Resource Selection Function Models Research Articles

Resource selection by Sarda cattle in a Mediterranean silvopastoral system.

Knowledge of how grazing cattle utilize heterogeneous landscapes in Mediterranean silvopastoral areas is scarce. Global positioning systems (GPS) to track animals, together with geographic information systems (GIS), can relate animal distribution to landscape features. With the aim to develop a general spatial model that provides accurate prediction of cattle resource selection patterns within a Mediterranean mountainous silvopastoral area, free-roaming Sarda cows were fitted with GPS collars to track their spatial behaviors. Resource selection function models (RSF) were developed to estimate the probability of resource use as a function of environmental variables. A set of over 500 candidate RSF models, composed of up to five environmental predictor variables, were fitted to data. To identify a final model providing a robust prediction of cattle resource selection pattern across the different seasons, the 10 best models (ranked on the basis of the AIC score) were fitted to seasonal data. Prediction performance of the models was evaluated with a Spearman correlation analysis using the GPS position data sets previously reserved for model validation. The final model emphasized that watering point, elevation, and distance to fences were important factors affecting cattle resource-selection patterns. The prediction performances (as Spearman rank correlation scores) of the final model, when fitted to each season, ranged between 0.7 and 0.94. The cows were more likely to select areas lower in elevation and farther from the watering point in winter than in summer (693 ± 1 m and 847 ± 13 m vs. 707 ± 1 m and 635 ± 21 m, respectively), and in spring opted for the areas furthest from the water (963 ± 12). Although caution should be exercised in generalizing to other silvopastoral areas, the satisfactory Spearman correlations scores from the final RSF model applied to different seasons indicate resource selection function is a powerful predictive model. The relative importance of the individual predictors within the model varied among the different seasons, demonstrating the RSF model's ability to interpret changes in animal behavior at different times of the year. The RSF model has proven to be a useful tool to interpret the spatial behaviors of cows grazing in Mediterranean silvopastoral areas and could therefore be helpful in managing and preserving ecosystem services of these areas.

Leveraging limited data from wildlife monitoring in a conflict affected region in Venezuela

Efficient monitoring of biodiversity-rich areas in conflict-affected areas with poor rule of law requires a combination of different analytical approaches to account for data biases and incompleteness. In the upland Amazon region of Venezuela, in Canaima National Park, we initiated biodiversity monitoring in 2015, but it was interrupted by the establishment of a large-scale mining development plan in 2016, compromising the temporal and geographical extent of monitoring and the security of researchers. We used a resource selection function model framework that considers imperfect detectability and supplemented detections from camera trap surveys with opportunistic off-camera records (including animal tracks and direct sighting) to (1) gain insight into the value of additional occurrence records to accurately predict wildlife resource use in the perturbated area (deforestation, fire, swidden agriculture, and human settlements vicinity), (2) when faced with security and budget constraints. Our approach maximized the use of available data and accounted for biases and data gaps. Adding data from poorly sampled areas had mixed results on estimates of resource use for restricted species, but improved predictions for widespread species. If budget or resources are limited, we recommend focusing on one location with both on-camera and off-camera records over two with cameras. Combining camera trap records with other field observations (28 mammals and 16 birds) allowed us to understand responses of 17 species to deforestation, 15 to fire, and 13 to swidden agriculture. Our study encourages the use of combinations of methods to support conservation in high-biodiversity sites, where access is restricted, researchers are vulnerable, and unequal sampling efforts exist.

Resource Selection Function-Adjusted Carrying Capacity Informs Bison Conservation Management in the Imperilled Mixed Grassland Ecosystem

