Responses of wildlife to tourism and glacial recession in Glacier Bay National Park, Alaska

Abstract

Wildlife have varying responses to disturbances depending on the duration, severity, and type of disturbance event. Some disturbances modify wildlife habitat, and can impact community assembly and patterns of diversity, while others can modify wildlife behavior. Human disturbance often elicits two opposite behavioral response from wildlife, one in which they can exhibit “fear effects”, where they avoid humans, and the “human shield effect”, where wildlife are attracted to centers of human activity and use human presence as a buffer against predation. However the level of human disturbance that causes a detectable change in wildlife behavior remains unknown. Many disturbances that alter landscapes and modify wildlife habitat are expected to increase in frequency and severity with climate change, and post-disturbance successional patterns of wildlife communities remain poorly understood. In this thesis, I investigated the responses of wildlife to two types of disturbances: human activity (Chapter 1) and glacial recession (Chapter 2) in Glacier Bay National Park, Alaska (GLBA). GLBA is a remote park with relatively low but increasing levels of visitation, and it is centered around a marine fjord that is the product of the most rapid glacial recession in modern times. To understand the responses of brown bears (Ursus arctos), black bears (Ursus americanus), moose (Alces alces), and wolves (Canis lupis) to human activity, I used camera traps to document wildlife activity and a paired-plot, crossover experimental design to manipulate human visitation during summers 2017 and 2018 (n = 5 pairs of sites). Single-season occupancy models and activity overlap analyses indicated that brown bears were unaffected by human use in GLBA, black bears exhibited fear effects temporally but the human shield effect spatially, moose utilized human presence as a shield temporally, and wolves exhibited fear effects temporally. Traditional occupancy models assume logit-linear relationships between predictor and response variables; however, visual inspection of the detection data showed a threshold in which detections did not exceed four per week for any species unless human activity was absent. The camera trap data was supplemented with small mammal trapping and vegetation surveys to investigate community-level successional patterns of mammals in GLBA and understand the roles of habitat and time-since-disturbance in shaping community assembly and diversity. Mammal communities in GLBA were more strongly influenced by time-since-disturbance than habitat, and dispersal ability had a strong influence on mammal colonization patterns, as indicated by both PERMANOVA and beta diversity analyses. My findings have several implications: first, despite low visitation to GLBA, I detected spatiotemporal responses of wildlife to human disturbance, indicating that the threshold level of human activity for sites to function as baseline controls in studies of anthropogenic impacts may be lower than those found in most protected areas. Additionally, these results support “land sparing” management techniques whereby visitors are concentrated in certain parts of protected areas to maintain spatial refugia for wildlife. Second, these results highlight the importance of incorporating landscape connectivity and dispersal ability metrics into wildlife conservation efforts following disturbances. Furthermore, caution should be used making inferences about changes in wildlife communities following disturbance based on changes in habitat, as my results indicate that wildlife and vegetation communities do not respond to disturbance in the same way. This knowledge may improve predictions of mammalian community assembly following major disturbance events.