-
Title
-
Eng
Estimating Black Bear Population Size in Gustavus, Alaska: Implications for Determining the Effect of Human Caused Mortality on Population Size
-
Date
-
2013
-
Creator
-
Eng
Pinjuv, Kyle
-
Subject
-
Eng
Environmental Studies
-
extracted text
-
ESTIMATING BLACK BEAR POPULATION SIZE IN
GUSTAVUS, ALASKA: IMPLICATIONS FOR DETERMINING
THE EFFECT OF HUMAN CAUSED MORTALITY ON POPULATION SIZE
By
Kyle Pinjuv
A Thesis
Submitted in partial fulfillment
of the requirements for the degree
Master of Environmental Studies
The Evergreen State College
June 2013
©2013 by Kyle Pinjuv. All rights reserved.
This Thesis for the Master of Environmental Studies Degree
By
Kyle Pinjuv
Has been approved for
The Evergreen State College
By
________________________
Dina Roberts, Ph.D.
Member of the Faculty
________________________
Martha L. Henderson, Ph.D.
Member of the Faculty
_______________________
Tania Lewis, M.S.
Wildlife Biologist, Glacier Bay National Park
______________________
Date
ABSTRACT
Estimating Black Bear Minimum Population Size in
Gustavus, Alaska: Implications for Determining
the Effect of Human Caused Mortality on Population Size
Kyle Pinjuv
Wildlife managers rely on accurate population estimations to contribute to the working
knowledge of game and non-game animals. Until recently, there have been no regional
population studies on black bears (Ursus americanus) in southeast Alaska, yet harvest
rates are set at 10% of the total population. This research is based on an in-depth
literature review of the history of bear harvest and the use of noninvasive genetic
sampling and mark-recapture methodology. Noninvasive genetic sampling techniques
using rub trees and hair traps were used during the spring, summer, and fall months of
2011 and 2012 to estimate the minimum number of black bears within the Gustavus,
Alaska forelands. I collected harvest records from the Alaska Department of Fish &
Game and compared the population estimate with the average number of harvested bears
from 1990 – 2011 to determine if harvest rates were either set at unsustainable or overly
conservative levels. I collected 196 bear hair samples and marked 33 black bears and 14
brown bears over two field seasons with hair collected off of 25 rub trees and 8 scented
hair traps. Using the Huggins linear logistical model in program DENSITY, I estimated
the population of black bears to be 54.5 ± 10.3 (95% CI=41.6 – 84.8). The average
number of black bears killed by humans annually from 1990 – 2011 was 3.68 indicating
that the current harvest level meets the 10% harvest objective but future research is
encouraged to obtain more precise population estimates allowing wildlife managers to
monitor black bear population trends and continue to make responsible management
decisions.
TABLE OF CONTENTS
LIST OF FIGURES…………………………………………………………………….vii
LIST OF TABLES……………………………………………………………………..viii
ACKNOWLEDGEMENTS…………………………………………………………….ix
CHAPTER I: LITERATURE REVIEW
INTRODUCTION……………………………………………………..………….1
BEAR HARVEST HISTORY………………………………………...…………..5
BEAR HARVEST TRENDS IN GMU01C……………………………………….8
HISTORY OF NONINVASIVE GENETIC METHODOLOGY OF BEAR
SPECIES………………………………………………………………………....13
NONINVASIVE SAMPLING METHODS……………………………..13
STUDY DESIGN TECHNIQUES………………………………………15
OPEN & CLOSED POPULATION MODELING FOR MARK-RECAPTURE
ANALYSIS……………………………………………………………………....17
CONCERNS WITH GENETIC SAMPLING…………………………….……..20
CONCLUSION……………………………………………………………..……21
CHAPTER II: BLACK BEARS IN GUSTAVUS, ALASKA: POPULATION AND
HARVEST MANAGEMENT.
INTRODUCTION………………………………………………………..…..….24
STUDY AREA…………………………………………………....….….27
METHODS………………………………………………………………...….…29
FIELD TECHNIQUES………………………………………………..…29
METHODS OF ANALYSIS……………………………………….....…34
GENETIC ANALYSIS……………………………………….....34
STATISTICAL ANALYSIS…………………………………….35
HARVEST RECORDS………………………………………….37
RESULTS..............................................................................................................37
SAMPLING EFFORT…………………………………………………...37
GENOTYPING SUCCESS…………………………………………..….38
MARK-RECAPTURE ANALYSIS………………………………..…....38
HARVEST DATA…………………………………………………....….40
DISCUSSION…………………………………………………………………....43
STATISTICAL ANALYSIS AND STUDY DESIGN…….…………….44
HARVEST RATES………………………………………………….…..46
MANAGEMENT IMPLICATIONS………………………………………….…46
CHAPTER III: CONCLUSIONS AND RECOMMENDATIONS…………………...…49
RECOMMENDATIONS…………….................................................…………..50
INTERDISCIPLINARY EFFORT…………………………………………....…51
LITERATURE CITED………………………………………………………………......54
vi
LIST OF FIGURES
FIGURE 1.1 GMU 01C-01D. Game management unit 01C - an area comprising over
1,700 square miles of high quality bear habitat and containing Federal (Glacier Bay
National Park), State, and private lands (ADF&G n.a.)……….………….……………11
FIGURE 1.2 Numbers of black and brown bears harvested by year in GMU 01C, 19962011 (Neil Barten, Ryan Scott ADF&G)..……………………………...………….……12
FIGURE 1.3 Number of bears killed by year in Gustavus by category: road kill, defense
of life or property (DLP), or harvested/hunted, 1996-2011 (Neil Barten, Ryan Scott
ADF&G)…………………………………………………………………………………12
FIGURE 2.1 Study Area - Gustavus forelands with 4 km² grid overlay. Total area within
the study area is 200 km². Source: ESRI software, ArcGIS………………......................33
FIGURE 2.2 Rub Tree and Trap Locations - 8 Scented hair traps were deployed May,
2012. One hair trap was deployed within each 4 km² cell. 25 rub trees were equipped
with barbed wire throughout the study area June and July of 2011……………………..34
FIGURE 2.3 Number of black bears harvested in Gustavus, Alaska by year. The trend
line showed a slight increase in harvest over the course of 21 years with an R² value of
0.032 (ADF&G 2011b). Even with the large harvest in 2002 removed from the data, the
trend line continued to show a slight annual increase…………………………………....41
vii
LIST OF TABLES
TABLE 2.1 Number and dates of sampling sessions for rub trees and baited hair traps for
2011 and 2012. Each sampling session was 10-14 days in length where all rub trees and
hair traps were checked one time per session. Session 1 and 8 were also representative of
installation dates of rub trees and hair traps……….……………………………………..32
TABLE 2.2 Variability of microsatellite markers used to determine individual identity of
black bears in the Gustavus, Alaska forelands…………………………………………...39
TABLE 2.3 Capture results for trapping sessions 1 – 17 from the data set including 2011
and 2012 and capture results for trapping sessions 1 – 10 from the data set including only
2012. 32 total black bears were captured between the two sampling years of 2011 and
2012 and 25 total bears were captured during the single sampling year of 2012. There
was only one marked bear harvested and sealed by ADF&G throughout the sampling
seasons in 2012……………………………………..……………………………………40
TABLE 2.4 Bear harvest rates by category in Gustavus, Alaska 1990-2011. The total
number of bears killed in the last 21 years was 81 with the average of 3.68 bears killed
annually. (ADF&G 2011b)………….………………………………………….…...…42
viii
ACKNOWLEDGMENTS
My sincerest appreciation goes out to all those who helped me through this
process. The list is long, but every person contributed a great deal to the success of this
project. A very special thank you to Tania Lewis for the support as a supervisor and
friend, and to Dina Roberts and Martha Henderson for helping me through the writing
process and giving me continued encouragement.
I would also like to thank all those who volunteered with me during the field
season in Alaska collecting bear fur off rub trees and hair traps. Thank you to Christopher
Behnke, Leslie Skora, Kevin Colson, Ryan Scott, Kevin White, Anthony Crupi, Lewis
Sharman, Lisa Etherington, T.J. Flanagan, Eric Riordan, Sarah Betcher,Laurel Carver,
Adam Clardy-Oneal, J.T. Mclaughlin, Mike Stone, Andrew Gertge, Hiroko Yoshii,
Sharon Grant, Barbara Morford, Barb Bruno, Amy Brodbeck, Neil Barten, Rob Fisk,
Gene Fiebich, Christina Rinas, Shahed Dowlatshahi, Diana Raper, Patrick Mershon,
Lauren LaRocca, Kaylin Elder, Katie Unertl, Stephanie Blumhagen, Kelley Graham, and
my incredibly supportive parents John and Debbie Pinjuv for all of your help!
And of course a great deal of gratitude must go out to the bears who so kindly let
me arm their rub trees with barbed wire and let me draw them into barbed wire corrals
with no reward aside from the smell of something tasty. I couldn’t have done it without
their participation.
ix
I: LITERATURE REVIEW
Introduction
For over a century, wildlife managers have increasingly sought to understand
population dynamics of game and nongame animals in order to sustainably manage these
populations faced with ever-growing threats. Human caused mortality and multiuse
demands of these natural resources has created a need to obtain more rigorous data on
population size, rate of population growth, and age and sex distributions. Obtaining
quantitative data to inform the impacts of humans on game and nongame animals, as well
as practicing adaptive management as a systematic approach to sustainably managing
populations is vital for the conservation of wildlife species. The emerging need for
quantitative data to inform wildlife managers on population trends has led to the
development of management actions such as bag limits and seasonal hunting restrictions
on game animals. More recently, wildlife viewing and other non-consumptive use of
natural resources has added to the value of game and nongame wildlife found throughout
the North American continent, especially as many of these species become increasingly
rare.
In Southeast Alaska, bears are widely valued for their general biological, cultural
and economic importance to the region. But until recently, the specific population trends
of these animals have been unknown throughout most of the region. One early estimate
found 4,155 brown bear populations on Chichagof, Baranof, and Admiralty islands, an
area comprising of 6,115 km². This estimate comprised approximately 70% of the total
population of brown bears in Southeast Alaska (Titus and Beier 1993). Research shows
1
brown bear density on Admiralty and Kodiak islands to be some of the highest in the
world, at 2.6 bears per km² (Schoen and Beier 1990). In contrast, studies on brown bears
on the North Slope, located in northern Alaska, show densities as low as 1 bear per 300
square miles (ADF&G 2013).