In Canada, plains bison (Bison bison bison) was assessed by The Committee on the Status of Endangered Wildlife in Canada as threatened. While bison are no longer at risk of demographic extinction, conservation programs remain challenged by the rarity of large populations and most bison are found in small, isolated, and confined herds. In this context, proper assessment of ecological carrying capacity is critical to inform habitat management and conservation efforts for species recovery. Although estimated food-limited carrying capacity is influenced by forage availability, forage requirements, and offtake proportion, it should also consider habitat selection by animals, especially inside confined settings to help avoid overgrazing. To support bison management in Grasslands National Park of Canada, we integrated remote sensing, geographic information systems, and resource selection functions (RSFs) to examine variables that were potentially associated with bison habitat selection and return a comprehensive estimate of bison carrying capacity. Relevant variables were then integrated with estimates of forage availability using remote sensing and extrapolated to the bison containment scale. Factors of relevance to the RSFs include vegetation landscape units, slope, distance to water, fence, and road. In particular, bison selected for upland and sloped grasslands, which were characterized by the highest forage availability (1 064.5 kg ha−1 and 1 238.5 kg ha−1), while avoiding water in both growing and dormant seasons. The top-performing RSFs models in growing and dormant seasons were assessed using k-fold cross validation and achieved good predictive capacity (Spearman rank correlation [rs] ≥ 0.83, P < 0.01). Application of traditional clipping biomass samples and remote sensing derived variables is helpful in estimating annual forage quantity for bison (R2 = 0.75, P < 0.05). When accounting for bison resource selection, our model resulted in a carrying capacity estimate of about 0.0424 bison ha−1 or 764−770 bison (each requiring 12.2 kg forage/d), compared with 0.0587 bison ha−1 or 1 062 animals estimated when considering only the availability of forage. Such potential to increase carrying capacity to more than 1 000 individuals stresses the importance of testing management tools to shape bison grazing and increase forage utilization across a greater proportion of the 18 000-ha containment area. By improving our understanding of the interaction of this species with the mixed-grass prairie ecosystem, the integration of RSFs with estimates of carrying capacity can help inform conservation management of bison and multi-species at-risk habitat.

Opportunistic citizen science data reveals habitat selection of lesser black‐backed gulls in Central Europe

AbstractThis study is the first attempt to find a relation between the opportunistic observation data of an easily identifiable but locally rare bird, in our case the lesser black‐backed gull (Larus fuscus fuscus), and the Corine Land Cover habitat map by applying a resource selection function model to reveal the habitat selection of the species. We provide a scientific analysis using the valuable citizen science dataset sourced from the www.birding.hu website, collected by hundreds of volunteers over a span of more than a decade in Hungary. Birds selected for landfills, river and lake habitats and showed a significant but much smaller attraction towards urban areas, while they selected against arable lands. We found that in the case of rare and moderately common birds, we can obtain a representative picture of the habitat selection of the species, even if the data are collected by non‐standardized means.

Habitat Selection of Female Rio Grande Wild Turkey Influenced by Vegetation and Anthropogenic Structures

The Rio Grande wild turkey (Meleagris gallopavo intermedia) is a popular game species throughout the southcentral United States. It has experienced recent population declines in some areas of its distribution, though the cause is not clear. To better understand Rio Grande wild turkey (hereafter, turkey) ecology and inform management, we attached Global Positioning System transmitters to 34 females in 2016 and 2017. We evaluated habitat selection at both the second-order (landscape) and third-order (home range) scale for the breeding season and second-order selection for the nonbreeding season. Our resource selection function (RSF) models evaluated vegetation, management, and anthropogenic development including oil and gas infrastructure and roads. We found that turkey space use was strongly associated with tall woody cover at both the landscape and home range scale during breeding season and at the landscape scale during nonbreeding season. Additionally, turkey space use was negatively related to the density of oil/gas wells at both scales during the breeding season and at the landscape scale during the nonbreeding season. Turkeys also selected areas close to roads at the landscape scale but avoided high traffic roads within their home ranges during the breeding season. Turkey also chose areas that were >24 mo post fire and close to water sources during the breeding season at both scales. Although turkeys exhibited consistent habitat selection patterns throughout the year relative to trees and oil and gas, our results suggest that their habitat requirements in western Oklahoma are seasonally variable. Management for the Rio Grande subspecies of turkey should maintain tree cover that provides roosting and loafing cover, as well as consider the potential negative impacts of oil and gas infrastructure when locating future energy development.

