At high latitudes, the thermal protection of snowpack is important for many small-bodied mammal species during the winter. As depth and persistence of snow changes due to climate change, species that rely on snowpack’s insulation must increasingly depend on alternative thermal refuges. For the American marten (Martes americana), winter rest sites are predominantly in ground microsites. In some regions, however, martens primarily use tree cavities. Differences are hypothesized to be driven in part by regional differences in climate, which can be tested by comparing the use of different rest site types in areas with varying winter conditions. Here, we evaluated the roles of ambient temperature, snowpack, tree cavity temperature, and ground temperature on marten winter rest site use in two contrasting areas in the Western Great Lakes Region. Use of tree cavities in the lower peninsula of Michigan, USA appears to largely be linked to the area’s shallow snow depths and relatively warm ambient temperatures. Conversely, the use of ground microsites in northern Minnesota, USA was linked to deep snow depths and cold ambient temperatures. However, factors like microsite availability are also likely driving microsite use in both populations. Our results suggest that with rapid changes in snowpack and ambient temperature, tree cavities and subterranean burrows may become increasingly important alternatives to ground microsites that were historically covered by snow. Consequently, rapid and effective strategies will be needed to maintain or increase the availability of these alternatives.

Many ectothermic vertebrates have predictable seasonal activity and habitat use patterns, but variable patterns at fine temporal scales (e.g., minutes to days) that are likely influenced by thermoregulatory demands on behavior. Wood turtles (Glyptemys insculpta) are freshwater turtles known to use terrestrial environments during their active period. However, little research has been conducted to quantify the influence of environmental factors on diel habitat use patterns. We used fine-resolution global positioning system (GPS) tracking and temperature data to quantify wood turtle aquatic-terrestrial habitat-use patterns from May through August 2015 and 2016, focusing on the influence of temporal variables and environmental temperature. We found that temporal variables and open canopy air temperature had strong explanatory power for wood turtle aquatic-terrestrial habitat use patterns, but temperature was a stronger predictor. Terrestrial activity was positively associated with air temperature, resulting in a consistent pattern of daytime terrestrial and nighttime aquatic activity during the pre-nesting period. Male and female activity patterns diverged during the post-nesting activity period, with most males returning to the river at night and most females remaining terrestrial, influenced by females moving farther from the river. The results of our study provide valuable information to inform population survey design, habitat management planning, and potential responses to climate change for this unique species of conservation concern.

Habitat suitability models (HSMs) play an important role in conservation planning by identifying areas for habitat management and guiding surveys for population discovery and monitoring. The wood turtle Glyptemys insculpta is a riverine turtle that uses both riparian and upland environments. Habitat loss and degradation have resulted in range-wide declines of this species throughout the United States. Previous HSMs have been developed for wood turtles; however, no published HSM we are aware of exists for their range in the midwestern United States (including parts of Iowa, Michigan, Minnesota, and Wisconsin), or has incorporated terrestrial, aquatic, and climatic variables. We developed a wood turtle HSM across their distribution in the midwestern United States using Random Forest. Modeling units consisted of 800-1,000-m stream segments and a 300-m buffer surrounding each segment. We obtained verified occurrence records from 2000-2024 from state agencies, concurrent research projects, and the community science platform iNaturalist, resulting in 1,833 segments with documented occurrence to serve as model training data. Assessment metrics indicated good model performance (area under the operator curve = 0.92, true skill statistic = 0.6704). Stream order, width, and gradient were the most influential aquatic variables, with lower stream order, lower widths, and lower gradient being associated with higher probability of occurrence. The most influential terrestrial variables were proportion of still water in the terrestrial environment, landscape condition index, and topographic ruggedness index, with lower proportion of still water in the terrestrial environment, moderate to high landscape condition, and lower topographic ruggedness being associated with higher occurrence probability. The majority of stream segments (68.4%) with >= 0.8 probability of occurrence currently lack documented wood turtle populations, allowing for prioritization of these streams for future surveys. These results can assist with guiding habitat management and restoration efforts in their midwestern United States range.