Compared to the growing number of studies on brown bear populations, black
bear populations have not been widely studied in Alaska. Peacock (2011) was the first to
estimate black bear populations on Kuiu Island in Southeast Alaska concluding that
populations are probably larger than originally thought. The population of black and
brown bears in Glacier Bay National Park (GLBA) and the surrounding areas has never
been determined (National Parks Conservation Association 2008). The Alaska
Department of Fish & Game (ADF&G) set black and brown bear harvest rates at a
certain percentage of the total population in designated management areas, but without
current population research it is possible to set the harvest rate at either unsustainable or
overly conservative levels. Information collected from harvested bears could offer
valuable insight about population trends and harvest sustainability if used in conjunction
with empirical population estimates (Barten 2004). Black bear estimates for the region
are based on research conducted in the 1960’s and 1970’s in Washington State (Poelker
and Hartwell 1973). ADF&G also uses anecdotal information from hunters and harvest
data to assess black bear populations (ADF&G 2011a). While both of these sources
provide valuable insight into bear population trends, quantitative data about bear
populations in Southeast Alaska would prove invaluable in the bear management decision
making process. Kendall (2008) discusses the value of estimating population trends as
2
they correlate to landscape changes, and how understanding these changes will allow
managers to “reverse negative, or enhance positive changes in the population.”
Glacier Bay National Park lies within one of the largest protected wilderness areas
in the world (National Park Service 2010). Wildlife within GLBA is managed by the
National Park Service (NPS) in accordance to the Organic Act of 1916 (National Park
Service n.d.). Both black and brown bears frequently move beyond the boundaries of the
National Park or protected area and the State lands, where most hunting occurs, requiring
interagency cooperative bear management. Wildlife on the adjacent private and state land
to the south of GLBA is under the management of ADF&G. Bears are important for
wildlife viewing opportunities within GLBA as well as provide sport hunting
opportunities in areas surrounding the park, such as adjacent to the town of Gustavus.
This small town, located approximately 40 miles west of Juneau, Alaska, has a yearround population of approximately 450 people (U.S. Census Bureau 2010), which
increases dramatically during the summer due to the high rate of seasonal employment.
Tourism, including wildlife viewing, is a large industry for Gustavus because it is the
gateway to Glacier Bay National Park and also offers hunting and wildlife viewing
opportunities in the area.
In order to maintain sustainable populations of bears for both wildlife viewing and
hunting, both State and Federal bear managers are concerned about potential increases in
bear harvest in Gustavus due to increased access to the community via the Alaska ferry
system, as well as the recent transfer of land ownership at Falls Creek from federal
National Park land to State land. The previously protected Falls Creek area is a high
quality spring habitat for bears that is now open for hunting. Additional human caused
3
Defense of Life and Property (DLP) bear mortalities add to the number of animals
removed from the environment. Understanding bear populations and trends allows
wildlife managers to understand how these mortalities impact trends within the local bear
populations.
Within the past 15 years, many studies on black and brown bears have been
conducted throughout the North American continent. This research has provided
information on previously unknown populations and guided management (Bittner et al.
2002, Boersen et al. 2003, Boulanger et al. 2002, Kendall et al. 2008, Poole et al. 2001);
for Alaska, however, significant information gaps remain, and establishing precise
regional population estimates of black bears is a central objective for sustainable
management.
Recent advances in genetic tagging have allowed wildlife managers to gain
insight into bear populations using noninvasive methodology. Using a combination of rub
trees and baited hair traps, hair samples can be collected providing information on species
distribution, sex (Taberlet 1993), genetic population structure, and individual genealogies
(Woods et al. 1999). From these techniques, scientists can extrapolate population
densities, abundance, and trends using mark-recapture statistics. This is a noninvasive,
relatively inexpensive, and accurate method for determining minimum population counts
in comparison to live capture and collaring techniques (Stetz et al. 2010, Kendall et al.
2008). With the advancement of these new methodologies, research on bear and other
elusive species has increased dramatically. These methods can be applied in Southeast
Alaska allowing further insight into bear population trends within important management
units and inform management protocols.
4
The main objectives of this research were to explore the history of bear harvest
management in Southeast Alaska, with particular emphasis on documenting harvest
trends of 1990-2011 in Gustavus, Alaska. Following this a review of the development of
the noninvasive genetic sampling and mark-recapture analysis of bear species and its’
implications on bear management, potential management implications. Secondarily, to
determine a minimum population count and density estimate of black bears within the
Gustavus forelands using noninvasive genetic sampling, and, finally recommendations
for wildlife managers for sustainable harvest rates of black and brown bears in the
Gustavus forelands management zone.
Bear Harvest History
Throughout the United States and Canada bear hunting has been a common
practice in many communities. With westward European expansion, grizzly bear
populations in the lower 48 States were hunted almost to the point of extinction leaving
only small island populations in the states of Washington, Montana, Wyoming, and Idaho
(Knibb 2008). The benefit of grizzly bear recovery has been widely debated over the past
few decades and continues to be a contentious topic. The listing of grizzly bears as an
endangered species in 1975 ceased the hunting of the species in the lower 48 states, but
grizzly bear hunting in Alaska remains a large industry throughout the state. Black bears
continue to be hunted throughout the United States and Canada and continue to have
robust populations throughout a wider area of the continent. Although most black bear
ranges are limited to forested areas with low human density (Pelton 1982), they have also
adapted to living among humans when their presence is tolerated (Powell et al. 1997).
Most large carnivores are susceptible to anthropogenic influences due to their large home
5
ranges, smaller population sizes, long generation times, and increased interactions (often
with negative outcomes for the bears) with humans (Noss et al. 1996). There is a direct
correlation between surviving populations of grizzly bears and human scarcity (Mattson
and Merril 2002). Because of limited urban areas, high quality and plentiful habitat, and
miles of inaccessible terrain Alaska is home to some of the densest populations of both
black and brown bears on the planet (Peacock 2011, Scott 2009). The unlikely event of
full urban development in these remote Alaskan locations makes hunting pressure one of
the top anthropogenic forces affecting bear populations (Milner et al. 2007, Harris et al.
2002).
Bear harvest levels vary across game management units (GMU) throughout
Alaska due to available habitat, human access, jurisdictional regulations, geography, and
species distribution. Southeast Alaska, primarily the Alexander Archipelago and the
adjacent mainland, contains some of the densest populations of bears allowing for a
larger historical harvest rate. For example, in GMU 1C, a unit comprised of 1,700 square
miles in Southeast Alaska (Figure 1.1), and an area of abundant high quality bear habitat,
approximately 68 black bears and seven brown bears were harvested in 2011. In
comparison, GMU 20A, located in the interior of Alaska near Fairbanks and spanning
down to Denali National Park, an area with abundant habitat but a lesser abundance of
high quality available food such as salmon, approximately 28 black bears and zero brown
bears were harvested the same year (www.secure.wildlife.alaska.gov). The large variation
between populations throughout Alaska creates a need for area-specific population
research to determine the sustainability of ADF&G’s harvest rate region to region.
6
Bears have been regularly hunted for sport and subsistence in Southeast Alaska
for many generations (Barten 2004). Within the past century, attitudes towards bears
throughout the United States and Canada have shifted where the conservation of bear
populations outweighs the ideology that bears are a nuisance animal that must be
eradicated to promote development and human safety (Miller 1989). While brown bear
conservation continues to be an issue of debate in the lower 48 because of the greater
likelihood of negative bear-human interactions due simply to larger human populations,
brown bears populations in Alaska are considered healthy and, in most areas, not
considered to be in danger of critical population depletion. As a result, harvest
regulations have evolved in an attempt to allow for subsistence and sport hunting at
sustainable levels.
Up until the early part of the 20th century, bears were hunted in Alaska without
regulation. The first bag limit set on coastal brown bears occurred in 1902 (Miller 1989)
offering them a small amount of protection. Hunting brown bears continued at
unsustainable levels until 1925, when a new game law was passed that eliminated market
hunting of big game (Thornton 1992). The historical viewpoint towards brown bear
harvest management well into the early 20th century was that bears were a “nuisance” or
hindrance to development in Alaska. Thomas Riggs, Alaska Governor from 1915-1921,
spoke freely about the impediment of the brown bear on development and actively sought
to extirpate the species from the state (Sherwood 1979).
Throughout the decades to follow, with the development of the Alaska
Department of Fish & Game, as well as the Alaska Board of Game, bag limits were
reduced and hunting seasons were introduced. The development of quantitative science
7
and rigorous wildlife ecology research enabled wildlife managers and citizens alike to
demand more responsible game management and wildlife conservation from public and
state land managers. Beginning in 1973, it was required that harvested black bears be
sealed, a process in which a sealing officer places a locking seal on the skull and hide
until the skull is measured, a tooth is extracted, and date, location and cause of mortality
is recorded. This enables game managers to record trends on harvested animals in
designated game management units (GMU). Sealing continues to occur with harvested
animals to this day and is the primary contributor to the working knowledge of harvested
bears and their population trends (Barten 2004). Game management in Alaska continues
to develop in an attempt to maintain sustainable harvest levels. In the case of bear
management in Southeast Alaska, more precise population estimates would allow for a
more accurate prediction of population trends and allow wildlife managers to implement
regulation accordingly.
Bear Harvest Trends in GMU 01C
Since 1973, when sealing became a requirement for harvested bears, trends have
shown a steady increase in harvested black bears with a mean of 47 in the 1970’s, 73 in
the 80’s, and 96 in the 90’s (Barten 2004). Brown bears are hunted less often in GMU
01C with averages nearing 10-20 bears annually. This is in part due to lower densities of
brown bears and the availability of nearby hunting opportunities in GMU 04 which
contains some of the largest populations of brown bears in Southeast Alaska (Scott 2009).
Black bear harvest has continued to increase through the 2000’s but the upward trend has
slowed since the beginning of the decade (Figure 1.2).
