Aversive conditioning of grizzly bears produces high probabilities of retreat from human-bear conflict locations

• Author / Creator

  Edwards, Claire

• Parks and protected areas provide important refugia for source populations of threatened grizzly bears in Alberta, where high human-use and recreation can cause human-bear conflict. Many jurisdictions in Alberta and beyond use hazing and aversive conditioning (hereafter AC) programs to deter bears from potential conflict, but there are few established standards with which to design and evaluate these programs. I examined data from an AC program in Kananaskis Country that occurred from 2000 to 2019 to (a) describe and summarize the scope of the program, (b) quantify the type and frequency of tools used to treat bears and (c) evaluate correlates of conditioning success, measured as retreat by bears from locations of potential conflict as conditioning technicians arrived, immediately following each action, and after the conditioning event (consisting of a series of 1-5 conditioning actions) was completed. AC events employed four conditioning tool types, each with two or three deployment methods that included noise (vehicle, human, Karelian bear dog), approach (vehicle, human), projectiles (no-contact, contact) and pursuit (human only, human and Karelian bear dog). In addition to conditioning variables, I evaluated several demographic and contextual variables and accounted for the identity of individual bears.

  In the 20 years of data I analyzed, technicians applied AC to 48 marked bears and an unknown number of unmarked bears in 6,539 conditioning events with an average of 104 events per marked bear. When bear identity was known, conditioning was over 50 times more likely to be applied to adult females than adult males, and in 99% of events where females were accompanied by cubs, cubs were young of year or yearlings. The frequency of conditioning events significantly declined with bear age. Conditioning events most often targeted bears that were feeding on natural vegetation (72%) and conditioning locations were most often at roadsides (65%) and campgrounds (13%).

  Among 3,613 events on 39 marked grizzly bears when a response to technician arrival was recorded, the average likelihood of retreat by bears was 32%. As the arrival response, retreat probability increased with the number of actions in the previous conditioning event, the number of conditioning events in the preceding two weeks, and the presence of cubs; retreat likelihood upon technician arrival decreased when noise or projectiles were used in the previous event, when bears were resting or feeding, and with increasing bear age. Among 4,959 events on 46 marked individuals where a response to conditioning was recorded, bears almost always retreated from conditioning technicians (93%) and rarely approached them, either upon technician arrival (1%) or after conditioning commenced (<0.001%). Conditioning events consisted of an average of 2.2 conditioning actions with a total of 14,323 actions, comprised in rank order, by noise (49%), approaches (25%), projectiles (21%) and pursuits (5%). Bears were more likely to retreat from entire conditioning events when the pursuit tool was used, when there were more actions in the event and with increasing distance to cover. Retreat likelihood decreased when noise tools were used, distance to the technician increased, or cubs were present. Within tool types, bears were equally likely to retreat from all three noise deployment methods (human, vehicle, or Karelian bear dogs) and from both approach types (human and vehicle). However, bears were significantly more likely to retreat from contact than no-contact projectiles and from pursuit with humans alone than pursuits with humans and Karelian bear dogs. Responses to both technician arrival and conditioning events were highly variable among individual bears, with bear ID accounting 30-60% of the variance explained by these models.

  The distribution of bears and conditioning events in our study suggest that in Kananaskis Country, female bears are disproportionately drawn to feed on natural vegetation, where they can avoid conspecifics via use of a human shield. Conditioning results suggest that bears are more likely to retreat from tools with higher aversive intensity and that they learn over time to avoid areas where conditioning is likely to occur, reducing their use of the human landscape during peak reproductive years for female bears. This, along with the other conflict management tools used in Kananaskis Country (bear-proof garbage, attractant management etc.), may contribute to the high reproductive success of this population. I suggest future AC programs to support human-wildlife coexistence should (a) consider the drivers of conflict (such as human shield dynamics and availability of natural attractants) on the landscape when setting program goals and defining program success metrics (such as an increase in wariness, or a decrease in developed site use), (b) develop conditioning and data collection protocols to adhere to learning theory and specific rules of effective punishment (such as higher intensity treatments and events), and (c) incorporate strategic planning recommendations of assessment and modification of program metrics on a regular basis to enable reporting of future programs.

• Subjects / Keywords

  • Aversive conditioning
  • hazing
  • human-bear conflict

• Graduation date

  Spring 2024

• Type of Item

  Thesis

• Degree

  Master of Science

• DOI

  https://doi.org/10.7939/r3-hjz2-nw96

• License

  This thesis is made available by the University of Alberta Libraries with permission of the copyright owner solely for non-commercial purposes. This thesis, or any portion thereof, may not otherwise be copied or reproduced without the written consent of the copyright owner, except to the extent permitted by Canadian copyright law.