account for false negatives in the evaluation data set. We found that wetland hydroperiod (the length of time that a wetland holds water), as well as the occurrence state in the prior year, were generally the most important factors in determining occupancy. The model with habitat-only covariates predicted species occurrences well; however, knowledge of wetland use in the previous year significantly improved predictive ability at the community level and for two of 12 species/species complexes. Our results demonstrate the utility of multispecies models for understanding which factors affect species habitat use of an entire community (of species) and provide an improved methodology using AUC that is helpful for quantifying the uncertainty in model predictions while explicitly accounting for detection biases.

Species conservation at a broad spatial scale: Reproductive success of golden‐cheeked warblers across their breeding range

AbstractResearch conducted near the time a species is listed as threatened or endangered can help inform management guidelines. However, these studies are typically limited in temporal and spatial scope and typically address a limited set of ecological questions. Application of such information can result in misleading management guidelines or ineffective conservation if species–habitat relationships vary across space or if preliminary information fails to provide reliable assessments of habitat quality. We evaluated existing management guidelines that addressed reproductive success of the golden‐cheeked warbler (Setophaga chrysoparia), a federally endangered songbird endemic to central Texas, USA. We monitored 1,382 territories in 87 patches of woodland across the breeding range from 2007 to 2011. Ours was the first evaluation of management guidelines addressing reproductive success with data collected across the warbler's breeding range. Our results did not consistently confirm the commonly accepted management guidelines that habitat quality, based on the presence of offspring in a territory, was positively associated with patch size and territory canopy cover, and negatively associated with patch edge. The relationship between probability of a territory fledging ≥1 offspring and patch size, patch edge‐to‐area ratio, and territory canopy cover varied by geographic region (level‐IV eco‐region). Our results suggested future conservation plans that address habitat quality should consider variation among regions to improve conservation efforts and that habitat associations identified with warbler occurrence were not necessarily indicative of reproductive success. © 2012 The Wildlife Society.