8
Brown bears have just recently begun showing up in small numbers within the
Gustavus forelands since their extirpation in the early 1900’s. 2010 was the first year in
decades where multiple brown bear sightings occurred within the town and sightings
continued to increase annually (Lewis 2012). It is possible that brown bear populations
will continue to rise as they inhabit areas of the Gustavus forelands. There is high quality
brown bear habitat with abundant food resources available, such as miles of shoreline
containing grasses and sedges, an essential spring and early summer food, as well as
multiple salmon bearing streams and rivers. These resources may allow brown bear
populations to increase during the next few years.
Within the town of Gustavus and bordering state lands, human caused mortality
of bears can be divided into three categories: harvest, defense of life and property (DLP),
and road kill. The majority of human caused bear mortality in the Gustavus area is
classified as DLP. In the case of a DLP kill, nuisance bears are shot and the hides of the
animals are forfeited to ADF&G. As a result, licensed hunters frequently use their tags on
DLP kills in order to keep the hides (Lewis 2009). These occurrences must be considered
when determining the primary cause of bear mortality within the Gustavus area. DLP
kills are common in the area but records show that they only account for a small
percentage of the total annual mortality (Figure 1.3). Unfortunately, it is impossible to
determine the number of instances this occurs annually because there is no method of
reporting in place that can adequately represent these occurrences. 80% of hunting in
GMU 01C for both black occurs in the spring months when food availability consists of
vegetation growing near the intertidal zone allowing easier access by boat and visual
identification by hunters (Barten 2008).
9
The availability of hunting opportunities in the Gustavus forelands has expanded,
coinciding with the establishment of ferry access beginning in 2011, the recent transfer of
ownership of Falls Creek, and the increases in the local brown bear population. It is now
more important than ever to keep detailed records of harvest trends and begin more indepth research on bear populations. With rising annual harvest rates and lack of empirical
population data, it is unknown whether current human-caused mortality of bears in
Gustavus is at a sustainable level. Brown bears encroaching on previously black bear
dominated areas could also have a dramatic impact on species distribution and total black
bear numbers. The current density of black bears in Southeast Alaska is estimated by
extrapolating from a bear population study conducted by Poelker and Hartwell (1973)
from western Washington State. In addition, skull seals and anecdotal information from
local residents and hunters contribute to the total working knowledge of bear populations
in the area. Noninvasive genetic sampling of bears in the area will increase this
knowledge allowing for a more precise population estimate that will enable wildlife
managers to predict current and future impacts of human-caused mortality.
10
Figure 1.1. GMU 01C-01D.
Game management unit 01C - an area comprising over 1,700 square miles of high
quality bear habitat and containing Federal (Glacier Bay National Park), State, and
private lands (ADF&G n.a.).
11
Figure 1.2. Numbers of black and brown bears harvested by year in GMU 01C, 19962011 (Neil Barten, Ryan Scott ADF&G).
Figure 1.3. Number of bears killed by year in Gustavus by category: road kill, defense of
life or property (DLP), or harvested/hunted, 1996-2011 (Neil Barten, Ryan Scott
ADF&G).
12
History of Noninvasive Genetic Methodology of Bear Species
Recent advances in noninvasive genetic tagging have enabled wildlife managers
around the world to obtain more precise estimates of bear populations (Kendall et al.
1999, Paetkau 2004, Boulanger 2002, Boulanger et al. 2004, Wegan et al. 2012, Waits
and Leberg 2000, Taberlet et al. 2009, Stetz et al. 2010, Sawaya et al. 2012, Pool et al.
2001). DNA analyses of animal hair dates back to the early 1990’s (Morin and Woodruff
1992), but these techniques were not used for bear population estimates until the late
1990’s (Woods et al. 1999, Poole et al. 2001, Boulanger et al. 2002, Paetkau 2003,
Boersen et al. 2003). Prior to the use of noninvasive genetic tagging, wildlife managers
frequently used methods involving live capture and collaring techniques. While effective
in determining movement and population size, these techniques proved to be costly, and
time consuming. Moreover, in regions characterized by dense forests, other techniques
used to estimate populations, such as aerial surveys, are not feasible and pose the risk of
identification error. The challenges of estimating wildlife populations with large home
ranges in remote locations continue to impede wildlife managers’ ability to obtain
accurate abundance and density estimations and thus the search for more robust
techniques has continued to develop.
Noninvasive sampling methods
One common form of sampling hair from bears for genetic analysis comes from
the use of bear rub trees. Bear rubs are naturally occurring and can be easily identified by
claw and bite marks found on the bark of trees along bear and human use trails. Both
black and brown bears rub on trees throughout the summer season but studies have
shown bear rubs peak during the months of May and June (Green and Madson 2003).
13
This behavior is thought to occur at higher rates during the molting and breeding season
though rubbing can continue through the summer and fall until hibernation. To increase
sample quality, two foot-long strands of barbed wire can be installed on a rub tree. Bear
rubs are a repeatable data source frequented by ursine over an extended time frame and a
single tree is commonly used by more than one bear (Green and Mattson 2003). The sole
use of bear rubs for mark-recapture analysis is not recommended due to the bias caused
by unequal frequency of use between differing sexes and species (Sawaya 2012).
Another method developed for sampling bear hair deploying hair traps. This
method was developed within the past 15 years and greatly increases sample size and
reduces capture heterogeneity. Hair traps are barbed wire corrals surrounding scent lure,
usually a mixture of rotten cow’s blood, emulsified fish oil, and glycerin. The hair traps
are set up by encircling a group of trees using 30 meters of barbed wire at approximately
50 cm above the ground. When bears investigate the scent lure in the center of the trap,
their hair is snagged on the barbed wire. These samples are then analyzed by a genetics
lab where specific loci extracted from the roots of the hair are then amplified using
polymerase chain reaction (PCR). By using multiple methods of obtaining black and
brown bear hair samples, there is a reduction in the heterogeneity of capture probability
(Boulanger et al. 2008). This implies that using data sources such as rub trees and hair
traps limits bias based on variables such as species, sex, or age by allowing the
probability of capture to be equal.
14
Study Design Techniques
Woods et al. (1999) was the first study to use noninvasive genetic sampling on a
large scale for estimating black and grizzly bear genetic variability in the Columbia River
basin of British Columbia, Canada. In this study they researched new methodology for
obtaining genetic samples from free ranging bears using three separate field trials. These
trials show baited barbed wire enclosures (i.e. hair traps) are the most successful in
obtaining high numbers of samples from a diverse population of bears in the study area.
Following this study the methodology behind the genetic tagging of black and grizzly
bears has undergone many alterations in an attempt to alleviate biases associated with
capture probability. For example, Sawaya et al. (2012) showed high detection of female
grizzlies and male and female black bears, but a low detection of male grizzlies using
only hair traps. Conversely, bear rub trees had a higher detection of both male and female
grizzlies as opposed to black bears. Therefore, a combination of both barbed wire
enclosures and barbed wire installations has proven to reduce capture heterogeneity and
limit gender bias (Boulanger et al. 2008, Sawaya et al. 2012). This is a single example of
the evolution of genetic tagging and in the past decade there have been a growing number
of articles published addressing similar issues related to the refining of field and DNA
extraction techniques (Paetkau 2003, Boulanger et al. 2002, Boulanger et al. 2008,
Sawaya et al. 2012).
Research design for optimal capture probability has also recently undergone much
scrutiny with researchers attempting to balance cost with sampling distribution and
intensity (Stetz 2008). Most studies involving the noninvasive genetic sampling of bears
use a grid overlay system where each grid across the study area represents the minimum
15
home range of a female black bear (Otis et al. 1978). Within each cell a single hair trap is
deployed and checked at a frequency designated by the study design. Wegan et al. (2012)
found that to avoid biased population estimates, and to increase the cost effectiveness of a
project, the sampling of bear hair should occur between late spring and early summer
while the hair samples are of higher quality, coinciding with annual molt. To increase
capture probability and avoid closure violations, simple random sampling is not required
for collecting mark-recapture data (Williams et al. 2002). More commonly a stratified
random approach is used by installing traps in high quality bear habitat within one of the
randomly selected cells. The frequency of the sampling sessions as determined by Woods
et al. (1999) is typically 10-14 days.
Using baited hair traps in combination with barbed wire installations and naturally
occurring rub trees, biologists can collect fur samples from individual barbs. DNA
extracted from the roots of 3-5 guard hairs per sample are analyzed by identifying a suite
of a minimum of six microsatellite markers. This number was determined by Paetkau
(2003) when analyzing proper methods to reduce genotyping errors in mammalian hair
analysis. The individual markers given to individual animals through this process can
then be used in mark-recapture analysis to determine abundance, density, and total
population size respectively. This method was developed within the past two decades in
previous studies of brown and black bear populations in Canada, Alaska, and the North
Continental Divide (Kendall et al. 2009, Wilder 2003, Woods et al. 1999, Boulanger
2002). There are many factors that must be considered when conducting this research in
order to provide accurate results. For example, rain and sun can degrade the hair sample
rendering it useless for DNA extraction. Another consideration is to avoid cross
16
contamination between samples. If two animals are detected within a single sample, it
will be unusable (Long and Zielinski 2008). These factors have been studied throughout
the past 20 years making this methodology one of the most widely used in field research
surrounding the population estimates of rare and elusive species (Woods et al. 1999,
Boulanger 2002, Boulanger et al. 2004, Kendall et al. 2009).
Open and Closed Population Modeling for Mark-Recapture Analysis
Understanding animal population size and trends has been important to biologists
for centuries in order to understand trends and make informed management decisions.
One way researchers have obtained these population estimates is through mark-recapture
methods. This process involves the capture, marking, and releasing of a portion of a
population. Later another portion of the population is captured and the previously marked
animals are counted. In theory, the number of marked animals captured during the second
capturing occasion should be proportional to the number of marked animals in the whole
population (Amstrup et al. 2005). Methodology surrounding mark-recapture analysis has
continued to develop and our understanding of wildlife populations now includes detailed
analysis of population structure of many species in a wide variety of environmental
conditions. Until recently, the use of mark-recapture methods on rare and elusive species
has been limited due to the inability of gathering information from those species, or our
inability to reliably “mark” them. Modern field methods described above allow
researchers to successfully mark and recapture animals efficiently and reliably. Markrecapture methods have been used to estimate population sizes by taking the total number
of animals captured during two or more sampling sessions. Mark-recapture using
noninvasive methods uses this same technique but also includes the associated
17
probability of detecting the individual animals (Long and Zielinski 2008). In these
studies, the animal is not physically marked but rather has a specific genetic mark, or id,
associated with each individual. Boulanger et al. (2002) conducted one of the first studies
investigating the efficacy of noninvasive genetic sampling using bear rub trees and baited
hair traps to determine a population estimate of grizzly bears in British Colombia and the
United States. In his analysis, he describes three principle concerns when applying markrecapture methods to black and grizzly bears:
1. Widespread movement of bears can violate the assumption of population closure
therefore positively biasing the given population estimate.