Multiscale influences on Elk (Cervus canadensis) calving site selection across a fragmented Appalachian landscape

Abstract Due to the vulnerability of neonatal ungulates, selection of a birth site can have important implications for offspring survival and thus fitness of the mother. We studied parturition site selection in cow Elk in southeastern Kentucky, United States, using a use–availability framework to evaluate the effects of landscape variables sampled at multiple spatial grains on the relative probability of use of parturition sites. We identified 81 Elk parturition sites during May–Aug 2020–2022 and fit several candidate resource selection function models using a sample of 24,314 random locations to characterize habitat availability. Using an information-theoretic approach to rank candidate models, we identified two top-performing models (cumulative ωi = 0.97), which indicated that at fine spatial grains (~10- and 30-m pixels) parturient cow Elk selected for intermediate canopy cover and lower terrain ruggedness compared to available locations. At coarse grains (250–1,000-m buffers/neighborhoods), Elk selected against vegetation greenness/biomass, for higher topographic positions, for closer proximity to major roads, and with higher solar radiation potential. We also found evidence that Elk responded positively to forest/open edge densities at coarse grains, demonstrating that calving habitat throughout the Kentucky Elk Restoration Zone is associated with heterogeneous landscapes resulting from surface mine reclamation. As a result, habitat management actions should aim to increase patchy, early successional shrub cover on gentle topography.

Habitat Selection: Autumn and Winter Behavioral Preferences of Water Deer (Hydropotes inermis) in Northeast China

The wild water deer (Hydropotes inermis) population has declined rapidly over recent decades and has reached an endangered status in China. Therefore, it is important to understand their habitat selection to effectively protect both existing and emerging populations. This paper used the data of 11 habitat factors in Baishan Musk Deer National Nature Reserve in the autumn and winter from 2018 to 2019 to conduct a habitat selection study of water deer by resource selection function analysis. The results indicated that in both the autumn and winter, water deer preferred grasslands at sunny and middle slopes, dominated by Artemisia carvifolia and A. argyi, respectively. In addition, the resource selection function showed that the height of dominant herbage, hiding cover, distance from water, and distance to human settlements greatly contribute to the habitat selection of water deer in the cold season. The correct prediction rate of the resource selection function model exceeded 80%, highlighting its suitability for predicting the habitat selection of water deer. The outcomes of this study provide an effective scientific basis for the conservation and restoration of water deer, and valuable enlightenment for implementing a sustainable development strategy in northeast China.

A multi-level modeling approach to guide management of female feral hogs in Great Smoky Mountains National Park

AbstractWe trapped, anesthetized, and fit 16 female feral swine (Sus scrofa) with Global Positioning System (GPS) collars in Great Smoky Mountains National Park (GRSM) to develop predictive summer and winter models for more effective population control efforts. Given the highly diverse habitat and topography in GRSM and the spatial extent of our dataset, we employed Step Selection Function (SSF) to evaluate resource selection at the 3rd-order level and Resource Selection Function (RSF) models at the 2nd-order level for both summer and winter seasons. The summer SSF and RSF models suggested relatively similar levels of selection, whereas the winter models differed by method. We created a straightforward consensus model to better visualize the agreement and constraints of each set of models. In summer, feral swine used lower slopes regardless of elevation, especially those closer to human-dominated spaces such as along paved and gravel roadways. In winter, feral swine maintained preference for lower slopes but preferred oak-dominated forest areas and selection for human development was less than in summer. Wildlife managers can use these models to better focus feral swine surveillance and management in GRSM. Managers can identify areas of high use by season and plan control activities that are both accessible and highly efficient. The combination and consensus framework presented here can be applied to other systems where species’ habitat selection may result in incongruous results across different levels of selection or seasons of interest.

