Resource Selection and Bait Response by Female Eastern Wild Turkey in Alabama

Abstract

A key element for the successful management of a wildlife species is understanding the

specific resources an animal is using and at what time of the year. Resource selection by animals is typically addressed in the terms of use versus availability. I defined availability for individual female eastern wild turkeys as a circular buffer with a radius equal to the largest distance an individual moved from its capture site. I outfitted 38 female eastern wild turkeys with GPS loggers from 2015 – 2019 on seven study areas. My objectives were to describe the use of land cover and landform types during the reproductive period, hunting season, and non-reproductive periods. During the non-reproductive period I found that hardwood forest were 1.29 times more likely to be selected compared to their availability on the landscape (95%; CL = 1.04 – 1.41). Mixed and pine forests were both 1.36 times more likely to be selected in the non-reproductive period, however the confidence limits for both estimates overlapped zero indicating that selection occurred in proportion to its availability (95%; CL = 0.89 – 2.09) (95%; CL = 0.92 – 2.02). Further, turkeys selected south-facing forested slopes 1.73 times more than their availability in the non-reproductive period (95%; CL = 1.17 – 2.56). During the reproductive period, I hypothesized that open areas would be important to turkeys for nesting and brood- rearing habitat. However, I found that the use of open areas was much lower than its availability on the landscape for most of the year. Selection for all land covers was in proportion to availability throughout most of the reproductive period including bottomland cover types. During the turkey-hunting season, I hypothesized that turkeys gathered in mixed-sex flocks for mating would use open areas. However, I found that open areas continued to be avoided during this period. Decisions about species management should be informed by accurate estimates of population size and structure. Baited camera surveys are a valuable tool for estimating population size and structure; however, sources of error need to be understood for appropriate survey design. There is little published information about whether the use of bait could alter the movements and home range of Eastern Wild Turkeys (Meleagris gallopavo silvestris, hereafter turkey), or lead to biased estimates of occupancy and density. I marked thirteen adult female turkeys on five study areas in Alabama using transmitters with onboard GPS loggers to monitor movements. I examined turkey response to bait by calculating weekly probabilities of use using a Kernel Density Estimator and location data. I estimated the probability of use P(useijw) of each individual (i) at each camera site (j) during each week (w). I used the sign of the difference between P(useijw) and P(useijw+1) as a binomial indicator of an increase in probability of use and estimated the probability of an increase in P(useijw) using generalized linear mixed models. P(useijw) was 1.78 (0.87-3.62; 95% C.L.) times more likely to increase when bait was present versus when bait was absent, but confidence limits included 1.0 indicating that bait did not have a conclusive effect. My results indicate that the use of bait on cameras trapping surveys will not bias our demographic estimates and does not violate the assumptions of our analytic techniques. Further analysis should examine whether use in response to bait differs between sites inside and outside of a turkey’s home range.