The moose (Alces alces; mooz in Anishinaabemowin, Ojibwe language) population has recently declined in Minnesota, USA, and gray wolf (Canis lupus; ma'iingan) predation is likely a contributing factor. We analyzed diet composition of gray wolves in northeastern Minnesota during 2011-2013 to evaluate the importance of moose as prey and seasonal and regional variations in wolf diet. We identified frequency of occurrence of prey items and biomass consumed in 1,000 wolf scats collected on and adjacent to the Grand Portage Indian Reservation and Voyageurs National Park and within the 1854 Ceded Territory (greater northeastern Minnesota). White-tailed deer (Odocoileus virginianus; waawaashkeshiwag [plural]), moose, and beaver (Castor canadensis; amikwag [plural]) composed the majority of wolf diet, with moose as the primary prey in Grand Portage (35-54% of diet by biomass across seasons) and deer the primary prey in the 1854 Ceded Territory (46-62%) and Voyageurs (63-79%). Relative importance of prey species differed by study area and season. Moose calves were an important prey item in spring in the 1854 Ceded Territory (12% of diet by biomass) but not in Grand Portage or Voyageurs. Although calves were not a majority of wolf diet by biomass, many calves were preyed upon by wolves (30% of calves born each year in Grand Portage), thus affecting recruitment in a declining moose population. Deer fawns composed 12% of wolf diet in spring and 10% in summer in Grand Portage and 19% in summer in Voyageurs. Beaver composed 16% of wolf diet by biomass in spring and 14% in summer in Grand Portage and composed 22% of wolf diet in spring and 30% in summer in Voyageurs. At most prey densities, moose were preferred and deer avoided in Grand Portage and the 1854 Ceded Territory and beaver were preferred in Voyageurs. Our results can be used in conjunction with predation and prey studies to evaluate the effect of wolves on prey populations.

Wood turtles (Glyptemys insculpta) are a species of conservation concern throughout their geographic distribution. Several studies have investigated individual-level habitat selection of wood turtles in the Upper Midwest in the United States, but the effects of habitat characteristics on abundance are poorly understood. This information is needed to improve landscape-level habitat management and conservation initiatives for the species. Our study aimed to identify important aquatic and terrestrial habitat characteristics and quantify their influence on abundance dynamics of adult wood turtles in the Laurentian Mixed Forest Province ecoregion of Wisconsin and Minnesota, USA. We collected standardized population survey data at 57 sites within the ecoregion between 2016 and 2022. We used N-mixture models with a multi-stage model selection procedure to assess the influence of aquatic and terrestrial predictors on abundance, including several 3-dimensional forest structure metrics derived from airborne Light Detection and Ranging (LiDAR) data. Several aquatic and terrestrial habitat characteristics influenced local abundance patterns of adult wood turtles. The most influential aquatic predictors were stream velocity and stream width, and the most influential terrestrial predictors were mean return height and vertical coefficient of variation of height. Abundance was high at sites containing comparatively narrow streams with moderate velocities. The most supported terrestrial predictors were derived from LiDAR and indicate that complex forest structures support larger wood turtle populations. Our results can be used in forest management strategies to improve habitat quality for wood turtles, such as selective tree harvesting to increase structural diversity, and potentially identify robust populations in under-surveyed areas.

Reliable population estimates are important for making informed management decisions about wildlife species. Standardized survey protocols have been developed for monitoring population trends of the wood turtle (Glyptemys insculpta), a semi-aquatic freshwater turtle species of conservation concern throughout its distribution in east-central North America. The protocols use repeated active search surveys of defined areas, allowing for estimation of survey-specific detection probability (p) and site-specific abundance. These protocols assume population closure within the survey area during the survey period, which is unlikely to be met as wood turtles are a highly mobile species. Additionally, current protocols use a single-pass design that does not allow for separation of availability (pa) and detectability (pd). If there are systematic influences on pa or pd that are not accounted for in the survey design or data analysis, then resulting abundance estimates could be biased. The objectives of this study were to determine if pa is a random process and if pa and pd are influenced by demographic characteristics. We modified the wood turtle survey protocol used in the upper Midwest to include a double-pass design, allowing us to estimate pa and pd using a robust design capture-recapture model. The modified protocol was implemented at 14 wood turtle monitoring sites in Minnesota and Wisconsin between 2017 and 2022. Our results indicated that pa was non-random and that pd increased with turtle carapace length. Our study suggests that model assumptions for current wood turtle population models may be violated, likely resulting in an overestimation of abundance. We discuss possible protocol and modeling modifications that could result in more accurate wood turtle abundance estimates.

