Assessing Bear Response to Human Activity
at Kenai Fjords National Park:
It's Nature, Frequency and Costs

USGS PRINCIPLE INVESTIGATORS:

Tom S. Smith, USGS, Alaska Science Center, Biological Science Office, Anchorage.
Steve Partridge, USGS, Alaska Science Center, Biological Science Office, Anchorage.

NPS COLLABORATORS:

Terry DeBruyn, NPS, Alaska Support Systems Office, Anchorage.
Ian Martin, NPS, Kenai Fjords National Park and Preserve, Seward.

RESEARCH OBJECTIVES:

The focus of this study is to: determine the short term effects of land and sea-based human activity on coastal bears, investigate the long-term effects of human activity on coastal bears, and estimate the energetic costs of human displacement of bears.

This research will provide several products:

1. a field procedure for determining bear response to a variety of human activities occurring along coastal Alaska,
2. a profile of bear responses to the various types of human activity presently occurring at Kenai Fjords National Park (KEFJ),
3. an estimate of the frequency of bear-human interactions,
4. an estimate of the costs of human disturbance to bears,
5. recommendations for minimum approach distances for land and sea-based human activities,
6. recommendations regarding backcountry camping within coastal areas of KEFJ, and
7. a research publication that presents results of this work.

SPECIFIC RESEARCH TASKS:

The following tasks will be completed:

1. formulate a study plan (this document), circulate for review, revise and finalize
2. use the 2002 field season primarily for research protocol testing and revision,
3. incorporate 2002 findings into a final study plan, circulate for peer review, revise and file,
4. conduct bear-human interactions study in the summer of 2003,
5. analyze data and write final report by December 2003.

Importantly, the 2003 field season was used to test the feasibility of research methodology outlined in this study plan. Insights gained while working in the field in 2002 provide additional approaches that would strengthen the study, these will be added to the final study plan. The goal of this project is to extend our understanding of the nature, frequency, and energetic costs associated with bear-human interactions and suggest ways in which resource managers at KEFJ may effectively manage human activity in these important coastal areas.

PROJECT BACKGROUND AND JUSTIFICATION:

Steadily increasing human activity along Alaskan coastal areas has made it a high priority to determine how this activity may be affecting wildlife. Both the coastal brown bear (Ursus arctos) and American black bear (Ursus americanus) are of particular concern because of their reliance on coastal habitats as well as the threat they pose to humans. Resource managers need information regarding bear response to human activity in order to: 1) develop and establish management guidelines that will maintain 'natural and healthy' populations of bears; 2) minimize human disturbance, particularly displacement from key resources; 3) minimize the chances of bear conflict while allowing for public enjoyment of coastal areas; and 4) develop and implement coastal bear management plans. Research protocols for measuring bear response to a variety of human activities presently occurring in coastal areas (e.g., bear-kayak, bear-vessel, bear-campers, etc.) do not exist and this work seeks to create and implement them for collecting these data.

STUDY AREA:

Kenai Fjords National Park is an extensive wilderness encompassing 230,000 hectares (567,000 million ac) of tidewater glaciers, timbered islands, winding fjords and a unique assemblage of marine and terrestrial life (Fig. 1). Coastal mountains ascend from the ocean to over 1,300 m (4,000 ft), with rock, ice and barren terrain the largest components of the terrestrial ecosystem. Consequently, KEFJ's most productive terrestrial habitats are immediately adjacent to the sea, sandwiched between ascending mountains and beach. This area, heavily used by bears and people, is the focus of this work.