Territory establishment and microhabitat selection of Bachman's sparrows in a Mississippi pine forest

AbstractBachman's sparrows (Peucaea aestivalis) are an obligate southern open pine forest species linked to disturbance regimes (e.g., fire, windthrow) that reduce canopy cover and promote herbaceous groundcover needed for nesting, foraging, and concealment from predators. Early successional groundstory characteristics suitable for Bachman's sparrows are ephemeral in nature, raising questions about the perceived visual cues created by vegetation structural characteristics and how these characteristics influence multi‐scale habitat selection. We quantified habitat selection for 40 breeding and post‐breeding male Bachman's sparrows within 5 separately managed open pine management sub‐units at Sam D. Hamilton Noxubee National Wildlife Refuge in east‐central Mississippi, USA, from 2020–2021. We captured and color‐banded 20 individuals each year and used weekly resight locations and resource selection function models to assess factors influencing territory establishment (i.e., second‐order selection) within our study area, and microhabitat selection (i.e., third‐order selection) within territories. Bachman's sparrow territory establishment was positively influenced by evergreen cover and fire disturbance, and negatively influenced by tree canopy cover, and distance to unpaved roads. Within territories, sparrows selected areas with a greater basal area of pine (18 m2/ha), more available perching options (&gt;9), and an abundance of native herbaceous groundcover (&gt;36%). They also selected recently burned areas (1–3 yr) with low visual obstruction (&lt;40 cm), and a low percent coverage of leaf litter (&lt;33%). A better understanding of territory establishment and microhabitat selection of the Bachman's sparrow underscores the importance of frequent fire and is helpful in guiding conservation and management efforts for this declining species.

Mesopredator predation risk limits northern bobwhite nesting habitat

AbstractPredators can significantly limit the amount of available habitat for vulnerable prey. Encouraging spatial segregation of predators and prey through habitat management prescriptions may be a useful tool to increase productivity and abundances of prey species of conservation concern. We estimated the total area of low‐risk nesting habitat by interacting predator habitat selection models with northern bobwhite (Colinus virginianus) nest‐site selection models. We used trail cameras, transect sampling, spotlighting transects, and incidental tracking to survey and map occurrences of mesopredators at a site in western Oklahoma, USA. We created broad‐scale resource selection function models for bobwhite nest‐site and mesopredator habitat selection using generalized linear models to create maps of occurrence probability in ArcGIS. Bobwhite nests were more likely to be depredated in areas of high selection by coyotes (Canis latrans) and striped skunks (Mephitis mephitis) compared to areas with low mesopredator selection. Approximately 392 ha of bobwhite nesting habitat (17%) had low probabilities of selection by both coyotes and striped skunks. Our results demonstrate the need for incorporating species' interactions into estimations of available habitat. Based on the total area of highly selected cover by northern bobwhite, approximately 29% of the study area would have been deemed suitable for nesting. However, only 5% of that area is predicted to have a low predation risk.

Effects of grazing strategy on facultative grassland bird nesting on native grassland pastures of the Mid-South USA.

Understanding how livestock grazing strategies of native warm season grasses (NWSG) can impact facultative grassland bird nesting can provide insight for conservation efforts. We compared pre and post treatment effects of rotational grazing (ROT) and patch-burn grazing (PBG) for facultative grassland bird species nest success and nest-site selection on NWSG pastures at three Mid-South research sites. We established 14, 9.7 ha NWSG pastures and randomly assigned each to either ROT or PBG and monitored avian nest-site selection and nest success, 2014–2016. We collected nesting and vegetation data in 2014, before treatment implementation, as an experimental pre-treatment. We implemented treatments across all research sites in spring 2015. We used a step-wise model selection framework to estimate treatment effect for ROT or PBG on avian nest daily survival rate (DSR) and resource selection function (RSF) at the temporal scale and within-field variables. Daily survival rates were 0.93% (SE = 0.006) for field sparrow (Spizella pusilla), 0.96% (SE = 0.008) for red-winged blackbird (Agelaius phoeniceus), and 0.92% (SE = 0.01) for indigo bunting (Passerina cyanea). Model support for PBG treatment and vegetation height were indicated as negative and positive influences for field sparrow DSR, respectively. Red-winged blackbirds’ DSR were negatively influenced by ROT while vegetation height positively affected DSR, and DSR for indigo bunting did not differ among treatments. Combined RSF models indicated nest-site selection for all species was positively related to vegetation height and only weakly associated with other within-field variables. We provide evidence that ROT and/or PBG effects vary by species for DSR for these three facultative grassland birds, and vegetation characteristics affected their nest-site selection in the Mid-South USA. A lack of disturbance in Mid-South grasslands can lead to higher successional stages (i.e., mix shrub-grassland), but some combination of ROT, PBG, and unburned/ungrazed areas can provide adequate nesting habitat on small pasture lands (∼1.8 –7.8 ha) for various facultative grassland birds and potentially offer the opportunity to simultaneously maintain livestock production and grassland bird nesting habitat.