2. Capture probability can be biased based on sex and age classes of bears.
3. Obtaining adequate sample sizes can be challenging given typical bear densities.
A closed population is one where it is assumed that the total number of
individuals within a study area is unchanging. This means there is no emigration,
immigration, deaths, or births (Amstrup et al. 2005). Recent models demonstrating markrecapture analysis rely heavily on this assumption as well as the assumptions that the
animals will not lose their tags, that all tags are recorded correctly, and that the animals
act independently. Any violation of these assumptions can lead to biased results, typically
a positive bias creating a population estimate that is larger than the ‘true’ population.
Poole et al. (2001) accounted for this inflated population estimate of grizzly bears in
northeastern British Columbia, Canada. Male grizzly bears have a larger home range than
females (Boulanger and McLellan 2001) and therefore, by reducing the total estimated
population by a percent proportional to the estimated number of male grizzly bears, Poole
and colleagues were able to reduce the bias created from closure violation. Closure
18
violation usually results in an inflated estimate because animals move in and out of the
mark-recapture grid and are therefore counted as part of the total population (Boulanger
and Mclellan 2001). Because the natural world does not function by standards such as
these, and having a truly closed population is uncommon, it is difficult to find a species
or study area that does not violate closure on some level. By limiting the duration of the
study scales (both temporally and spatially), researchers can limit the effects of
emigration, immigration, births, and death on population estimation (Boursen et al. 2003,
Bittner et al. 2002).
In recent decades open population models have become more accurate and have
been used more frequently by researchers in the field. Open population models allow for
birth, death, emigration, and immigration but require that the subjects have equal
probability of capture. The Jolly-Seber (Jolly 1965, Seber 1965), and other additions to
this model, are the most commonly used models when investigating open populations.
There are few time restrictions with this model as long as intervals between trapping
sessions remain constant. For example, each session must occur once every day, week,
month, or year, without adding or subtracting sessions during the study. After occasion 1,
unmarked and marked animals that are caught are recorded and the unmarked animals are
then marked and released back into the population (White 1998). This process continues
for the entire sampling period. Once completed, analyses using the Jolly-Seber model can
help determine minimum population, apparent survival, and capture probability. It is
important to account for capture heterogeneity, or probability of capture between sex,
age, and other demographic classifications, in order to avoid inflated population results as
discussed above with closure violations.
19
Concerns with Genetic Sampling
DNA sampling techniques are based on Paetkau’s (2003) research regarding the
use of microsatellite molecular markers to obtain multilocus genotypes of individual
animals. These individual markers are then used in mark-recapture analysis to determine
abundance, density, distribution, genetic relatedness, and overall population. The
noninvasive nature of this type of study offers many advantages to conventional
population analysis involving “hands on” techniques but care must be taken during DNA
extraction and analysis because of genotyping errors. These errors commonly occur when
there is low DNA quantity or quality, or poor extract quality within a single sample
(Taberlet et al. 2009). These errors are known as allelic dropout, or a misidentification of
alleles because only one allele of a heterozygous individual is detected or because of
extreme DNA degradation of the single sample (Taberlet and Luikart 1999). Allelic
dropout is one of the more common types of genotyping error associated with
noninvasive genetic sampling but other considerations must also be made when analyzing
genetic material.
During the process of DNA extraction, a number of individual loci are extracted
and analyzed giving a unique “mark” to each sample. It is possible for individuals to have
identical genotypes at a limited number of loci examined (Mills et al. 2000). This error is
known as a “shadow effect” and is most commonly seen within closely related
populations and populations with little genetic variation (Mills et al. 2000). The problem
most commonly associated with the shadow effect is the failure to identify certain
individuals in a population, resulting in a negatively biased population estimate and a
positively biased survival rate. The single solution to this problem is to simply analyze a
20
greater number of loci per sample. Waits and Leberg (2000) discuss the potential
drawbacks of adding loci stating the genotyping error rate increases with increased loci
per sample which has the opposite effect as the shadow effect, creating an inflated
population estimate. The balance between number of loci analyzed and the probability of
increased genotyping error has been well researched during the past decade and improved
methodology continues to decrease errors associated with genetic tagging.
Conclusion
Black and brown bears in Southeast Alaska have a large cultural, economic, and
biological impact on the areas where they reside. Throughout the 20th century, and
progressing into the 21st century, research on bear biology and behavior in Southeast
Alaska has been slowly increasing. Brown bears have been widely studied in specific
regions mainly due to extremely high densities and opportunities for sport hunting. There
has been very little research conducted on black bear populations throughout the region
and until recently all population estimates were based on studies done in Washington
State in the 1960’s and 1970’s. Alaska Department of Fish and Game collect data from
harvested bears as well as anecdotal information from recreationists and hunters to
inform them of current population trends of both black and brown bears. Recent harvest
records have shown a gradual increase in harvest of both black and brown bears
throughout the Southeast region of Alaska. In order to assure sustainable harvest rates,
accurate population counts would prove invaluable for wildlife managers. Recent
advances in noninvasive genetic tagging have allowed biologists to acquire accurate
population counts of ursine species throughout the North American continent using markrecapture modeling. Research methodology has evolved to account for errors associated
21
with field techniques, genetic extraction, mark-recapture statistical modeling, and capture
heterogeneity based on species, gender, and age of the bears.
Bears located in Gustavus, Alaska, bordering Glacier Bay National Park,
frequently cross between state, federal, and privately owned land. There has never been a
population study done on either black or brown bears in Gustavus and its’ surrounding
areas. Bears are important for wildlife viewing and sport hunting opportunities that draw
people from all over the world. Federal and state wildlife managers are concerned about
the impacts of potential annual bear harvest increases on bear populations in Gustavus;
and without a baseline population estimate it is impossible to know at what level to set
the maximum sustainable harvest rate. It is possible that harvest rates are either set too
high, where the harvest is greater than the reproduction rate; or too low, where there is
the potential for increased harvest without danger of depleting the current population.
Peacock’s (2011) research on black bear populations on Kuiu Island, Alaska, shows that
current population estimates based off of harvest information and anecdotal reports from
recreationalists can be misleading by having an inaccurate population estimate. This
misinformation could have a large impact on bear populations. With the recent
advancement in population monitoring technology, researchers and wildlife managers
should continue to pursue information on bear populations throughout Southeast Alaska
to reduce the negative impacts of human-caused mortality. The development of more
robust population estimates will contribute and inform multiple agencies and
organizations decisions about how to implement sustainable bear harvest and
management. The continued refinement of these methods encourages adaptive
22
management practices within the agencies that will promote healthy populations of bears
and responsible use of one of southeast Alaska’s most valuable natural icons.
23
II: BLACK BEARS IN GUSTAVUS, ALASKA: POPULATION AND
HARVEST MANAGEMENT.
Estimating wildlife population size and understanding population trends are
central concerns for species management. Noninvasive sampling techniques have been
developed over the past two decades allowing wildlife managers the ability to obtain data
on elusive wildlife species’ population size, distribution, and genetic variance, within and
among populations (Woods et al. 1999, Boulanger et al. 2008, Boulanger et al. 2002,
Mowat and Strobeck 2002, Kendal et al. 2009). These noninvasive methods have been
increasingly used by researchers to study the population trends and assess the effects of
human-caused mortality on certain animal species. Obtaining quantitative data to inform
the impacts of humans on game and nongame animals, as well as practicing adaptive
management as a systematic approach to sustainably managing populations is vital for the
conservation of wildlife species. Even in large, remote areas of North America, excessive
animal harvest can have large impacts on the game animals’ population, and, as a result,
impact the larger ecosystem as a whole (Levi et al. 2012).
Within the past 15 years, a large amount of research has been conducted on black
(Ursus americanus) and brown (Ursus arctos) bears throughout North America providing
information on previously unknown populations and their management implication
(Bittner et al. 2002, Boersen et al. 2003, Boulanger et al. 2002, Kendall et al. 2008, Poole
et al. 2001). These studies have shown the need for more precise regional population
estimates of black bears in Alaska. Black and brown bears in Southeast Alaska are
commonly hunted for subsistence and sport purposes and are also valued for viewing
opportunities. While brown bear populations in the region have been widely studied, until
24
recently (Peacock 2011), no population studies existed for black bears in Southeast
Alaska. Current population estimates of black bears are extrapolated based on results
from research completed in Washington State in the 1960’s and 1970’s (Poelker and
Hartwell 1973). Studying population dynamics of black bears in designated game
management units, where harvest levels varies, can increase the understanding of impacts
of human-caused mortality on bear species by providing insight on population trends.
This knowledge will contribute to adaptive management and conservation strategies and
actions proposed by resource agencies, conservation organizations, and other
stakeholders.
The Gustavus forelands, a 200 km² area of land comprising federal, state, and
private lands, are an example where bear populations move between National Park
Service land where they are protected and state and privately owned land where an
annual harvest limit of 10% the total population is in place (Figure 2.1) (ADF&G 2011a).
During a 2010 Region I board meeting, members of ADF&G discussed regional
estimates of black bear populations and the potential for maximum black bear harvest.
Using the mean number of bears harvested per year and per area from 2007 – 2009 to
assess the harvestable surplus, the board members decided on 10%. Because the black
bear population of the Gustavus forelands is currently unknown, it is impossible for
managers to know for certain whether the harvest objective of 10% has been met. For
bear management decisions to be effective, it is essential that a cooperative process exists
between both the National Park Service (NPS) and Alaska Department of Fish and Game.