RESEARCH APPROACH AND HYPOTHESES:

Research Approach

Sea kayaking is the predominant recreational activity in KEFJ's extensive marine backcountry. Kayakers often stay several nights, camping in the narrow belt of supra-tidal land at the base of steep-walled fjords. Both brown and American black bears inhabit KEFJ, although black bears are numerous while brown bears rare. These bears’ seasonal activities are concentrated in low elevation, coastal areas because beaches not only provide convenient movement corridors, but also marine and terrestrial foraging opportunities. Consequently, the potential for bear-human interactions, as well as the likelihood that human activity may displace bears from important forage resources or interfere with their travel, is highest in these coastal areas. By devising, testing and applying research protocols that specifically measure bear responses to human activities, this research effort will assemble information regarding the nature of bear-human interactions occurring in coastal areas with regard to the following questions:

1) What are the short-term effects of human activity (e.g., camping, kayaking, and sightseeing on vessels, etc.) on coastal bears?
2) What are the long-term effects of human activities on resident bear populations?
3) What are the energetic costs of human disturbance to coastal bears?

To assess bear responses to a variety of human activities, this research builds upon the work of a number of scientists addressing similar human disturbance concerns for a variety of species: nesting bird colonies (Rodgers and Smith 1995), flamingos (Galicia and Baldassare 1997), mule deer (Freddy et al. 1986), eagles (Grubb and King 1991), and seals (Boren et al. 2001, Lewis and Mathews 2000), to name a few. These approaches either passively observed, or actively perturbed, these species and recorded their responses to an array of human-related stimuli (e.g., a variety of aquatic and terrestrial vehicles, aircraft, and pedestrians). In this work we plan to combine passive, opportunistic observations of bear-human interactions along with intentional direct approach scenarios in order to elicit, and measure, bear responses to a variety of human activities.

In this research, individual bears are the experimental units; the dependent variable is the bear's response associated with each interaction; other variables expected to influence a bear's response are independent variables and include: type of activity (e.g., human activity mode; distance between people and bear; speed of approach to the bear; directness of approach to bear; number of people involved in the interaction; boisterousness/loudness of the people present; bear cohort involved; bear species involved; the bear's previous history of interactions with people; time of day; and season of year to name a few of the likely main factors that influence the bear's response). The population that will be investigated by this work will be those bears that researchers are able to opportunistically encounter, approach and record data.

Research Hypotheses

This research will attempt to test the following hypotheses (H), associated predictions (P), assumptions (A), and limitations (L):

H1: Bears respond predictably to specific human activities.

This hypothesis suggests a number of predictions (P) that can be tested with observations in the field:

P1: Bear responses to each human activity mode will be consistent in the following ways:

1. each mode of human activity will elicit predictable, and unique, patterns of responses from bears (e.g., sea kayaking may prove to be consistently more disruptive to bears than large cruise vessels),
2. the louder and more boisterous people are, the more intense the bear response,
3. the distance between bears and people is expected to have a marked effect upon the resulting response; the shorter the distance the greater the response,
4. the larger the group of people the more likely the bear will respond more intensely (i.e., larger groups will affect bears at greater distances than smaller groups),
5. direct approaches to bears will elicit a greater reaction than other movements made by people,
6. females with cubs will respond more strongly than single bears,
7. brown bears will respond more strongly than American black bears,
8. the more open and exposed the environment in which the bear-human encounter occurs, the more strongly the bear's response is expected to be, and
9. whenever the bear initiates the interaction, the response will be less intense than when initiated by people. For example, if a bear walks into a camp it will respond less strongly to people than if hikers walk up to a bear.

The following assumptions (A) are associated with this research approach and hypothesis:

A1: Bear responses to specific classes of stimuli (e.g., approach by kayaks) are genetically controlled to a large degree, hence consistency in response will be observed within species and age-cohorts, and with respect to classes of independent variables.

A2: Inconsistencies in response to human activity (e.g., bears of the same species and age-sex cohort may allow close approaches while others do not) will occur, will be the result of genetic variation and/or learning, and will be uncommon.

A3: Research protocols are appropriate for measuring bear response in an accurate and consistent manner.

A4: Observable behaviors are an accurate indicator of stress in coastal bears.

The following limitations (L) are associated with this research approach and hypothesis:

L1: Researchers are unable to assess the inner state of the bear, recording only observable changes in their behavior to human activity. However, bears interacting with people may be highly stressed yet not display it outwardly.

L2: The unknown histories and genetic variation of individual bears may confound results.

L3: Bear-human interactions may be too difficult to obtain for statistically sound conclusions to be drawn.