Influence of interspecific competition on mule deer birthing and rearing site selection

AbstractUngulates often alter behavior and space use in response to interspecific competition. Despite observable changes in behavior caused by competitive interactions, research describing the effects of competition on survival or growth is lacking. We used spatial modeling to determine if habitat use by female mule deer (Odocoileus hemionus) was affected by other ungulate species prior to, during, and after parturition. We conducted our study in the Book Cliffs region of eastern Utah, USA, during 2019 and 2020. We used resource selection function (RSF) analysis to model space use of 4 ungulate species that potentially competed with mule deer: bison (Bos bison), cattle, elk (Cervus canadensis), and feral horses. We incorporated RSF models for competing species into a random forest analysis to determine if space use by mule deer was influenced by these other ungulate species. We used survival and growth data from neonate mule deer to directly assess potential negative effects of other ungulates. Habitat use by elk was an important variable in predicting use locations of mule deer during birthing and rearing. The relationship was positive, suggesting interference competition was not occurring. Survival of neonate mule deer increased as the probability of use by elk increased (hazard ratio = 0.185 ± 0.497 [SE]). Further, probability of use by elk in rearing habitat had no influence on growth of neonate mule deer from birth to 6 months of age, suggesting that exploitative competition was not occurring.

Temporal mismatch in space use by a sagebrush obligate species after large‐scale wildfire

AbstractThe increase in size and frequency of wildfires in sagebrush steppe ecosystems has significant impacts on sagebrush obligate species. We modeled seasonal habitat use by female greater sage‐grouse (Centrocercus urophasianus) in the Trout Creek Mountains of Oregon and Nevada, USA, to identify landscape characteristics that influenced sage‐grouse habitat selection and to create predictive surfaces of seasonal use 1 and 7 years postfire. We developed three resource selection function models using GPS location data from 2013 to 2019 for three biologically distinct seasons (breeding, n = 149 individuals: 8 March–12 June; summer, n = 140 individuals: 13 June–20 October; and winter, n = 94 individuals: 21 October–7 March). For all seasons, by the fourth or fifth year postfire, sage‐grouse selected for unburned patches more than all other burn severity patches and the use of unburned areas in comparison with burned areas increased through time. During the breeding season, sage‐grouse selected for low‐sagebrush (Artemisia arbuscula)‐dominated ecosystems and areas with low biomass (normalized difference vegetation index). During summer, sage‐grouse selected for areas with higher annual and perennial grasses and forb cover, and areas that had higher biomass. During winter, sage‐grouse selected for areas of intact sagebrush on less rugged terrain. For the winter and breeding season, there was a positive linear relationship between annual grasses and forb cover through time. Seven years postfire (2019), the area predicted to have a high probability of use in each seasonal range decreased (breeding: 16.4%; summer: 12.2%; and winter: 4.2%), while the area predicted to have low or low‐medium probability of use increased (breeding: 14.5%; summer: 22.5%; and winter: 22.8%) when compared to the first year following the wildfire (2013). Our results demonstrated a 4‐ to 5‐year time lag before female sage‐grouse adapted to a disturbed landscape began avoiding burned areas more than intact, unburned habitats. This mismatch in ecological response may imply declines in habitat availability for sage‐grouse and may destabilize population vital rates. Spatially explicit models can aid in identifying priority areas for restoration efforts and conservation actions to mitigate the impacts of future disturbances.

