Brian Roell - Academia.edu (original) (raw)

Papers by Brian Roell

Biological conservation, Mar 1, 2024

Canadian field-naturalist, 2001

Human–Wildlife Interactions, 2011

Wolves (Canis spp.) have recolonized the Great Lakes region and expanded into agricultural areas ... more Wolves (Canis spp.) have recolonized the Great Lakes region and expanded into agricultural areas where there is increasing concern of conflict with livestock. We documented 121 wolf predation events on captive or domestic animals in the Upper Peninsula (UP) of Michigan between April 1996 and April 2009. We investigated the relationship between annual wolf abundance and predation events, seasonality of predations on livestock, and the association between previous winter severity and predations on livestock. The annual number of predations on livestock increased with wolf abundance, and overall, predations on cattle and calves increased during calving season. We observed a direct relationship between the annual number of predations on livestock and previous winter severity. We observed no relationship between the annual number of domestic dogs killed by wolves and wolf abundance. If the observed trends persist, wolf-livestock conflict in the UP will continue to increase, elevating management costs and likely reducing human tolerance for wolves. Managers should be prepared for continued conflicts as wolf populations increase and eventually are delisted in the region.

Wildlife Biology, Mar 1, 2011

Wolves Canis spp. in the Great Lakes region have expanded into rural areas where livestock produc... more Wolves Canis spp. in the Great Lakes region have expanded into rural areas where livestock production occurs, resulting in an increase of conflicts. We applied a predictive spatial model for livestock predations by wolves developed by Treves et al. (2004; hereafter the '2004 model') to the Upper Peninsula (UP) of Michigan. The 2004 model did not satisfactorily discriminate between townships with (57.1%) and without (65.7%) wolf predations (61.4% overall) that occurred during 15 April 1996-14 April 2009. Consequently, we adapted the 2004 model based on deer density and spatial data derived from the UP to maximize the model's predictive ability in the UP. We used matched pair analysis of six landscape variables significant in the 2004 model. Our adapted model improved on the 2004 model, and overall discriminated 70% of townships in our sample (N ¼ 70). Affected townships (i.e. townships with predations) in the UP displayed a consistent set of landscape variables, including relatively higher proportions of pasture/hayfield and crops, and relatively lower proportions of coniferous forest. We extrapolated from the 35 affected townships to the entire UP to generate two maps, available to managers for assistance in predicting townships at higher risk of livestock predation by wolves. As wolves continue to recover in the Great Lakes region, predicting livestock predations by wolves can assist managers in limiting the number of conflicts, as well as costs of control and compensation.

Scientific Reports, Aug 8, 2022

Using existing data can be a reliable and cost-effective way to predict species distributions, an... more Using existing data can be a reliable and cost-effective way to predict species distributions, and particularly useful for recovering or expanding species. We developed a current gray wolf (Canis lupus) distribution model for the western Great Lakes region, USA, and evaluated the spatial transferability of single-state models to the region. This study is the first assessment of transferability in a wideranging carnivore, as well as one of few developed for large spatial extents. We collected 3500 wolf locations from winter surveys in Minnesota (2017-2019), Wisconsin (2019-2020), and Michigan (2017-2020). We included 10 variables: proportion of natural cover, pastures, and crops; distance to natural cover, agriculture, developed land, and water; major and minor road density; and snowfall (1-km res.). We created a regional ensemble distribution by weight-averaging eight models based on their performance. We also developed single-state models, and estimated spatial transferability using two approaches: state cross-validation and extrapolation. We assessed performance by quantifying correlations, receiver operating characteristic curves (ROC), sensitivities, and two niche similarity indices. The regional area estimated to be most suitable for wolves during winter (threshold = maximum sensitivity/specificity) was 106,465 km 2 (MN = 48,083 km 2 , WI = 27,757 km 2 , MI = 30,625 km 2) and correctly predicted 88% of wolf locations analyzed. Increasing natural cover and distance to crops were consistently important for determining regional and single-state wolf distribution. Extrapolation (vs. cross-validation) produced results with the greatest performance metrics, and were most similar to the regional model, yet good internal performance was unrelated to greater extrapolation performance. Factors influencing species distributions are scale-dependent and can vary across areas due to behavioral plasticity. When extending inferences beyond the current occurrence of individuals, assessing variation in ecology such as habitat selection, as well as methodological factors including model performance, will be critical to avoid poor scientific interpretations and develop effective conservation applications. In particular, accurate distribution models for recovering or recovered carnivores can be used to develop plans for habitat management, quantify potential of unoccupied habitat, assess connectivity modeling, and mitigate conflict, facilitating long-term species persistence. Understanding which factors limit species distributions is a foundational question in ecology and conservation, and species distribution models (SDMs) have become an important tool to map and predict species occurrences 1-5. Use of SDMs has proliferated in the past several decades due to increasing accessibility and quantity of species occurrence data, the development of robust modeling algorithms (e.g. 6), and the improvement of software and technological resources 7-9. In addition, SDMs have been used successfully for on-the-ground conservation and management initiatives 10. For example, information regarding species occurrence and distribution is critical for natural resource decision making, including International Union for Conservation of Nature plans 11 , invasive species management 12 , state wildlife action plans 13,14 , and conservation of endangered