H2: Bears exposed frequently to human activity will habituate, that is, their response to anthropogenic stimuli will wane.

This hypothesis suggests a number of predictions (P) that can be tested with observations in the field:

P1: Bears in areas not frequented by people (e.g., Nuka Bay) will respond much more strongly to human activities than those in areas heavily used by people (e.g., Aialik Bay).

The following assumptions (A) are associated with this hypothesis:

A1: Bear-human interactions are largely benign, hence habituation will occur.

A2: Observable behaviors are an accurate indicator of stress in coastal bears.

A3: Research protocols are appropriate for measuring bear response in an accurate and consistent manner.

The following limitations (L) are associated with this research approach and hypothesis:

L1: Researchers are unable to assess the inner state of the bear, recording only observable changes in their behavior to human activity. However, bears interacting with people may be highly stressed yet not display it outwardly.

L2: The unknown histories and genetic variation of individual bears may confound results.

L3: Bear-human interactions may be too difficult to obtain for statistically sound conclusions to be drawn.

PROCEDURES:

Two populations of bears will be studied: one with high levels of human activity (Aialik Bay), and one with negligible levels of human activity (Nuka Bay). We will assess the effects of a number of different modes of human activity on these populations of bears using a combination of behavioural observations and experimental approaches.

Measuring Bear Response to Direct (Intentional) Approaches
By varying the activity modes (i.e., by land, kayak or motor boat), the number of people involved, noise level during an approach, the distance and angle of approach, a series of intentional approaches will be conducted.

Approaches on Land
By approaching bears on the coast on foot, we will record two measurements: 1) the distance at which the bear first noted the presence of people and 2) the distance at which persons’ approach elicited a response that represents a change in the bear’s behavior. To accomplish this, research crews will monitor coastal areas by boat in search of bears that can be approached on foot. For consistency, crews who approach bears will dress similarly for all trials, likely in forest green raingear so that variation in sight-ability is minimized. Once a bear is located, biological technicians will be put on the beach, preferably out of sight of the bear and at a distance > 100 m. The crew will slowly (< 4 km/h) approach the bear until it first responds to their presence, or until it is felt it is not safe to continue. The age class, sex, and activity of the test animal prior to the approach will be recorded, as well as the distance between the bear and the persons when the bear first responds to their presence. The distance at which the bear first responds to the approaching persons is defined as the moment the bear is clearly aware of their presence, as indicated by its fixed gaze. The distance (m) between people and bear will be determined by use of a laser range finder (LRF). After the initial response to the approach, the crew will continue approaching until a change in the bear’s behavior is elicited, or until the crew felt it was unsafe to approach any closer. The bear’s response will be recorded and ranked:

Neutral Response - the bear is aware of the approaching persons, as determined by having gazed momentarily in their direction but there is no physical response (e.g., no movement of head, ears or redirection of movement path).

Slight Response - the bear is clearly aware of approaching people (e.g., lifts its head or reorients ears), and changes its activity mode (e.g., ceases to forage, gets up from resting position, etc.).

Moderate response- the bear is clearly aware of people, changes activity mode, and either moves steadily towards or away from approaching persons.

Strong response- bear is clearly responding to the presence of people, changes activity mode and quickly leaves area or approaches people.

For safety, the minimum group size to be tested will be 2 persons, and both people will be equipped with bear deterrents (pepper spray and signal flares, both highly effective for discouraging curious/pushy bears) and trained in their use. Details regarding the approach, time for the bear to return to is pre-disturbance behavior, and any additional factors that may have influenced the bear’s response (e.g., windiness, other bears nearby, etc.), will be recorded.

Effect of Group Size on Bear Responses
We will test the effect of group size for groups approaching the bear on land. Group size (2-6) will be randomly determined prior to approach. The age class, sex, and activity of the test animal prior to the approach will be recorded, as well as the distance between the bear and the persons when the bear first responds to their presence. Criteria for determining the bear’s initial response and for strength of response is the same for all approaches and is detailed above.