• We describe a theoretical model for estimating tree cavity temperature over time. • The model uses first principles of heat and mass transfer equations. • Predicted temperature values were accurate for natural and artificial cavities. • The model is a useful tool for understanding animal behavior and physiology.

Given recent and abrupt declines in the abundance of moose (Alces alces) throughout parts of Minnesota and elsewhere in North America, accurately estimating statewide population trends and demographic parameters is a high priority for their continued management and conservation. Statistical population reconstruction using integrated population models provides a flexible framework for combining information from multiple studies to produce robust estimates of population abundance, recruitment, and survival. We used this framework to combine aerial survey data and survival data from telemetry studies to recreate trends and demographics of moose in northeastern Minnesota, USA, from 2005 to 2020. Statistical population reconstruction confirmed the sharp decline in abundance from an estimated 7,841 (90% CI = 6,702-8,933) in 2009 to 3,386 (90% CI = 2,681-4,243) animals in 2013, but also indicated that abundance has remained relatively stable since then, except for a slight decline to 3,163 (90% CI = 2,403-3,718) in 2020. Subsequent stochastic projection of the population from 2021 to 2030 suggests that this modest decline will continue for the next 10 years. Both annual adult survival and per-capita recruitment (number of calves that survived to 1 year per adult female alive during the previous year) decreased substantially in years 2005 and 2019, from 0.902 (SE = 0.043) to 0.689 (SE = 0.061) and from 0.386 (SE = 0.030) to 0.303 (SE = 0.051), respectively. Sensitivity analysis revealed that moose abundance was more sensitive to fluctuations in adult survival than recruitment; thus, we conclude that the steep decline in 2013 was driven primarily by decreasing adult survival. Our analysis demonstrates the potential utility of using statistical population reconstruction to monitor moose population trends and to identify population declines more quickly. Future studies should focus on providing better estimates of per-capita recruitment, using pregnancy rates and calf survival, which can then be incorporated into reconstruction models to help improve estimates of population change through time.

This paper provides outlines of presentations made during a special session devoted to GPS-based telemetry systems at the 32nd North American Moose Conference and Workshop, Banff, Alberta, April2024, 1996. The intent of this session was to provide an overview of GPS technology and to relate early experiences with GPS-based telemetry systems. Performance of these new systems under both controlled and field situations were discussed. Consideration was given to issues of data management and analysis, including estimation of home range characteristics using existing software packages and integration of GPS location data with remotely sensed digital habitat data in a Geographic Information System. Cost comparisons among telemetry systems used in moose studies were made. Future directions of GPS-based telemetry systems were also discussed.. This special session demonstrated that GPS-based animal location systems have the potential to set new standards for habitat-resource utilization studies of large animals over the next 5-l 0 years. The need for an efficient, cost-effective way to collect large amounts of highly accurate location data from free-ranging animals has prompted the development of new telemetry devices based on the NA VST AR Global Positioning System (GPS) (Rodgers et al. 1996). These novel animal location systems provide data 24-hours-per-day under all weather conditions and, most importantly, with an accuracy and precision that is at least two orders of magnitude improvement over previous methods. GPS collars are the most significant advance in wildlife telemetry in more than 20 years. Given the potential significance of GPS telemetry for studies of moose and other large mammals, a special session devoted to this new technology was held at the 32nd North American Moose Conference and ALCES VOL. 33 (1997) pp.203-209 Workshop, Banff, Alberta, April 20 24, 1996. The primary objectives of this session were to provide an overview of GPS technology and to relate early experiences with GPSbased telemetry systems. Performance of these new systems under both controlled and field situations were discussed. Consideration was given to issues of data management and analysis, including estimation of home range characteristics using existing software packages and integration of GPS location data with remotely sensed digital habitat data in a Geographic Information System (GIS). Cost comparisons among telemetry systems used in moose studies were made. Future directions of GPS-based telemetry systems were also discussed. Outlines of conference presentations covering these topics are provided in the following sections. The author 'Order of authorship determined by the sequence of presentations at the 32nd North American Moose Conference and Workshop, Banff, Alberta, April20-24, 1996.