Both agencies value the animals and strive to promote a healthy and sustainable bear
population, even if each agency has different management objectives. Both State and
25
Federal game managers are concerned about the potential increase in bear harvest due to
increased access to the community via the Alaska Ferry, which began service to Gustavus
in 2010, and the recent transfer of land ownership from federal National Park land to
State land for the Falls Creek hydroelectric project. With this potential increase in access
to bear harvest, it is essential that wildlife managers begin monitoring black bear
populations within the area.
Non-invasive mark-recapture methods using barbed wire installations on rub trees
and scented hair traps are relatively inexpensive and accurate ways to determine
population estimates and densities of bears in comparison to live capture and collaring
techniques (Stetz et al. 2010, Kendall et al. 2008). Rub trees are a repeatable data source
frequented by black and brown bears over an extended time frame (Green and Mattson
2003). Hair traps add rigor to the study by increasing sample size and reducing bias based
on gender and species (Sawaya 2012). Hair collected from sample trees and hair traps can
be analyzed to determine individual identification, sex (Taberlet, 1993), species, genetic
population structure, and individual genealogies (Woods et al. 1999). The genetic data
collected from the rub trees and hair traps can be used to assess population trends and the
impacts of certain management decisions such as allowable annual harvest. Kendall
(2008) discusses the value of estimating population trends as they correlate to landscape
changes, and how understanding these changes will allow managers to “reverse negative,
or enhance positive changes in the population.” Providing a baseline population structure
of black bears will allow future researchers comparable data that could inform shifts in
species distribution, population shifts, and human caused impacts on both species. The
objectives of this research are to 1) determine a population estimate of black bears within
26
the Gustavus forelands, 2) establish a monitoring protocol for evaluating population
trends and future research, 3) explore sources of human-caused bear mortality and
harvest trends, and 4) explore possible interagency bear management actions to ensure
sustainable bear harvest levels.
Study Area
Glacier Bay National Park is comprised of 3.3 million acres of protected federal
land and is adjacent to four other protected lands making it one of the largest protected
areas in the world at over 25 million acres (NPS 2013). The 200 km² study area of this
project includes most of the Gustavus forelands located in Southeast Alaska at the
southern border of Glacier Bay National Park approximately 40 miles west of Juneau,
Alaska (Figure 2.1). This small triangle of land is surrounded by the waters of Icy
Straight to the south, the Sitakaday Narrows to the north and west, and the alpine and
sub-alpine peaks of Excursion Ridge to the east. The forelands are jurisdictionally
divided east to west by state and private land, and National Park Service land. The town
of Gustavus, located in the southern section of the study area has approximately 450
year-round residents, but this population increases dramatically during the summer
season due to the high amount of seasonal employment (U.S. Census Bureau 2010).
Gustavus is the gateway to Glacier Bay National Park and is the only developed area
adjacent to the park and the surrounding lands and therefore is the main jumping off point
for hunting and wildlife viewing opportunities. All access to Gustavus is by boat or
airplane, as there are no roads that connect to other towns, only 10 miles of paved road
connecting the town to the National Park. Tourism is a large industry due to the
proximity of Gustavus to Glacier Bay National Park. Most tourism to the area is by cruise
27
ship or tour boat where the visitors do not set foot in Gustavus, but still thousands of
people visit the town itself for recreation, hunting, and sightseeing (NPS 2012).
The Gustavus forelands are the largest flat plain in all of Southeast Alaska
(ADF&G 2013) with a primarily sandy substrate and a spruce dominated forest. The vast
forested area includes multiple salmon-bearing streams and rivers providing excellent
habitat for robust bear populations and other species such as moose (Alces alces) and
wolf (Canis lupus) (White et al. 2006).
METHODS
During the months of July – October of 2011 and May – September 2012, I
collected hair from 25 opportunistically found rub trees within the study area. Scented
hair traps were deployed from May - September 2012 only (Table 2.1). Scented hair trap
stations were designed using protocols from Woods et al. (1999). Sawaya et al. (2012)
found high detection of female grizzlies and male and female black bears, but a low
detection of male grizzlies using only barbed wire corrals. Conversely, bear rub trees had
a higher detection of both male and female grizzlies as opposed to black bears. I used a
combination of both barbed wire enclosures (scented hair traps) and barbed wire
installations (rub trees) in an attempt to reduce capture heterogeneity and limit gender
bias.
DNA-based marking for capture-mark-recapture (CMR) studies are beneficial
because marks given to an animal from a DNA sample cannot be lost like physical
markers such as ear tags or collars. Sampling design to ensure limited bias and maximum
capture probability was modeled after Boulanger et al. (2004) and Woods et al. (1999).
The DNA sampling technique uses microsatellite molecular markers that provide genetic
28
structure and unique marks for individual animals. The results of the genetic analysis
were organized and analyzed using program DENSITY to estimate a minimum
population of black bears within the study area.
Field Techniques
Bear Rub Trees
Rubbing is a naturally occurring behavior among both male and female bears
(Sawaya et al. 2012). Both black and brown bears rub on trees throughout the summer
season but studies have shown bear rubs peak during the months of May and June (Green
and Madson 2003). This behavior is thought to occur at higher rates during the molting
and breeding season though rubbing can continue through the summer and fall until
hibernation. Bear rub trees are found at different densities and can be located by
following game or human use trails. They are identifiable by the scratch and bite marks in
the bark inflicted by the bear. It is common for a rub tree to be used by multiple bears
throughout the season and some trees may be used for many generations (Green and
Madson 2003).
I located 25 rub trees within the study area. Multiple rub trees were commonly
found within one square mile of one another and therefore, to maximize the likelihood of
capture and reduce the chance of sampling redundancy, I chose only 3-4 trees in a given
area. I installed two, 1.5 ft. strands of barbed wire at approximately three feet and five
feet on the tree at a diagonal angle. This method helps ensure capture of younger and
smaller bears by putting the strands of barbed wire lower on the tree where they are more
likely to come in contact with it.
29
Each tree was checked on a 12-14 day interval from July - October 2011 and May
- September 2012 (Table 2.1). Each 12-14 day interval is considered a single sampling
session. There were a total of 7 sampling session in 2011 and 10 in 2012. This frequency
of return decreases errors associated with genotyping by ensuring hairs in a single sample
are from one individual (Taberlet 2009). Each clump of hair was treated as an individual
sample. Hair samples were collected using tweezers and were placed in a small paper
coin envelope and given a unique identifier including location of the tree, hair sample
number, and the date. If the hair samples were wet or moist they were dried at low heat
using a food dehydrator and then placed directly into a small paper envelope and plastic
bag containing silica desiccant. A cigarette lighter was used to burn off any remaining
hair on the barbs after the samples had been collected to avoid sample contamination
from future rubs.
Scented Hair Traps
I deployed 8 hair traps at one time within the study area. The locations of the traps
were determined using a grid system overlay of the study area where each cell within the
grid represented the minimum home range of a female black bear (Otis et al. 1978). This
grid design balances effort across the area and reduces capture variation (White et al.
1982). All bears must have an equal opportunity of capture in order to reduce the
occasion of underestimating the population size (Pollock et al. 1990). I subjectively chose
sites within the cells for trap placement to maximize capture probability (Woods et al.
1999). Traps were deployed in areas with bear sign and high quality bear habitat. Because
there is no information on minimum home ranges of female black bears within the study
area, I used the ranges of black bears documented by the Juneau region Alaska
30
Department of Fish and Game (ADF&G). Minimum female black bear home ranges in
the region span from 10-25km and therefore each cell within the study area is
approximately 4km x 4km. There were eight cells within the study area. It is suggested
by White et al. (1982) and Woods et al. (1999) that there be a trap set in each cell per
sampling session.
Hair trap design and location was modeled after Woods et al. (1999). Samples
were collected from May - mid-September (Table 2.1). Hair traps were approximately 30
meters in circumference with a single strand of barbed wired around several trees 5075cm above the ground running the circumference of the unit. The center of the trap was
baited with a liquid scent deposited on a pile of wood. This scent was a two to one ratio
of rotted cow’s blood and fish oil mixed with 150ml of glycerin. This method ensured
there is no food reward for the bear that enters the trap and also bait will not have to be
replaced between sampling occasions. I changed the scent throughout the season by
subtracting the fish oil and applying skunk essence oil to nearby trees to reduce the
chance of trap response by the bears. All depressions in the earth below the barbed wire
corral were filled with debris to ensure the bears could not get to the bait without contact
with the barbed wire. Any individual hair clump found was given an identifier associated
with the trap number and considered a single and unique sample. All traps were set at a
minimum of 500 meters from human development including trails, home sites and
campgrounds. There was signage near any hair trap alerting any passerby of its presence.
Because of restrictions of personnel and time restraints traps remained in the same
spot throughout the study. As with the rub trees, the hair traps were checked every 10-14
31
days. These intervals represent a single trapping session. Throughout the study there were
7 trapping sessions in 2011 and 10 sessions in 2012.
Table 2.1 Number and dates of sampling sessions for rub trees and scented hair traps for
2011 and 2012. Each sampling session was 10-14 days in length where all rub trees and
hair traps were checked one time per session. Session 1 and 8 were also representative of
installation dates of rub trees and hair traps.
2011(Rub trees only)
Sampling
Dates within Sampling
Session #
Session
1
6/28/2011 - 7/15/2011
2
7/16/2011 - 7/30/2011
3
8/1/2011 - 8/15/2011
4
8/16/2011 - 8/30/2011
5
9/1/2011 - 9/15/2011
6
9/15/2011 - 9/30/2011
7
10/1/2011 - 10/15/2011
Sampling
Session #
8
9
10
11
12
13
14
15
16
17
2012
Dates Within Sampling
Session
5/1/2012 - 5/15/2012
5/16/2012 -5/30/2012
6/1/2012 - 6/15/2012
6/16/2012 -6/30/2012
7/1/2012 - 7/15/2012
7/16/2012 - 7/30/2012
8/1/2012 - 8/15/2012
8/16/2012 - 8/30/2012
9/1/2012 - 9/15/2012
9/15/2012 - 9/30/2012
32
Figure 2.1. Study Area - Gustavus forelands with 4 km² grid overlay. Total area within
the study area is 200 km². Source: ESRI software, ArcGIS.