Biological Conservation, May 1, 2021

Abstract Cougars (Puma concolor) have lost substantial portions of their historical range yet inc... more Abstract Cougars (Puma concolor) have lost substantial portions of their historical range yet increased sightings suggest potential for re-establishment in some regions; greater understanding of potential distribution and connectivity is necessary to make sound management and policy decisions. Specifically, the Great Lakes region of the USA will likely be an important area for cougar range expansion into the Midwest and Eastern USA. We used verified cougar observations to model and predict potential distribution and connectivity in the Great Lakes region. We compiled all confirmed observations of cougars from Michigan, Minnesota, and Wisconsin (2010–2020); which resulted in 180 reports (154 images/videos, 20 signs, 6 mortalities). We developed an ensemble distribution model (1 km res) based on three machine learning methods. We used weighted cost-distances to identify linkages between core areas and circuit theory to model overall connectivity potential. We calculated selection ratios for land covers (30 m res) at fine and coarse scales. The ensemble distribution model had good performance (ROC of 0.94). Suitability was positively associated with increasing vegetation structure, lower distance to natural cover, and mid-high terrain ruggedness. Forest covers were always selected for regardless of scale, and human development was avoided only at the coarser scale. We identified 191 core patches and 362 linkages connecting them; only 50.1% of them were located within legally protected areas. We identified high regional connectivity in a generally west-east direction. Successful conservation of large carnivores like cougars will depend on conserving not only habitat patches and linkages but also efforts to facilitate long-term coexistence.

PLOS ONE, Apr 17, 2013

Background: The influence of policy on the incidence of human-wildlife conflict can be complex an... more Background: The influence of policy on the incidence of human-wildlife conflict can be complex and not entirely anticipated. Policies for managing bear hunter success and depredation on hunting dogs by wolves represent an important case because with increasing wolves, depredations are expected to increase. This case is challenging because compensation for wolf depredation on hunting dogs as compared to livestock is less common and more likely to be opposed. Therefore, actions that minimize the likelihood of such conflicts are a conservation need. Methodology/Principal Findings: We used data from two US states with similar wolf populations but markedly different wolf/hunting dog depredation patterns to examine the influence of bear hunting regulations, bear hunter to wolf ratios, hunter method, and hunter effort on wolf depredation trends. Results indicated that the ratio of bear hunting permits sold per wolf, and hunter method are important factors affecting wolf depredation trends in the Upper Great Lakes region, but strong differences exist between Michigan and Wisconsin related in part to the timing and duration of bear-baiting (i.e., free feeding). The probability that a wolf depredated a bear-hunting dog increases with the duration of bear-baiting, resulting in a relative risk of depredation 2.12-7.226 greater in Wisconsin than Michigan. The net effect of compensation for hunting dog depredation in Wisconsin may also contribute to the difference between states. Conclusions/Significance: These results identified a potential tradeoff between bear hunting success and wolf/bearhunting dog conflict. These results indicate that management options to minimize conflict exist, such as adjusting baiting regulations. If reducing depredations is an important goal, this analysis indicates that actions aside from (or in addition to) reducing wolf abundance might achieve that goal. This study also stresses the need to better understand the relationship among baiting practices, the effect of compensation on hunter behavior, and depredation occurrence.

Ecosphere

Land use and climate change alter species distributions worldwide, and detecting and understandin... more Land use and climate change alter species distributions worldwide, and detecting and understanding how species ranges shift can facilitate conservation planning and action. Following extirpation from most of the contiguous United States, gray wolves (Canis lupus) have partially recolonized former range in the western Great Lakes region, but it is unknown how land use and climate change may alter amounts of wolf habitat. Using wolf observation data collected during winters 2017–2020 in Minnesota, Wisconsin, and Michigan, we created ensemble models to predict how land use and climate change may affect the amount of wolf habitat within these states. A projection model for the western Great Lakes region suggested three of four scenarios of land use and climate change will lead to 9%–35% increases in wolf habitat, while a solely climate‐based projection model supported our expectation that changes in climate, in isolation, will have limited effect on current wolf range. Our results suppo...

