Loading…
US-IALE 2018 has ended
Back To Schedule
Wednesday, April 11 • 3:45pm - 4:00pm
WILDLIFE MANAGEMENT: A Hybrid Approach to Modeling Habitat Selection Reveals Preferences of Brown Bears on the Kenai Peninsula, Alaska, USA

Sign up or log in to save this to your schedule, view media, leave feedback and see who's attending!

AUTHORS: Andrew Mashintonio*, Kutztown University; Grant Harris, U.S. Fish & Wildlife Service; Sean Farley, Alaska Department of Fish and Game; Gareth Russell, New Jersey Institute of Technology, Rutgers University

ABSTRACT: Understanding animals’ use of habitat often steers their conservation and management. Recently, models of species-habitat relationships have begun to incorporate numerous spatial scales to account for the fact that individuals may perceive their environment at scales that differ from the available data (e.g., satellite imagery). Building these models becomes more challenging when individuals confront different habitat choices due to the heterogeneous nature of their surrounding environment. We identified habitat preferences and generated habitat quality maps for brown bears (Ursus arctos) on the Kenai Peninsula (Alaska, USA) by first generating multiple spatial scales for each environmental variable. We then ranked each predictor variable in order of importance using the least absolute shrinkage and selection operator (lasso), which enabled us to fit models of increasing complexity and compare them using Akaike Information Criterion (AIC). Bears responded to their surroundings across multiple spatial scales in all seasons and regions. When information describing preferences was aggregated across multiple bears, it revealed patterns of spatial similarity useful for predicting the areas bears use and their habitat characteristics. Such predictions can help wildlife biologists manage bear populations by reducing encounters between bears and humans on the Peninsula. Our method of combining lasso with an information-theoretic approach via AIC is the preferable method for modeling animal habitat preferences when numerous variables are included in the modeling exercise. This method is widely applicable to studies incorporating many predictor variables.

Wednesday April 11, 2018 3:45pm - 4:00pm CDT
LaSalle 1 (7th Floor)