Human-bear interactions in the North Slope oilfields of Alaska (USA): characteristics of grizzly bear sightings and use of infrared for bear den detection

Abstract

Minimizing unsafe human-bear (Ursus spp.) interactions in the North Slope oilfields of Alaska (USA) requires knowledge of where they occur and methods to prevent them. My research goals were to characterize the spatial and temporal dynamics of grizzly bear (U. arctos) sightings during the non-denning season around industrial infrastructure in the North Slope oilfields over the past 25 years (Chapter 2), and to evaluate the efficacy of forward-looking infrared (FLIR) systems to detect grizzly bears and polar bears (U. maritimus) in their winter dens (Chapter 3). I used reports (n = 2,453) of summer grizzly bear sightings collected by oilfield security officers from 1990-2014 to estimate how the spatial distribution of sightings for food-conditioned (FC) and natural food (NF) bears changed following restriction of bear access to anthropogenic food waste (to be known hereafter as "treatment") in 2001. I found that concentrations of FC bear sightings shifted toward the landfill with medium-low effect (Hedges' g = 0.41), one of the only remaining areas with available food waste, after the treatment. The treatment also decreased NF bear sighting distances to landfill with low effect (Hedges' g = 0.15). My findings suggested that grizzly bear access to food waste should be prevented to minimize negative human-bear interactions and that an active bear reporting system facilitates adaptive management of human-bear interactions. During the winter, grizzly bears and pregnant female polar bears enter dens in areas that overlap anthropogenic activity. FLIR techniques have been used to locate occupied dens by detecting heat emitted from denned bears. However, the effects of environmental conditions on den detection have not been rigorously evaluated. I used a FLIR-equipped Unmanned Aircraft System (UAS) to collect images of artificial polar bear (APD) and grizzly bear (AGD) dens from horizontal and vertical perspectives from December 2016 to April 2017 to assess how odds of detection changed relative to den characteristics and environmental conditions. I used logistic regression to estimate effects of 11 weather variables on odds of detection using 291 images. I found that UAS-FLIR detected APDs two times better than AGDs, vertical perspective detected 4 times better than horizontal, and that lower air temperatures and wind speeds, and the absence of precipitation and direct solar radiation increased odds of detection for APDs. An increase of 1°C air temperature lowered the odds of detection by 12% for APD, and 8% for AGDs, but physical den characteristics such as den snow wall thickness determined detectability of AGDs. UAS-FLIR surveys should be conducted on cold, clear days, with calm winds and minimal solar radiation, early in the denning season. UAS-FLIR detectionof bear dens can be effective but should be confirmed by a secondary method.