Movement patterns, behavior, and habitat use of female moose on Joint Base Elmendorf-Richardson, AK
Date
2016
Authors
Battle, David C., author
Rittenhouse, Larry, advisor
Farley, Sean, advisor
Meiman, Paul, committee member
Peel, Kraig, committee member
Journal Title
Journal ISSN
Volume Title
Abstract
Joint Base Elmendorf-Richardson (JBER), which is a combined United States Army/Air Force installation, and neighboring Anchorage, Alaska, support a population of moose Alces alces (Linnaeus, 1758) that inhabit a fragmented landscape of habitat types interspersed with human development. Because development plans in support of the military mission may have significant impacts on moose movement in the area, JBER and Alaska Department of Fish and Game (ADF&G) biologists began a study of moose habitat use and behavior on JBER. In order to help identify behaviors in wild radio-collared moose captured on JBER, we tested Telonics tri-axial accelerometers for accuracy in the detection of activity and the identification of behaviors in radio-collared moose. Direct observations of three captive animals fitted with radio collars containing accelerometers allowed us to calibrate activity readings to observed behaviors. We developed four datasets in order to test whether readings from this type of accelerometer could identify specific behaviors (browsing, grazing, walking, standing, lying), behavior categories (feeding, traveling, resting), or simply when moose were active or inactive. Multiple threshold criteria were tested in order to maximize correlation to observed behaviors. The highest overall accuracy was achieved when using threshold criteria to characterize behaviors as active (92.29% accuracy) or inactive (90.64% accuracy). A Fisher’s Exact Test indicated that there was no significant difference between observed behaviors and those correctly classified using threshold criteria for either active (p = .9728) or inactive (p = .9431) behaviors, indicating that our threshold criteria is correctly classifying these behaviors. In the next phase of this study, we collected 244,957 GPS locations from 18 female moose captured on JBER and fitted with GPS collars equipped with the same model tri-axial accelerometer used in the captive trials. Data from the accelerometers were used to characterize moose behavior as active or inactive. GPS locations, along with behavior patterns and movement characteristics, were used to rank JBER habitat types. Turning angle and speed were calculated between successive locations for each animal across the animal’s home range. Values were pooled for all animals and used to assess movement characteristics by season and habitat type. The highest velocity recorded for a 60 minute period was 1.50 m/s (5.40 kph), and 99.50% of all steps had velocities < 0.26 m/s (0.94 kph). Turning angle groups did not vary among either habitat types (p = 1.00) or seasons (p = 0.99). A new, intuitive home range estimation method, Dynamic Potential Path Area (dynPPA), was used to incorporate behavioral states into the delineation of animal home ranges. We delineated dynPPA home ranges by season for each moose, and used this technique in combination with Jacobs Index (which measures utilization in relation to availability) to determine habitat preference. Seasonal dynPPA home range sizes averaged 15.28 km2 in summer (SD = 6.43) and 23.25 km2 in winter (SD = 7.97). Habitat types most often used by moose on JBER included mixed deciduous/conifer (38.23% of summer locations and 30.03% of winter locations occurred within this habitat type), shrublands (15.04% of summer locations and 28.57% of winter locations), and deciduous forest (21.89% of summer locations and 19.08% of winter locations). While individual moose differed in habitat selection (F = 1.73, df = 17, p < 0.01), the most preferred habitat (according to Jacobs Index) on JBER in relation to its availability within the home range was shrublands.