Midwest Fish & Wildlife Conference 2019

AUTHORS: Scott Bergeson, Purdue University-Fort Wayne; Joy O'Keefe, Indiana State University

ABSTRACT: There is a growing interest in the effects of timber harvest on forest-dwelling bats due to the potential for timber harvest to reduce available habitat. We conducted a study to determine how endangered Indiana bats (Myotis sodalis) select summer roosts within a Midwestern forest managed for timber. In the summers of 2012–2014, we tracked 4 male and 11 female Indiana bats to 49 roosts (n_male= 24, n_female= 25) in south-central Indiana, USA. We collected multi-scale data on roosts and associated available trees, randomly located throughout the same landscape. We generated 10 matched pairs conditional logistic regression models based on a priori hypotheses on roost selection and ranked them using Akaike’s Information Criteria. Plausible models explaining female roost selection included those coding for typical Indiana bat maternity roosts and typical tree-cavity bat roosts. Females selected roosts under exfoliating bark on large (17 ± 2 m in height and 34.8 ± 3.0 cm in diameter) standing dead trees and in bat boxes with high solar exposure (28.0 ± 6.0 % canopy closure above roosts). For males, the model coding for predator avoidance was the most plausible explanation of roost selection. Males selected for roosts under exfoliating bark on tall trees (23 ± 2 m; 71% snags) surrounded by snags (4.5 ± 0.7 snags/0.1 ha plot) and live trees (30.4 ± 2.7 live trees/0.1 ha plot). Females roosted in or 10 m from harvest openings and first-stage shelterwood cuts more than expected (15 of 25 roosts) based on their availability on the landscape. Males roosted in harvest openings as expected (3 of 24 roosts). Our results demonstrate that a managed Midwestern forest provides an array of roosts for Indiana bats and that Indiana bats do not actively avoid roosting near harvest openings in this forest.

AUTHORS: Melissa D. Starking, Michigan State University; Gary Roloff, Michigan State University; Michael Donovan, Michigan Department of Natural Resources

ABSTRACT: Northern hardwood forests cover > 2 million ha in Michigan and are managed for ecological and timber values. The mammal community of these forest systems provides important ecological and recreational functions. As part of a long-term study on forest regeneration techniques and wildlife interactions, we deployed an unbaited 25-camera grid (1 camera every 0.49 ha) across a 12 ha hardwood site in the Upper Peninsula, Michigan. We collected data from May 2017 through June 2018. We tagged photos to species and used standard photo verification processes. We documented a wide functional range of mammals, including small mammals (deer mice, flying squirrels), mesocarnivores (marten, fisher, bobcat, coyote), and larger herbivores (deer) and a carnivore (wolves). We quantified number of cameras needed to reliably detect white-tailed deer, snowshoe hare, black bear, and marten. We report on amount of sampling needed to represent detectable mammal species using relatively localized areas in managed northern hardwood forests in Michigan.

AUTHORS: Sara Adamczak, The Ohio State University; Anne Wiley, The University of Akron; Mažeika Sullivan, The Ohio State University; Stanley Gehrt, The Ohio State University

ABSTRACT: The North American river otter (Lontra canadensis) is an apex riverine predator that is adapted to hunting in the water. This lifestyle results in a diet focused on aquatic and semi-aquatic species.  We sought to determine river otter diet composition and trophic level using stable isotope analysis. Understanding diet and trophic level provides insight into how river otters might influence community dynamics and potentially reveal important prey species for river otters in various riverine systems. We collected tissue, whisker and nail samples collected from 108 river otters across Ohio during 2017-2018. We used these samples to estimate the ratios d¹⁵N and d¹³C, comparing river otter values to those of potential prey items. The d¹⁵N measurements serve as indicators of a consumer's trophic level, whereas d¹³C values are used to determine carbon sources in a trophic network. Our preliminary analyses revealed a higher d¹⁵N for males (mean = 14.58) than females (mean = 13.97), suggesting that male otters tend to feed at a slightly higher trophic level than female river otters. The trophic level of juvenile river otters (15N = 14.02) was the same as adults (= 2 years old; d¹⁵N = 14.01); possibly a result of juveniles mostly feeding from their mothers until weaning. Yearling river otters (1-2 years old) appeared to feed at a higher trophic level than adults. This age difference may be a result of younger individuals feeding at multiple trophic levels compared to adults. The feeding habits of younger individuals, and consequently their trophic level, will often change over time. Perch, sunfish, and crayfish were found the preferred prey species across all otter age classes. These results support the idea of stable isotopes as a useful tool when examining river otter ecology, providing insight on river otter dietary sources, as well as their trophic positioning.

AUTHORS: Edward. D. Davis, Western Illinois University; Tim C. Swearingen, Western Illinois University; Robert W. Klaver, U.S. Geological Survey; Christopher N. Jacques, Western Illinois University

ABSTRACT: Increased understanding of mortality of bobcats (Lynx rufus) is a prerequisite to successful management programs, particularly as it relates to population dynamics and the role of population models in adaptive species management. Survival and cause-specific mortality of bobcats have been well documented in predominantly forested landscapes, but limited information has been collected in agriculturally-dominated Midwestern landscapes. Thus, our objective was to evaluate survival and cause-specific mortality rates of bobcats across agriculturally dominated landscapes of west-central Illinois. We captured and radio-collared 38 (20 males, 18 females) bobcats from January 2016 to September 2018. We used known fate models with the logit link function in Program MARK to estimate annual survival of bobcats, which accommodated staggered entry and exit times of radiocollared bobcats during our analysis interval. Because mortality events were limited, covariate modeling was not conducted. Nevertheless, we constructed a survival model in which survival was constant (S{_constant}) between years and across sexes. We documented 11 deaths during our study; vehicle collisions was the leading cause of mortality and accounted for 5 (45%) mortality events. We attributed remaining deaths to harvest (n = 3; [1 legal, 1 illegal, 1 incidental harvest]), unknown (n = 1), other (n = 1), and capture-related factors (n = 1); we censored capture-related deaths from analyses. The estimated annual survival rate using model S{_constant} was 0.74 (95% CI = 0.55–0.87). Bobcat survival monitoring is ongoing through 2019 and will evaluate potential effects of intrinsic and habitat variables on seasonal and annual survival rates.