33
Figure 2.2. Rub Tree and Trap Locations - 8 Scented hair traps were deployed May,
2012. One hair trap was deployed within each 4 km² cell. 25 rub trees were equipped
with barbed wire throughout the study area during June and July of 2011.
Methods of Analysis
Genetic Analysis
Analysis of hair was conducted at Wildlife Genetics International, British
Columbia, Canada. DNA was extracted using QUIAGEN’s DNeasy Blood and Tissue
kits, following manufacturer’s instructions for use (Paetkau 2012). The roots of 10 guard
hairs per sample were analyzed by identifying the suite of seven microsatellite markers
34
G10B, G1D, G10J, G10M, MU50, MU59, and G10U. Typically, a suite of six
microsatellite markers are used, but the seventh, G10U, is to compensate for the
relatively low variability in marker MU50 in black bears within the study area. This
number was determined by Paetkau (2003) when analyzing proper methods to reduce
genotyping errors in mammalian hair analysis. This method has been used in previous
studies of brown and black bear populations in Canada, Alaska, and the North
Continental Divide (Kendall et al. 2009, Wilder 2003, Woods et al. 1999, Boulanger
2002). Once individuals were identified, any samples containing more than two alleles at
a locus were assumed to contain DNA from one or more individuals and were thus
discarded from the data to avoid population calculation errors.
Statistical Analysis
Capture histories for each unique individual genotype produced from the genetic
analysis were created for closed population mark-recapture analysis in program
DENSITY (Efford 2012). This model assumes there is geographic closure, where there is
no emigration or immigration on or off the study area, demographic closure, where there
are no births or deaths within the time frame of the study, no marks lost, and that every
animal has equal capture probability. Analysis of 2011 and 2012 capture histories were
run both separately and together. A population estimate of black bears within the study
was produced using the more conservative of the two analyses. This is to account for
possible overestimation in population due to violations in geographic closure, where
bears entering from outside the study area could have inflated the estimate. Humancaused deaths were accounted for by removing the individual from the statistical analysis
but including that individual in the final count (Otis et al. 1978).
35
I ran two analyses using the data collected from the field seasons. For the first
analysis, I divided the two sampling occasions between 2011 and 2012, to reduce the
chance of possible closure violation as a result of births and deaths within the population
between sessions 7 and 8 (Oct. 15, 2011 – May 1, 2012). I ran analyses on the 2012 only
because it contained data from both rub trees and hair traps. I used the closed population
Huggins (1989) linear logistical model with no time covariate because the sampling
sessions were constant. This model assumes there was little to no migration to and from
the study site. Geographic closure can be assumed on two of the three sides of the study
area because two sides of the triangular study site are large bodies of water. Natural
mortality of black bears should be negligible within a single sampling year due to their
relatively high survival rate, and births are unlikely to occur multiple times throughout
the season. One black bear captured during this study in was killed by a hunter in 2012.
As advised by Otis et al. (1978), I omitted the capture history from analysis, but added
the number of lost animals (one) to the final estimate.
For the second analysis, I used the data set including both 2011 and 2012 and
used the Huggins Linear Logistic model with time covariate to account for time in
between sampling seasons to estimate population size. Both models assumed no
behavioral effect on capture, meaning once an animal was captured once, the probability
of being captured again would not change.
Harvest Records
I compared the results of the genetic and statistical analyses to ADF&G’s bear
harvest records, looking at trends in human-caused black bear mortality over a 21-year
period from 1990-2011 to determine if the community of Gustavus is a population sink
36
for black bears. A population sink as defined by Pulliam (1988) is where within-habitat
reproduction is insufficient to balance local mortality. I used the population estimate
obtained through the statistical analysis and compared it with the harvest trends from the
past 21 years to determine if harvest rates are currently set at either overly conservative
levels or unsustainable levels. By assessing the average number of black bears harvested
per year I was able to produce a percentage of the total estimated population of bears
taken annually.
Because of the low sample size of bear hair collected from brown bears from the
two field seasons, as well as the historically low rate of brown bear hunting that occurs
within the study area, I limited my analysis to understanding the effects of human-caused
mortality on black bear rather than including both bear species.
RESULTS
Sampling Effort
A total of 196 bear hair samples were collected from rub trees, hair traps, and
opportunistically during the two field seasons. During the months of June - November of
2011, 58 hair samples were collected from rub trees and two samples were collected
opportunistically within one of the study area’s 4km x 4km cells. From May – September
of 2012, 134 total samples were collected. 82 samples were collected from rub trees, 49
from hair traps, and three opportunistically found samples in areas located within the
study area.
Genotyping Success
Of the 196 total samples, 57 (29%) did not contain sufficient DNA material to
assign an individual id. The first batch of samples collected in 2011 were sent in June of
37
2012 and had a lower success rate (36%) than that of the second and third batch (93%),
sent within a month after collection in the field. The low success rate in 2011 is largely
attributed to the rainy conditions when samples were being collected in the fall as well as
the delayed analysis of 7 months which likely contributed to the degradation of the
samples. Of the 139 successful samples, 51 were assigned to 14 individual brown bears,
and 88 were assigned to 33 individual black bears. Four samples lacked suitable material
for DNA extraction, one (0.5%) sample was excluded based on appearance (thought to be
ungulate), 51 (26%) samples failed to provide sufficient DNA for analysis, and one
(0.5%) sample showed evidence of multiple species’ hair. Mean observed heterozygosity
among the seven locus used in individual identification was 0.62 (Table 2.2). The
expected heterozygosity was 0.65 with an average difference of 0.03. DNA analysis of
foot pad samples collected from harvested bears in Gustavus showed that only one bear
was shot and sealed in 2012 that was also marked in this study.
Mark-Recapture Analysis
The Huggins Linear Logistical model for the pooled data set of 2011 and 2012
resulted in a population estimate of 54.5 ± 10.3 (95% CI=41.6 – 84.8) black bears within
the study area (Population size N-hat ± SE (95% CI). The capture probability (P – hat)
was equal to 0.0508. A total of 17 trapping sessions, 47 captures, and 32 individual bears
were documented in this analysis (Table 2.3). This indicates approximately 58% of the
total population of black bears were captured within the two year study.
The analysis using the Huggins Linear Logistical model for the 2012 only data set
resulted in a population estimate of 72.7 ± 27.8 (95% CI=41.5 – 162.7) black bears
within the study area. The capture probability for this analysis was P – Hat=0.0413. A
38
total of 10 trapping sessions, 30 captures, and 25 individual bears were marked in this
analysis (Table 2.3).
Table 2.2. Variability of microsatellite markers used to determine individual identity of
black bears in the Gustavus, Alaska forelands.
Locus
K
N
HE
HO
HE-HO
G10B
4
75
0.617
0.573
0.044
G1D
4
75
0.719
0.707
0.012
G10J
9
75
0.649
0.64
0.009
G10M
5
75
0.722
0.693
0.029
Mu50
4
75
0.411
0.32
0.091
Mu59
7
75
0.667
0.653
0.014
G10U
9
75
0.764
0.733
0.031
0.650
0.617
0.033
Average
K=number of alleles observed per marker
N=number of individual genotypes used in the calculation
HE= Expected Heterozysosity
HO=Observed Heterozygosity
39
Table 2.3. Capture results for trapping sessions 1 – 17 from the data set including 2011
and 2012 and capture results for trapping sessions 1 – 10 from the data set including only
2012. 32 total black bears were captured between the two sampling years of 2011 and
2012 and 25 total bears were captured during the single sampling year of 2012. There
was only one marked bear harvested and sealed by ADF&G throughout the sampling
seasons in 2012.
Mark Recapture Results 2011 & 2012
Mark Recapture Results 2012
Session ia
n(i)b
u(i)c
f(i)d
M(i + 1)e
lossesf
Session
ia
1
2
2
22
2
0
1
0
0
20
0
0
2
0
0
7
2
0
2
3
3
5
3
0
3
3
2
2
4
0
3
8
6
0
9
1
4
4
4
0
8
0
4
6
5
0
14
0
5
4
2
1
10
0
5
3
2
0
16
0
6
2
1
0
11
0
6
1
1
0
17
0
7
0
0
0
11
0
7
1
1
0
18
0
8
0
0
0
11
0
8
1
1
0
19
0
9
3
2
0
13
0
9
3
3
0
22
0
10
9
5
0
18
1
10
4
3
0
25
0
11
6
4
0
22
0
12
3
2
0
24
0
13
2
1
0
25
0
14
1
1
0
26
0
15
1
1
0
27
0
16
3
2
0
29
0
17
4
3
0
32
0
a
Number of trapping session
b
number of animals caught at time i
c
number of animals first caught at time i
d
number of animals caught exactly i times
e
number of marked animals after time i
f
number of animals lost after time i
n(i)b
u(i)c
f(i)d
M(i + 1)e
lossesf
Harvest Data
Between the years of 1990 and 2011 a total of 82 black bears were killed by
humans as documented by ADF&G. The mean annual harvest rate for these years was
3.68. The annual harvest rate and mean harvest rate showed a slight upwards trend
40
(Figure 2.3). There is some uncertainty with regard to the true level of harvest due to
underreporting by local hunters. Of the 81 total harvested bears from these two decades,
11 were DLP, 2 were road kill, and the remaining 68 were legally harvested (Table 2.4).
Figure 2.3. Number of black bears harvested in Gustavus, Alaska by year. The trend line
showed a slight increase in harvest over the course of 21 years with an R² value of 0.032
(ADF&G 2011b).
41
Year
DLP
Roadkill
Legal
Harvest
Total
1990
1991
1992
1993
1994
1995
1996
1997
1998
1999
2000
2001
2002
2003
2004
2005
2006
2007
2008
2009
2010
2011
0
0
1
1
0
0
4
4
0
0
4
4
0
0
0
0
0
1
1
2
0
4
0
0
1
1
0
1
0
0
1
0
1
2
0
0
0
11
total
average 0.52
0
0
0
0
0
0
0
1
0
0
0
0
0
0
0
0
0
0
0
2
0.09
4
1
3
6
4
4
12
2
0
3
2
3
1
1
3
9
68
2.8
8
1
3
7
6
4
13
2
0
4
2
4
3
1
3
9
81
3.68
Table 2.4. Bear harvest rates by category in Gustavus, Alaska 1990-2011. The total
number of bears killed in the last 21 years was 81 with the average of 3.68 bears killed
annually (ADF&G 2011b).