Resource selection functions of black‐tailed prairie dogs in Native nations of Montana

AbstractThe black‐tailed prairie dog (Cynomys ludovicianus) is a keystone species of North American short‐ and mixed‐grass temperate grasslands with critical importance for multiple species of conservation concern. Low population numbers due to habitat and social constraints as well as disease require conservation and management actions. To identify potential habitat and future relocation sites, we evaluated 6 statistical techniques to predict the optimal resource selection function model for black‐tailed prairie dog habitat in 3 Native nations of Montana. Random forest models performed best across our study area and included grass coverage and climatic and topographic variables as the most important predictors of prairie dog occupancy. Using predictive habitat models in conjunction with land‐use designations can refine conservation planning to support wildlife population objectives where there are foundations for long‐term success. Our analysis also serves as a guide for assessing habitat suitability, connectivity, and restoration for additional species dependent or interconnected with black‐tailed prairie dogs.

High dependence on protected areas by the endangered giant armadillo in Argentina

A combination of technics was used to assess landscape variables associated with the presence of the globally vulnerable giant armadillo (Priodontes maximus) in its southernmost distribution in the Chaco region of Argentina. Between 2017 and 2019, 369 km of foot and vehicle transects were run to search and record indirect evidence of the species (burrows and feeding excavations) in areas with different level of protection, including two national parks, three provincial parks, and non-protected areas. Distance sampling was used to estimate density of excavations and effective strip width sampled to estimate the surveyed area. Resource selection functions (RSFs) were used to assess the variables associated to its presence by comparing locations of excavations with available habitat within surveyed areas. In these RSF models covariates that may affect habitat use and survival of giant armadillos were included: three categorical ones: soil type, vegetation cover (forests vs. grasslands), protection level (national park, provincial park, non-protected), and five continuous variables: distance to nearest river, distance to nearest national park, distance to closest protected area, distance to nearest transformed area, and distance to natural grasslands. Habitat use was modelled pooling together burrows and feeding excavations, and models were validated with a pool of records previously excluded from the analysis. A total of 104 records of burrows and 344 of feeding excavations were recorded. The probability of occurrence of giant armadillo excavations was much higher within forests, within or near protected areas, and increased with the distance to rivers, to transformed areas, and to grasslands. In the study area, only 10% of the landscape surveyed has good- to high-probability of occurrence of giant armadillos. The maintenance of large surfaces of native forests and the consolidation, improvement, and connectivity among protected areas seem to be essential for the long-term survival of the endangered giant armadillo in Argentina.

Evaluating habitat selection models to predict critical habitat for mountain goats in northwest British Columbia

AbstractAnthropogenic activity imposes increasing pressure on wildlife populations globally; these pressures can affect habitat suitability and function, modify wildlife space use, and influence population viability. Native mountain goat (Oreamnos americanus) populations can be negatively affected by anthropogenic disturbance and modify their space use in response to land development and recreational activity. From 2018 to 2020, we studied space use of mountain goats northeast of Smithers, British Columbia, Canada, an area that is subject to increasing anthropogenic development and yearlong recreational activities. We aimed to generate models that would improve our ability to identify habitat for mountain goats relative to existing survey data and established ungulate winter ranges. Using resource selection function (RSF) analyses generated from global positioning system (GPS) collar data, we identified influential habitat covariates and compared these covariates and RSF values to existing habitat models. Additionally, we compared the extent to which our models were congruent with existing resource selection probability functions, were congruent with aerial survey data, and overlapped existing ungulate winter ranges previously derived from predictive models inside and outside of the study area. Overall, our models noted higher RSF values among GPS data relative to aerial survey data for winter months, while results for summer habitats were comparable. In extending our RSFs outside of the study area and evaluating the overlap with ungulate winter ranges in adjacent areas, values were similar, albeit lower, as is expected given that the models were developed elsewhere. Ultimately, these models, combined with existing methods, improve the accuracy and reliability of identified, important areas of habitat for mountain goats. We recommend that the RSF models generated here be used in conjunction with aerial survey data and existing methods to delineate ungulate winter ranges for mountain goats in similar eco‐regions in British Columbia. The models developed here support existing methods that have been used to delineate or validate ungulate winter ranges for mountain goats in British Columbia and help facilitate mitigation measures to support the continued use of important winter habitat and significant landscape features that play a role in ensuring population viability and resilience through time.