Despite the importance of the Parelaphostrongylus tenuis infection for moose (Alces alces) and white-tailed deer (Odocoileus virginianus) management, only one peer-reviewed study has evaluated the relationship between deer and moose densities and the potential for parasite-mediated competition between the species. Using polynomial-regression modeling, that study identified a deer-density threshold above which moose populations declined; however, the nature of the data and apparent outliers suggests the approach used to develop that threshold may not have been appropriate. We used the data from the original study to test whether alternative models, including linear models and negative binomial models would be less sensitive to outliers and could better explain that relationship. We found no evidence that moose density decreases as deer density increases. We concluded that, although the proposed moose-deer-P. tenuis relationship could be partially density dependent, additional factors, such as frequency dependence of disease transmission, gastropod abundance, and shared use of resources by moose and deer should also be considered.

Anthropogenic habitat change and moderating climatic conditions have enabled the northward geographic expansion of white-tailed deer, Odocoileus virginianus, and of the parasitic nematode (meningeal worm) it carries, Parelaphostrongylus tenuis. This expansion can have consequences in dead-end host species for other ungulates because meningeal worm reduces health, causes morbidity or direct mortality, and has been attributed to population declines. In northeastern Minnesota, which marks the southern extent of the bioclimatic range for moose (Alces alces), the moose population has declined more than 50% in the last decade, with studies detecting P. tenuis in 25–45% of necropsied animals. We assessed the factors that most commonly are associated with meningeal worm infection by linking moose movement ecology with known P. tenuis infection status from necropsy. We outfitted moose with GPS collars to assess their space use and cause-specific mortality. Upon death of the subject animal, we performed a necropsy to determine the cause of death and document meningeal worm infection. We then created statistical models to assess the relationship between meningeal worm infection and exposure to hypothesized factors of infection risk based on the space use of each moose by season. Predictors included land cover types, deer space use and density, environmental conditions, and demographics of individual moose (age and sex). Moose with autumn home ranges that included more upland shrub/conifer, and individuals with high proportions of wet environments, regardless of season, had increased infection risk. In contrast, the strongest relationships we found showed that high proportions of mixed and conifer forest within spring home ranges resulted in reduced risk of infection. The spring models showed the strongest relationships between exposure and infection, potentially due to moose foraging on ground vegetation during spring. By incorporating movement of moose into disease ecology, we were able to take a top-down approach to test hypothesized components of infection risk with actual spatial and temporal exposure of individual necropsied moose. The probability of infection for moose was not influenced by deer density, although deer densities did not vary greatly within the study area (2–4 deer/km2), highlighting the importance of also considering both moose space use and environmental conditions in understanding infection risk. We suggest management strategies that use a combination of deer and land management prescriptions designed to limit contact rates in susceptible populations.

Data, code, and metadata for: Temperature shapes movement and habitat selection by a heat-sensitive ungulate

The Wood Turtle (Glyptemys insculpta) is a species of conservation concern in the Upper Midwest of the USA. State agencies and partners in the region are working collaboratively to identify threats to G. insculpta populations and to implement effective management actions. One component of this conservation initiative is to improve our understanding of the impacts of different management strategies on long-term viability of populations. This requires estimates of population vital rates, which are currently lacking for most G. insculpta populations in the Upper Midwest and across the range of the species. In this study, we used individual-level monitoring data to estimate annual survival of adult G. insculpta in Iowa (two populations, four monitoring years, 52 individuals), Minnesota (one population, two monitoring years, 29 individuals), and Wisconsin (two populations, two monitoring years, 32 individuals). We estimated annual survival for each sex, population, and year using a known-fates analysis with a binomial model. Twenty-three (20%) of the monitored individuals died during the study. For 12 turtles for which we knew the cause, predation was responsible for most (n =9; 75%) mortalities. Estimated annual survival of males and females ranged from 0.49-1.00 and 0.64-1.00, respectively, among populations and years. Estimated survival of all individuals was 0.86 in Iowa for 2012-2015, 0.89 in Minnesota for 2015-2016, and 0.87 in Wisconsin for 2014-2015. This study increases our understanding of adult G. insculpta survival rates and causes of mortality in the western Great Lakes portion of the range of the species and provides useful vital rate estimates for population viability analyses.