42
DISCUSSION
This study produced the first estimation of black bear populations in Gustavus,
Alaska. The ability to compare data sets using rub trees only in 2011 and to compare rub
trees and hair traps from 2011 and 2012, enabled the calculation of a reliable baseline
population estimate of 54.4 ± 10.3 (95% CI = 41.6 – 84.8). This baseline population
estimate can be used by wildlife managers to assess black bear populations and harvest
management and as a comparison with future data. Using bear hair traps and rub trees
throughout the course of two sampling seasons, I was able to provide an estimate of black
bear populations within the Gustavus forelands using two separate statistical analyses in
program DENSITY. The most conservative statistical models were chosen by
systematically analyzing the closed population models found in program DENSITY. For
the purposes of this research, the objective was to determine if Gustavus is a population
sink for black bears, therefore the most conservative estimates of black bear populations
will likely inform us if further research and more precise estimates are required. As
described by Sawaya et al. (2012), the combination of rub trees and hair traps reduces
capture heterogeneity, increases sample size, and reduces gender bias. Kendall et al.
(2009) found that pooling data sources such as rub trees and hair traps in mark recapture
analysis greatly increases the precision of the estimate as opposed to running multiple
analyses for single sources of hair collection (Boulanger et al. 2008, Kendall et al. 2008).
For this reason I did not test models using rub tree or only hair trap data only.
43
Statistical Analysis and Study Design
The first analysis of data from both 2011 and 2012 using the Huggins Linear
Logistical model with time covariance resulted in a smaller population estimate (54.45 ±
10.3 (95% CI=41.6 – 84.8)) as compared to the second analysis of the 2012 only data
(72.7 ± 27.8 (95% CI=41.5 – 162.7). There were also 60 additional samples from 2011
used in the first analysis
It is important to acknowledge that because of the low capture probability (PHat=0.0508), estimates may be less precise than those with values greater than 0.1-0.25
as reported by Kendall et al. (2009) and 0.19 as reported by Poole et al. (2001). Although
it was assumed there was no behavioral change from the traps, the low capture
probability may be due to a trap response, where the animal either returns to the trap
frequently or avoids it all together. To account for this possibility and avoid further trap
response, I used different scent lures throughout the season. Due to logistical and time
constraints, however, hair traps were not moved between sessions as recommended by
Boulanger et al. (2006). Moving traps between sampling sessions can reduce the trap
response of individual bears, either enticing them to return (trap happy) causing
negatively biased population estimations, or trap avoidance (trap shy), causing a
positively biased population estimation (Nichols et al. 1982). If future mark-recapture
studies are done within this study area, moving trap locations between sessions would be
advised to increase capture probability and sample size. The large variation in the
confidence interval also indicates the need for conservative harvest rates, at least until
further, more precise estimates are made. There is still a lot unknown about the black
44
bear population within this study area, but this study enables future researchers a
conservative baseline population for comparative use.
This study was designed after Woods’ and colleagues’ (1999) and Kendall’s
(2009) research conducted in British Colombia, Canada and northwestern Montana on
black and brown bear populations and their distribution. Most studies using hair snaring
techniques for bears cover a vastly larger area (33,480 km²; Kendall et al. 2009) and
obtain a much larger sample size (33,741 hair samples; Kendall et al. 2009), but this
project was designed to cover a much smaller area with the expectation of acquiring far
fewer samples.
It was assumed throughout this study that the population was closed. As seen by
the results of the closed population modeling containing relatively low capture
probability and high confidence intervals, it is possible there were closure violations. The
models used for analysis account for certain violations such as permanent loss, but open
population models could provide more precise estimations on this population. One
assumption that may be violated with my approach is that the marked individuals must be
as likely to be re-marked as un-marked individuals. To account for this possible violation,
the use of rub trees in addition to the hair traps reduces the amount of capture
heterogeneity in the study. This study was designed to provide a baseline population
estimate of black bears located in the Gustavus forelands, an area of only 200 km².
Historically, evaluating population trends that will contribute to harvest management has
been a difficult task due to imprecise estimates and large coefficient of variation (CV>20;
Boulanger 2002). As mark-recapture analysis gains popularity with wildlife managers,
45
the inefficiencies within the study designs should decrease allowing for more precise and
more effective use of population estimates.
Harvest Rates
The current harvest objective for black bears in the Gustavus forelands is 10% of
the total population annually. Using the mark-recapture results obtained from this study
(total black bear population = 54 ± 10 (95% CI = 41.6 – 84.8) and the average annual
harvest (3.68), the current harvest rate falls within the 10% allowance. Although, it is
important to recognize that throughout the 21 year data set, five of those years had black
bear harvest that exceeded the 10% limit. Years with harvest rates that greatly exceed the
10% limit such as 2002 (13 harvested black bears) are a rare occurrence but can have a
large impact on a small population of bears. If considering the lower confidence interval
of 41.6 bears, the 10% management objective was exceeded 10 of the 21 years. With
harvest trends increasing due to increased access to hunting opportunities by the transfer
of Falls Creek from Federal to State land and the recently added ferry access it is possible
for the harvest rate to quickly exceed the allotted sustainable harvest rate. In 2011, nine
black bears were legally harvested. If the modeling is correct, even at the upper end of the
population estimate, around 85 bears, nine bears harvested annually still exceeds the 10%
management objective. In addition to population estimates, more information on birth
and survival rates would greatly improve wildlife managers’ ability to predict outcomes
on bear populations of annual harvest rates that occasionally exceed the allotted 10%.
MANAGEMENT IMPLICATIONS
Noninvasive mark-recapture methods can be used within game management units
in southeast Alaska to assess population trends as they relate to human-caused mortality.
46
This is the first study to use noninvasive techniques to assess black bear population status
in Gustavus, Alaska. Current harvest rates are set at 10% the total population but until
now, the total population was unknown. My results indicate that harvest rates could be
currently set at unsustainable levels where more than 10%, or 4.2 – 8.5 bears, are killed
annually. Although the method used in this study proved to be immediately useful to
wildlife managers, further research is recommended to produce results with higher
capture probabilities and, thus, more precise population estimates. In addition,
information on survival and birth rate will add to the precision of future studies
concerning negative harvest impacts on black bear populations. The statistical modeling
for the data can be examined further for more precise estimates by running open
population models and exploring the possible closure violations further, but due to
logistical and time constraints my examination of the data is limited to the two models
described above. If funding and logistics allows, an extension of the study area as well as
a devoted field crew could potentially increase sample size, decrease capture
heterogeneity, decrease possible violations in closure by extending the study area and
moving traps between sessions, and thus provide increasingly more precise population
estimates. Black bears are a slow to reproduce, reaching sexual maturity at the average
age of three and only reproducing every two to four years. Because of this low
reproductive rate there could be concern surrounding the depletion of the population over
extended time frames. In addition, human-bear interactions could increase as access and
visitation to Gustavus increases. Increased interactions between humans and bears could
result in higher annual defense of life and property (DLP) counts unless management
both on a Federal and State level are diligent about their outreach to visitors as well as the
47
community about bear safety and awareness. These are some of the factors surrounding
human-caused black bear mortality.
Brown bears were also present throughout the study area and may have a large
impact on black bear populations and distribution over the coming years as their
population increases. During my two year study I captured and identified 14 individual
brown bears. Mark-recapture analysis was not run on the brown bear population because
of small sample size and time constraints. Although, capturing 14 brown bears within the
study area is significant in and of itself. Anecdotally, brown bear presence within the
Gustavus forelands has been limited with a few sightings reported annually. Brown bears
were extirpated from the area during early settlement times but since 2010, more
sightings have been reported, increasing in 2011 and 2012 (NPS unpublished data). This
naturally occurring colonization of brown bears into the Gustavus forelands is an event
that should be looked at and studied closely by wildlife managers for a number of
reasons. Brown and black bears have substantial habitat and diet overlap (Herrero 1972,
Mattson 1988). There is continuing research on the effects of cohabitation by both black
and brown bears within a certain area (Mattson et al. 2005). The Gustavus forelands has
abundant food sources for both black and brown bears but it is unknown whether a larger
brown bear presence will cause diminished black bear presence, or whether brown bear
encroachment will be impeded by the high population of black bears. Continued research
studying populations of both species could be achieved by repeating this study in three to
five years. Natural causes of mortality and displacement of black bears such as disease,
habitat changes, and brown bear encroachment, compiled with human-caused mortality
48
only increases the need for further research to ensure the black bear populations are
sustainably managed.
III. CONLUSIONS AND RECOMMENDATIONS
Black bears are an important natural resource throughout southeast Alaska. Their
cultural and economic contributions to the area have great value both for the people that
reside in the area as well as their biological contributions to the natural lands they inhabit.
Quantitative data on black bear populations in southeast Alaska has been limited to a
single study published as recently as 2012 (Peacock 2011). While it is generally agreed
upon that black bear populations in southeast Alaska are at healthy levels and not in
immediate danger of decline, without empirical baseline population data this is a
dangerous assumption that could lead to poor management decisions concerning
allowable harvest. With harvest rates at 10% allowable take of the total population in
much of southeast Alaska, it is important for wildlife managers to have a firm
understanding of the total estimated population, trends, and distribution of bear species
throughout these game management units (GMU) in order to avoid local depletion of
black bears. Recent advances in noninvasive genetic tagging and the use of markrecapture techniques have allowed managers to acquire more precise estimates of bears
across the North American continent (Boulanger et al. 2002, Kendall et al. 2009, Mowat
and Strobeck 2002, Boersen et al. 2003). Wildlife managers must use an adaptive
management and interdisciplinary approach to successfully manage the population of a
game species. The multivariable environment for which these animals reside shows the
49
importance of focusing not only on the biological variables but also the cultural and
economic values as well.