Scientific Reports

Biological Conservation, Sep 1, 2022

Predicted probability of a wolf depredation on bear-hunting dogs (y-axis</i&... more Predicted probability of a wolf depredation on bear-hunting dogs (y-axis) versus the number of days since training with bait began (x-axis) in Wisconsin (upper line) and Michigan (lower line). Each point represents a day since training with bait began in Wisconsin (closed symbols) and Michigan (open symbols). Note that open symbols for Michigan are offset from (0) and (1) probability so as to not overlap symbols for Wisconsin. The odds of a depredation event occurring in Wisconsin were 3.57× greater than the odds in Michigan; a relative depredation risk 2.12–7.22× greater in Wisconsin.</p

Note. NA = not applicable.</p

Poisson log-linear relationship between annual totals of wolf depredation (i.e. kill or ... more Poisson log-linear relationship between annual totals of wolf depredation (i.e. kill or injury) on bear-hunting dogs (y-axis) and annual estimates of wolf abundance (x-axis) in Wisconsin (1990–2010; closed symbols) and Michigan (1996–2010; open symbols).</p

PLOS ONE

Carnivores are ecologically important and sensitive to habitat loss and anthropogenic disruption.... more Carnivores are ecologically important and sensitive to habitat loss and anthropogenic disruption. Here we measured trophic level and gut bacterial composition as proxies of carnivore ecological status across the Upper Peninsula, Michigan, for wild American marten (Martes americana; hereafter marten). In contrast to studies that have focused on omnivorous and herbivorous species, we find that marten, like other carnivore species without a cecum, are dominated by Firmicutes (52.35%) and Proteobacteria (45.31%) but lack Bacteroidetes. Additionally, a majority of the 12 major bacterial genera (occurring at ≥1%) are known hydrogen producers, suggesting these taxa may contribute to host energy requirements through fermentative production of acetate. Our study suggests that live trapping and harvest methods yield similar marten gut microbiome data. In addition, preserving undisturbed forest likely impacts marten ecology by measurably increasing marten trophic level and altering the gut mic...

Lower panel: cumulative wolf depredations on bear-hunting dogs (<i... more Lower panel: cumulative wolf depredations on bear-hunting dogs (y-axis) each month (x-axis) from 1980–2010 in Wisconsin (closed bars) and Michigan (open bars). Upper panel: General timing of bear-baiting, training, and hunting regulations (y-axis) in each state; x-axis and bar symbols are the same as in lower panel. In Wisconsin there is a pre-training baiting period of ∼2.5 months that does not exist in Michigan and baiting in Michigan begins ∼4 months later than in Wisconsin.</p

*Encounter = the ratio of bear hunting permits sold per wolf (see methods). Numbers in p... more *Encounter = the ratio of bear hunting permits sold per wolf (see methods). Numbers in parenthesis under explanatory factors are p-values for the five best-performing models.aAICC is Akaike’s information criterion, corrected for small sample size.bΔAICC is AICC for the model of interest minus the smallest AICC for the set of models being considered. We only considered models with ΔAICC ≤2.cW is the Akaike’s weight of each model. The ratio of one model’s weight to another estimates how many times more support the data provide for that model over the other.</p

Biological Conservation, 2021

Mink metadata, QIIME2 artifacts and demultiplexed EMP-paired end sequences from Argonne National ... more Mink metadata, QIIME2 artifacts and demultiplexed EMP-paired end sequences from Argonne National laboratory, R code for statistical analyses and figure generation, and QIIME2 pipeline for Lafferty et al. 2021.

Journal of Mammalogy, 2021

Gut microbiomes encode myriad metabolic functions critical to mammalian ecology and evolution. Wh... more Gut microbiomes encode myriad metabolic functions critical to mammalian ecology and evolution. While fresh fecal samples provide an efficient, noninvasive method of sampling gut microbiomes, collecting fresh feces from elusive species is logistically challenging. Nonfresh feces, however, may not accurately represent the gut microbiome of the host due to succession of gut microbial consortia postdefecation as well as colonization by microbes from the surrounding environment. Using American mink (Neovison vison) as a model species, we examined postdefecation microbial community succession to learn how ambient temperature and temporal sampling constraints influence the reliability of nonfresh feces to represent host gut microbiomes. To achieve our goal, we analyzed fresh mink feces (n = 5 females; n = 5 males) collected at the time of defecation from captive mink at a farm in the Upper Peninsula of Michigan and we subsequently subsampled each fecal specimen to investigate microbial com...

A New Era for Wolves and People