Population genetic analyses for moose have been based on DNA extracted from blood and other body tissues. Non-invasive sampling of fecal pellets is another potential source of DNA. We compared DNA extraction from blood, liver tissue, and fecal pellet samples from moose in Minnesota and Yellowstone National Park, USA. Extracted DNA from all source types was sufficient for genotyping using 15 microsatellites. DNA extracted from fecal pellets was of lower quality and quantity than DNA extracted from blood and tissue. We provide comparisons of efficiency and effectiveness of DNA extraction protocols for blood, tissue, and fecal pellets, and demonstrate the suitability of using DNA extracted from non-invasively sampled material in moose.

Global positioning system (GPS) telemetry units are now small enough to be deployed on terrestrial and semi-aquatic turtles. Many of these GPS units use snapshot technology which collects raw satellite and timestamp data during brief periods of data recording to minimize size. We evaluated locations from snapshot GPS units in stationary tests and on wood turtles (Glyptemys insculpta) in northeastern Minnesota. Stationary GPS units were placed in wood turtle habitat to evaluate location accuracy, fix success rate, and directional bias. The GPS fix success rate and accuracy were reduced in closed canopy conditions and when the stationary GPS unit was placed under a log to simulate wood turtle hiding behavior. We removed GPS location outliers and used a moving average calculation to reduce mean location error in stationary tests from 27 m (SD = 38) to 10 m (SD = 8). We then deployed GPS units and temperature loggers on wood turtles and collected 122,657 GPS locations and 242,781 temperature readings from 26 turtles from May to September 2015 and 2016. Location outliers accounted for 12% of locations when the GPS receiver was on a turtle. We classified each wood turtle location based on the GPS location and by comparing temperature profiles from river, sun, and shaded locations to the temperature logger on the turtle. We estimated that wood turtles were on land 68% (SD = 12) of the time from May to September. The fix success rate for land locations was 38% (SD = 9), indicating that wood turtles often use habitats with obstructed views of the sky. Mean net daily movement was 55 m (SD = 192). Our results demonstrate that snapshot GPS units and temperature loggers provide fine-scale GPS data useful in describing spatial ecology and habitat use of semi-aquatic turtles.

Coarse woody debris (CWD) is an essential component of forest ecosystems that provides habitat for diverse species, functions in water and nutrient cycling, and can be a potential surface fuel in wildfires. CWD detection and mapping would enhance forestry and wildlife research and management but passive remote sensing technologies cannot provide information on features beneath forest canopy, while field-based CWD inventories are not practical for mapping CWD over large areas. Airborne light detecting and ranging (LiDAR) is a remote sensing technology that provides detailed information on three-dimensional vegetation structure that could overcome limitations of passive remote sensing technologies and field-based inventories. Our objectives were to evaluate whether airborne LiDAR could be used to detect individual pieces of CWD. We measured 1679 pieces of CWD at 144 field plots from 2015 to 2016. We acquired high-density (similar to 24 first returns/m(2)) LiDAR data in 2014, filtered out canopy and sub-canopy returns using a height threshold based on field measurements of CWD, and used height-filtered data to determine which field-measured pieces of CWD were visible in the resulting point cloud. CWD pieces that were detected constituted 50% of plot CWD volume, and there was a strong, positive correlation between total plot CWD volume and volume of detected pieces (r = 0.96). Overall, we detected 23% of the individual pieces of CWD we measured. Large pieces of CWD were most likely to be detected, with the majority of pieces >= 30 cm diameter or >= 13.9 m long detected. Canopy density, shrub density, and forest type did not influence detection probability. CWD detection rates increased from 1 pulses/m(2) to 16 pulses/m(2), and CWD detection rate was constant from 16 pulses/m(2) to 24 pulses/m(2). Our results demonstrate that airborne LiDAR can be used to detect CWD. LiDAR-based detection and mapping of CWD will be most useful for applications that focus on larger and longer pieces of CWD or applications focused on total CWD volume.