Enabling safe passage: predicting river crossing hotspots for a threatened boreal ungulate susceptible to drowning

Abstract Many mammals cross large rivers to access seasonal habitats, yet river crossing behavior is poorly acknowledged or understood. Crossing large rivers is inherently risky, with vessel traffic and flooding events increasing the risk of drowning. We investigated river crossing behavior by threatened bison (Bison bison) in the boreal forest, using data from 11 GPS-collared animals to identify when and where they cross a major river. We used kernel densities of GPS locations to identify crossing hotspots, and developed resource selection function (RSF) models that used habitat characteristics to explain crossing behavior based on a priori hypotheses. We then predicted high-probability crossing areas along a 400-km stretch of river using the best-supported RSF model. Bison regularly crossed the river (x¯ = 8.6 crossings/100 days), particularly during summer months. Model selection results suggested that bison may have chosen river crossing areas that maximized access to forage. Bison preferred to cross the river near islands and areas with high seismic line densities, both of which were likely preferred summer foraging habitats. In addition, islands may have been used as temporary stopovers while crossing the river, or for relief from biting insects. Bison crossed the river where the channel was relatively narrow (x¯ = 400 ± 213 m [SD]) and chose days when river discharge was low, likely to reduce the risk of drowning, exhaustion, or hypothermia. While based on a small sample of collared bison, predictions about high-probability crossing areas may be used to inform mitigation measures aimed at reducing bison drowning caused by vessel traffic, which is expected to increase as resource development expands in the region. Our approach may be informative for identifying river crossing hotspots for other mammals that cross major rivers.

贺兰山西坡啮齿动物生境选择及种间关系

PDF HTML阅读 XML下载 导出引用 引用提醒 贺兰山西坡啮齿动物生境选择及种间关系 DOI: 10.5846/stxb202109152593 作者: 作者单位: 作者简介: 通讯作者: 中图分类号: 基金项目: 国家自然科学基金资助项目（31870512，32071649，32070519）；中央高校基本科研业务费资助项目（2572014CA03） Habitat selection and interspecific relationship of rodent in the west slope of Helan Mountain Author: Affiliation: Fund Project: National Natural Science Foundation of China [31870512, 32071649, 32070519],Fundamental Research Funds for the Central Universities [2572014CA03], 摘要 | 图/表 | 访问统计 | 参考文献 | 相似文献 | 引证文献 | 资源附件 | 文章评论 摘要:为明确对贺兰山西坡啮齿动物生境选择起关键作用的主要环境因子，探讨各鼠种之间的相互作用如何影响鼠种的生境选择，选取地理因子（海拔、坡度、坡向）、植被因子（植被盖度、植被高度、灌木/乔木高度）和其他因子（距水源距离、距道路距离、距灌木/乔木距离）等9种生境因子，利用资源选择函数模型和logistic回归模型对啮齿动物生境选择和种间关系进行分析。结果表明，在贺兰山西坡，捕获数量最多的6种啮齿动物对的生境选择存在差异，大林姬鼠（Apodemus peninsulae）偏好高海拔、多灌木（乔木）的生境；子午沙鼠（Meriones meridianus）偏好海拔较低的半荒漠稀疏生境，对道路和人为干扰不回避；灰仓鼠（Cricetulus migratorius）偏好距水源较远、多灌木（乔木）、距道路较近的生境；阿拉善黄鼠（Spermophilus alaschanicus）偏好低海拔、植被丰富的生境；短尾仓鼠（Cricetulus eversmanni）偏好高海拔、距道路较近的生境；社鼠（Niviventer niviventer）偏好海拔较高、坡度较大、距水源较远的生境。鼠种间的相互关系主要呈现回避或随机关系。研究结果有助于了解啮齿动物分布的变化及种群动态规律，可为贺兰山生物资源保护，区域性鼠害的预测、预报和综合治理提供基础资料和科学依据。 Abstract:To identify the main environmental factor that plays a key role in habitat selection of rodents in the western slope of Helan Mountains, and to explore how the interaction among rodent species affect habitat selection, we selected geographic factors (altitude, gradient, slope direction), vegetation factors (vegetation coverage, vegetation height, height of shrubs/arbors) and other factors (distance from the water, away from the road distance, distance from shrubs/arbors) to analyze the habitat selection and interspecific relationship of rodents by resource selection function model and logistic regression models. The results showed that there were significant differences in the habitat selection among six captured rodent species. Apodemus peninsulae preferred high-altitude habitats covered with dense shrubs/arbors. Meriones meridianus preferred sparsely vegetated semi-desert habitats at lower altitude, and did not avoid the road and human disturbance. Cricetulus migratorius preferred habitats which were far away from water, covered with dense shrubs/arbors and close to roads. Spermophilus alaschanicus preferred the low-altitude habitats where are rich in vegetation. Cricetulus eversmanni preferred the high altitude habitats close to roads. Niviventer niviventer preferred habitats with higher altitudes, higher slopes and far from water. The relationship between rodent species is mainly avoidance or random. Our study is contributing to understand the changes of rodent distribution and population dynamics. It can provide basic data and scientific basis for protecting biological resources, predicting and forecasting rodents damage, and comprehensive management of the regional damage that caused by rodent in Helan Mountains. 参考文献 相似文献 引证文献