Ocean Approaches
Bears on beaches, or in intertidal areas, will be intentionally approached in watercraft (e.g., kayaks, small (< 7 m length) and large (>7 m length) motorized boats), to determine the range and intensity of bear responses associated with each of these modes of human activities. The age class, sex, and activity of the test animal prior to the approach will be recorded, as well as the distance between the bear and the persons when the bear first responds to their presence. The distance at which the bear first responds to the approaching vessel is defined as the moment the bear is clearly aware of their presence, as indicated by its fixed gaze. The vessel will continue to approach the bear until a response is elicited. That response will be graded according to criteria outlined above (i.e., ranging from neutral to strong).

Noise Level of Approaches
We will examine how noise levels influence bear responses to approaching persons and vessels by altering the noise levels associated with approaches. Although a great number of potential noise scenarios exist, a simplistic approach must be taken for determining the influence of noise given the limited number of approaches that will be made throughout the testing period. Therefore, we will limit the testing to either quiet approaches (no talking) or noisy approaches (loud talking between persons, approximately 85 dB). Prior to each approach a coin toss will determine whether a noisy or silent mode will be used. The distance to the bear when it becomes initially aware of the approaching researchers, as well as the distance to first response, will be recorded using the LRF. Similarly, when testing the effect that approaching vessels have on bears using adjacent coastal areas, a coin toss will determine whether the approach will be noisy or silent.

Classification trees and/or other multivariate statistics will be applied to bear response to directed approaches data to determine if disturbance due to human presence is occurring, what form of human use is the most disruptive, and at which distances disturbance is greatest.

Measuring the Effect of Campsites on Bear Movements, Use and Behaviour

Backcountry Campsite Use
In order to determine what effect campsites may have on bear activity, levels of camper and bear activity will be monitored using remote videography. Areas to monitor that meet the following criteria will be selected: 1) the area will support a relatively high level of bear activity, 2) the area will be popular with campers, and 3) a camera can be positioned so that bear activity within the general area surrounding the campsite, as well as the campsite itself, can be monitored. Selected sites will be monitored with video cameras using time-lapse intervalometers (Sony DCR-TRV900), powered by 12-volt storage batteries kept charged with passive solar panels. The camera will be set so that it records the date and time on each frame of videotape exposed.

Cameras will be positioned so that the campsite and surrounding area (> 50 m on each side) are monitored. Cameras will be set at a resolution so that the presence of people and bears can be determined, but individual persons (e.g., John Jones of Anchorage, Alaska) and their specific activities (e.g., cooking dinner) cannot be identified, thus minimizing the invasive nature of remote video monitoring. Intervalometers will be set to record 6 frames/minute for the duration of the campsite observation period. At this frame capture rate a standard 120-minute VHS format videocassette will record continuously for 600 hr (25 days), thereby minimizing the need for upkeep and maintenance. Although we may opt to deploy cameras in popular bear and camper use areas, we may also choose to select a site known for bear activity but not for campers then put our own ‘camp’ there so that we can control precisely the time before, during and after people were present. If we choose to use that approach, the video monitoring period will be 3 weeks long: a week of observation prior to campsite use, a week of observation during use, and a week of observation after campsite use.

Video images will be analyzed for: 1) temporal patterns of site use by people and bears, 2) spatial use of the site by people and bears, and 3) activity budgets of bears with, and without, people present. Additionally, differences in total number of bears, total number of bear minutes of use (i.e., sum of all video frames with bears present, converted to time), use by time of day, or in bear activity budgets will be considered as measures of bear response to human activity (e.g., disturbance). Campsites in areas of light and heavy human use will be observed. Non-parametric tests will be used to test the relative presence and absence of people and bears, as well as how the presence of people affects bear activity and time spent at camera locations.

Moored Vessels
The impacts that vessels moored near beaches may have on bear activity will be addressed by this research. Time-lapse video cameras will be used to determine levels of bear activity on adjacent coastal habitats in the presence, and absence, of moored vessels. Vessel distance from shore, the size of vessel and other variables that may have an effect on bear activity (e.g., tidal stages, weather, etc.) will be recorded. A camera will be positioned in an area known for bear activity and for harboring vessels and run for 3 weeks at a time from June through August. Periodic checks will determine that the unit is functioning properly and allow for the tape to be replaced.