White-tailed deer (WTD; Odocoileus virginianus) are the definitive hosts of meningeal worm (Parelaphostrongylus tenuis) and liver fluke (Fascioloides magna); two parasites implicated in the decline of moose populations (Alces alces) in northern USA. Understanding which areas pose transmission risk may contribute to effective mitigation of these parasites in imperiled moose populations. Our objective was to predict areas of potential P. tenuis and F. magna transmission risk in terms of landscape features and deer density. Analyses were based on biogeographic and ecological factors related to both parasites. Using ecological niche modeling tools, remote sensing satellite data, field sampling, and estimated densities of WTD in Minnesota, we characterized current suitable environmental conditions for F. magna and P. tenuis across the historical range of moose in the state and identified potential areas for the occurrence of these parasites in unsampled areas. Our results help elucidate risky landscapes for F. magna and P. tenuis transmission by identifying geographic locations where WTD occur at high densities and with the landscape features suitable for the parasites. High-risk areas identified by our models may guide future surveillance, conservation, and management plans by identifying hotspots of potential infection of these parasites from WTD to moose populations. Our study shows the applicability of ecological niche modeling tools for investigating disease transmission risks of complex parasite systems for conservation purposes.

Numerous studies of tooth plaques and remains of gut contents of have confirmed that mastodon diet was composed of woody browse species, forbs, nuts, and fruits.However, fossil gut contents also suggest that mastodon diet included significant amounts of spruce, even though spruce is a low-quality, chemically-defended food.Most extant large mammals only browse on spruce when all other food sources are exhausted, and mastodon tusk growth increments indicate that mastodons were not food limited as they moved toward extinction (Fisher, 2009).Here we review the vegetation associated with mastodon habitat from the Great Lakes to the Gulf Coast, USA, over the period 18-10 ka cal BP using pollen assemblage data from 29 sites located near proboscidean fossil remains.Pollen data were acquired from the Neotoma Database and pollen abundance was converted into species biomass abundance using the Landscape Reconstruction Algorithm (LRA) of Sugita (2004aSugita ( , 2004b)).Although spruce was the dominant conifer throughout the Great Lakes Region until ca. 10 ka cal BP, deciduous species such as ash, oak, and elm comprised 50% or more of the vegetation assemblages even at the earliest and northernmost sites, and remained at similar levels until mastodon extinction.Many of these species have been found in mastodon gut contents.These vegetation assemblage reconstructions support the suggestion that mastodons were not food limited as they neared extinction.Moreover, these analyses of landscapes surrounding mastodon sites strongly suggest that the contemporaneous forest, composed of large amounts of spruce intermixed with ash, elm, and oak, was unlike the forests found in much of eastern North America today.

Predators directly impact prey populations through lethal encounters, but understanding nonlethal, indirect effects is also critical because foraging animals often face trade-offs between predator avoidance and energy intake. Quantifying these indirect effects can be difficult even when it is possible to monitor individuals that regularly interact. Our goal was to understand how movement and resource selection of a predator (wolves; Canis lupus) influence the movement behavior of a prey species (moose; Alces alces). We tested whether moose avoided areas with high predicted wolf resource use in two study areas with differing prey compositions, whether avoidance patterns varied seasonally, and whether daily activity budgets of moose and wolves aligned temporally. We deployed GPS collars on both species at two sites in northern Minnesota. We created seasonal resource selection functions (RSF) for wolves and modeled the relationship between moose first-passage time (FPT), a method that discerns alterations in movement rates, and wolf RSF values. Larger FPT values suggest rest/foraging, whereas shorter FPT values indicate travel/fleeing. We found that the movements of moose and wolves peaked at similar times of day in both study areas. Moose FPTs were 45% lower in areas most selected for by wolves relative to those avoided. The relationship between wolf RSF and moose FPT was nonlinear and varied seasonally. Differences in FPT between low and high RSF values were greatest in winter (-82.1%) and spring (-57.6%) in northeastern Minnesota and similar for all seasons in the Voyageurs National Park ecosystem. In northeastern Minnesota, where moose comprise a larger percentage of wolf diet, the relationship between moose FPT and wolf RSF was more pronounced (ave. across seasons: -60.1%) than the Voyageurs National Park ecosystem (-30.4%). These findings highlight the role wolves can play in determining moose behavior, whereby moose spend less time in areas with higher predicted likelihood of wolf resource selection.