Using LiDAR and Random Forest to improve deer habitat models in a managed forest landscape

Conservation strategies are hindered by a lack of accurate maps of important habitat for many wildlife species, but especially for species inhabiting managed forest landscapes. Prioritizing restoration efforts on Alaska’s Tongass National Forest from past extensive clearcut logging is extremely challenging given the difficulty in accurately mapping its remote, rugged temperate rainforest landscapes. We tested the application of airborne light detection and ranging (LiDAR) technology to build a winter habitat model for Sitka black-tailed deer (Odocoileus hemionus sitkensis), the primary herbivore in the coastal temperate rainforest. We analyzed the importance of geomorphometric and forest structure characteristics as predictors of deer winter habitat selection using Random Forest applied to a 3-year GPS relocation dataset collected from 40 adult female deer. The LiDAR-based habitat model had a predictive performance of 94% (Out-of-bag error = 6%), a 10% lower model error compared to air-photo interpreted polygons and modeled plot data. Random Forest also outperformed analogous resource selection function models based on a comprehensive k-fold cross-validation. Deer habitat selection patterns in the LiDAR-based model were nonlinear across geomorphometric and forest structure predictive variables, and generally supported existing studies of deer habitat selection. Besides improving deer conservation and management on the Tongass National Forest, our approach could greatly enhance the accuracy and resolution of habitat maps used for conservation and restoration planning across large managed forest landscapes.

Individual-based seasonal habitat selection in a forest-dwelling population of reintroduced bison Bison bison

Mapping habitat selection by threatened species provides critical information for conservation planning. For reintroduced populations, understanding habitat selection is also necessary to predict dispersal and inform selection of new reintroduction sites. Efforts to restore bison Bison bison to the boreal forest hinge on the persistence of geographically isolated populations that occupy diverse landscapes, and for many populations selected habitats are unknown. We used location data from GPS-collared bison to develop seasonal resource selection function (RSF) models and predictive maps for the reintroduced ‘Nahanni’ population. We accounted for variation in individual behaviour by calculating averaged population-level selection coefficients from individual RSFs, and we compared these results to a pooled RSF from all bison. Individual RSFs revealed variation in habitat selection that was not always captured by the pooled RSF, although there were some consistencies. Bison strongly selected forage-rich graminoid-dominated wetlands (fens) during winter, but less so in summer when there were potential tradeoffs with poor footing and biting flies. In summer, bison selected alternative sources of forage such as herbaceous, shrubby and fluvial habitats (i.e. riverine islands and gravel bars). The observed association with fluvial habitat may be an adaptation to low forage availability on this landscape. Bison also selected roads and anthropogenic clearings associated with resource development, demonstrating potential for human–wildlife conflict. Our predictive maps highlight areas of conservation interest, and should be considered in land use planning and environmental assessments. We demonstrate the value of foraging habitat for forest-dwelling bison, particularly in winter. Identifying forage-rich habitat patches, and connectivity between them, is important when considering sites for new reintroductions or expansion of existing populations. More broadly, our approach may be used to identify areas of high conservation interest, where resources do not allow extensive sample sizes of GPS-collared animals.