Video images recorded for this aspect of the project will be analyzed for: 1) temporal patterns of mooring site use by people and bears, 2) spatial use of the site by people and bears, and 3) activity budgets of bears with, and without, people present. Differences in total number of bears, total number of bear minutes of use, use by time of day, or in bear activity budgets will be considered as measures of bear response to human activity (e.g., disturbance). Non-parametric tests will be used to test the relative presence and absence of people and bears, as well as how the presence of people affects bear activity and time spent at camera locations.

Focal Scan Sampling of Bear Activity within Beach and Intertidal Zones
This effort will provide an estimate of the energetic costs of human disturbance by determining 1) activity budgets for bears in the presence, and absence, of people, 2) the nutrient value of a few selected bear forages, and 3) the rate at which bears ingest these foods (biomass ingested/unit time). The activity budgets of bears foraging in coastal habitats will be determined by focal scan sampling procedures (Altmann 1974). Observed bear activity will be entered into The Observer® behavioral software running on a palmtop PC computer. The specific age-sex cohort of the bear and the distance separating people and bears will be recorded. A selection of bear food items will be collected to determine gross energy, percent protein and digestibility. Laboratory techniques will follow accepted standards of proximate analysis (AOAC 1980) for determining crude protein (6.25 by Kjeldahl nitrogen) and gross energy content. The acid detergent fiber (ADF) content (Waldern 1971) will be determined to estimate forage digestibility (Bunnell and Hamilton 1983). Bear forage intake rates will be estimated by establishing an average intake per unit time for a limited number of food items. This will be accomplished by either direct observation (i.e., establishing bite rates for specific forages), or by investigating areas of where bears have been actively foraging and determining the biomass ingested by estimated the amount removed.

Analysis of focal observation data will indicate how bears partition their time in the presence, and absence, of people. Differences in foraging efficiency will be attributable to human disturbance and will provide a measure of the energetic costs of disturbance.

LITERATURE CITED

Altmann, J. 1974. Observational study of behavior: sampling methods. Behaviour 49:227-267.

Association of Official Analytical Chemistry (AOAC). 1980. Official methods of analysis, 13th ed. AOAC, Wash. D.C. 1018 pp.

Boren, L.J., K.J. Barton and N. J. Gemmell. 2001. Assessing the impacts of tourism on New Zealand fur seals (Arctocephalus forseri). Contact rep. num. LC0102/017. Dep. Conserv., Nelson Conservatory, Nelson, New Zealand. 51 pp.

Bunnell, F. L. and N. Hamilton. 1983. Forage digestibility and fitness in grizzly bears. Int. Conf. Bear Res. and Manage. 5:179-185.

Freddy, D. J., W. M. Bronaugh and M. C. Fowler. 1986. Response of mule deer to disturbance by persons afoot and snowmobiles. Wildl. Soc. Bull. 14:63-68.

Galicia, E. and G. A. Baldassare 1997. Effects of motorized tourboats on the behavior of nonbreeding American flamingos in Yucatan, Mexico. Conserv. Biol. 11(5):1159-1165.

Grubb, T. G. and R. M. King 1991. Assessing human disturbance of breeding bald eagles with classification tree models. J. Wildl. Manage. 55(3):500-511.

Lewis, T. M. and E. A. Mathews. 2000. Effects of human visitors on the behavior of harbor seals (Phoca vitulina richardsi) at McBride Glacier Fjord, Glacier Bay National Park. Final Rept. Glacier Bay N.P., Res. Mgt. Div. 22 pp.

Rodgers, J. A. and H. T. Smith. 1995. Set-back distances to protect nesting bird colonies from human disturbance in Florida. Conserv. Biol. 9(1):89-99.

Waldern, D. E. 1971. A rapid microdigestion procedure for neutral and acid detergent fibre. Can. J. Anim. Sci. 51:67.