This research is the first to obtain a population estimate of black bears within the
Gustavus, Alaska forelands using noninvasive genetic tagging and mark-recapture
analysis. By comparing the results with the harvest data given by the Alaska Department
of Fish & Game (ADF&G), I was able to infer that there is the potential for harvest levels
to be set at an unsustainable rate over the course of many years. In addition to the field
research, my investigation into the history of bear harvest both in Alaska and the lower
48 states, where bears have seen dramatic population and habitat declines, only
strengthens the argument that continued research of bear populations in Gustavus, Alaska
is imperative for informing continued bear management. A slow decline of the total
population of black bears by human-caused mortality can be avoided if managers
continue to monitor populations and to collect harvest data from legal, illegal, defense of
life and property (DLP), and road kills. As human encroachment continues and the
interface between bears and humans increases the need for adaptive management and
multiple agency cooperation also increases.
Recommendations
As this is only the first population estimate in this area, and because of logistical
and time constraints in the field, it is highly recommended that the study be repeated in
three to five years to compare data and assess possible trends. While the estimate
produced from my analysis may provide immediate management implications, the value
of this research is more a call for further study using similar techniques. More precise
mark-recapture statistics could be used if the multiple year data sets contained both rub
50
trees and hair traps. As mentioned above, moving traps in between each sampling session
would also increase the precision of the mark-recapture output as well. Analyzing sex
distribution and capture probabilities could add to our knowledge of capture
heterogeneity based on sex. Running mark-recapture analysis separating male and
female, then running analysis with both would give insight into the relatively low capture
probabilities produced from my research.
This study was developed between the needs of multiple agencies including
Alaska Department of Fish & Game and the National Park Service. The results indicate
the need for further management actions which include:
Continued monitoring of harvested black and brown bears.
Consider management action if the average annual harvest of black bears exceeds
4 bears.
Analyze results using open population models.
Repeat study using similar but refined methodology every 5-10 years.
-Move hair traps between sessions.
-Increase study area to include excursion ridge and bear track cove.
-Locate more trees to increase sample size and capture probability.
-If sample size allows, run population estimate on brown bears.
Continue to monitor brown bear activity in the Gustavus forelands.
Interdisciplinary Effort
The interdisciplinary nature of wildlife management involves the application of
scientific, social, technical, mathematic, and economic knowledge. Understanding how
these all work together is imperative when compiling research that may impact future
51
management decisions. The social and economic side of research concerning game
species matters greatly because many times the effort behind the scientific research may
have social and economic implications. Bear harvest regulation, for example, may be
informed by this study resulting in impacts on allowable harvest. These scientific
findings will contribute to the working knowledge of bear species, their distribution, and
their populations.
Black and brown bear populations and harvest management issues will continue
to be an issue of concern in the Gustavus, Alaska forelands. Research on bear species and
game management that reaches across institutional barriers and promotes active
contribution from wildlife managers and biologists of multiple agencies is sure to
improve future management decisions. Understanding the history of harvest and
accounting for the economic and social dimensions of game management allowed me to
gain further insight into the importance of an interdisciplinary approach to conservation
biology and wildlife management.
For many years wildlife managers have sought to understand population dynamics
in order to make informed management decisions that will ensure the health of game and
nongame wildlife species. The continued use of bear species both for consumptive and
non-consumptive purposes add value to the animals and increase the need for rigorous
management. This management must begin with the understanding of the animals’
population trends, sex, and species distribution. Multiple agency cooperation, as well as
shared information between agencies, will allow managers to make informed decisions
about sustainable harvest and maintaining healthy populations. My research has
contributed to the working knowledge of black and brown bears in the Gustavus, Alaska
52
forelands but it is imperative that there be continued work analyzing population trends as
they relate to allowable harvest. Bears can continue to enrich their environment as well as
provide sport and subsistence hunting, and wildlife viewing opportunities for years to
come as long as care is taken by wildlife managers to ensure that reality.
53
LITERATURE CITED
Alaska Department of Fish and Game. 2011a. Board of Game Direction on Black Bear
Guide Allocations and Harvest in Southeast Alaska. Alaska Department of Fish
and Game Division of Wildlife Conservation. Juneau, Alaska.
Alaska Department of Fish and Game. 2011b. ADF&G Bear Harvest Database 19602011.
Alaska Department of Fish and Game. www.secure.wildlife.alaska.gov. Date retrieved:
Sept. 2012.
Amstrup, Steven C., McDonald, Trent L. Manly, Bryan F. J. 2005. Handbook of CaptureRecapture Analysis. Princeton, New Jersey: Princeton University Press.
Barten N. 2008. Unit 1C Black Bear Management Report. Pages 30-47. Black Bear
Management Report of Survey and Inventory Activities 1 July 2004 – 30 June
2007. Alaska Department of Fish & Game. 17.0. Juneau, Alaska.
Barten N. 2004. Unit 1C Black Bear Management Report. Black Bear Management
Report of Survey and Inventory Activities 1 July 2001-30 June 2004. Alaska
Department of Fish and Game. 17.0. Juneau Alaska.
Bittner, S. L., T. L. King and W. F. Harvey. 2002. Estimating population size of
Maryland’s black bears using hair snaring and DNA analysis. Proceedings Annual
Conference of Southeastern Association of Fish and Wildlife Agencies 56: 31221.
Boersen, M.R., Clark, J.D., King, T.L. 2003. Estimating Black Bear Population Density
and Genetic Diversity at Tensas River, Louisiana, Using Microsatellite DNA
Markers. Wildlife Society Bulletin 31:197-207.
Boulanger, John, Katherine C. Kendall, Jeffrey B. Stetz, David A. Roon, Lisette P. Waits,
and David Paetkau. 2008. Multiple Data Sources Improve DNA-Based Mark–
Recapture Population Estimates of Grizzly Bears. Ecological Applications
18:577–589.
Boulanger J, Proctor M, Himmer S, Stenhouse G, Paetkau D, et al. 2006. An empirical
test of DNA mark-recapture sampling strategies for grizzly bears. Ursus 17: 149–
158.
Boulanger, John, White, Gary C., McLellan, Bruce N., Woods, John, Proctor, Mike,
Himmer, Sefan. 2002. A Meta-Analysis of Grizzly Bear DNA Mark-Recapture
Projects in British Columbia, Canada: Invited Paper. Ursus, 13:137-152. 2002.
54
Boulanger, John and McLellan, Bruce. 2001. Closure Violation in DNA-Based MarkRecapture Estimation of Grizzly Bear Populations. Can. J. Zool 79:642-651.
Efford, M.G. 2012. DENSITY 5.0: software for spatially explicit capture–recapture.
Department of Mathematics and Statistics, University of Otago, Dunedin, New
Zealand.
Harris, R. B., Wall, W.A., Allendorf, F.W. 2002. Genetic Consequences of Hunting:
What Do We Know and What Should We Do? Wildlife Society Bulletin. 30:634643.
Herrero, S. 1972. Aspects of evolution and adaptation in American black bear (Ursus
americanus Pallas) and brown and grizzly bears (U. arctos Linne´.) of North
America. International conference of Bear Research and Management 2:221–231.
Huggins RM 1989. On the statistical analysis of capture experiments. Biometrika 76:
133–140.
Jolly, G. M. 1965. Explicit Estimates From Capture-Recapture Data With Both Death and
Immigration: Stochastic Model. Biometrika 52:225-247.
Karanth, K. U. and Nichols, J.D. 2002. Monitoring Tigers and Their Prey. Centre for
Wildlife Studies, Bangalore, India.
Kendall, Katherine C., Stetz, Jeffery B., Roon, David A., Waits, Lissette P., Boulanger,
John B., Paetkau, David. 2008. Grizzly Bear Density in Glacier National Park,
Montana. Journal of Wildlife Management, 72(8):1693-1705.
Kendall, Katherine C., Stetz, Jeffrey B., Boulanger, John, Macleod, Amy C., Paetkau,
David, White, Gary C. 2009. Demography and Genetic Structure of a Recovering
Grizzly Bear Population. Journal of Wildlife Management, 73(1):3-17.
Knibb, David. 2008. Grizzly Wars: The Public Fight Over the Great Bear. Eastern
Washington University, Washington.
Levi T, Darimont CT, MacDuffee M, Mangel M, Paquet P, et al. 2012. Using Grizzly
Bears to Assess Harvest-Ecosystem Tradeoffs in Salmon Fisheries. PLoS Biol
10(4): e1001303. doi:10.1371/journal.pbio.1001303
Lewis, Tania. 2009. Glacier Bay National Park Bear Management Plan.
Lewis, Tania. 2012. Shoreline Distribution and Landscape Genetics of Bears in a
Recently Deglaciated Fjord: Glacier Bay, Alaska. University of Alaska,
Fairbanks. M.S. Thesis. 103 pp.
55
Long, Robert A. and Zielinski, William J. 2008. “Designing Effective Noninvasive
Carnivore Studies.” Noninvasive Methods for Carnivores. Washington D.C.:
Island Press.
Mattson, D.J., Herrero, S., Merrill, T. 2005. Are Black Bears a Factor in the Restoration
of North American Grizzly Bear Populations? Ursus 16(1):11-30.
Mattson, D.J., Merrill, T. 2002. Extirpations of grizzly bears in the contiguous United
States, 1850–2000. Conservation Biology 16: 1123-1136.
Mattson, D.J. 1998. Diet and morphology of extant and recently extinct northern bears.
Ursus 10:479–496.
Miller, Sterling D., 1989. Population Management of Bears in North America.
International Conference of Bear Restoration and Management. 8:357-373.
Mills, Scott L., Citta, John C., Lair, Kevin P., Schwartz, Michael K., Tallmon, David A.
2000. Estimating Animal Abundance Using Noninvasive DNA Sampling:
Promises and Pitfalls. Ecological Applications 10(1):283-294.
Milner J., Nilsen E. B., Andreassen, H.P. 2007. Demographic Side Effects of Selective
Hunting in Ungulates and Carnivores. Conservation Biology. 21:36-47.
Morin, P.A., and D.S. Woodruff. 1992. Paternity exclusion using multiple hypervariable
microsatellite loci amplified from nuclear DNA of hair cells. Paternity in
Primates: Genetic Tests and Theories. Karger, Basel, Switzerland.
Mowat, G., and Strobeck, C. 2002. Estimating Populations Size of Grizzly Bears Using
Hair Capture, DNA Profiling, and Mark-Recapture Analysis. Journal of Wildlife
Management 64:183-193.
National Parks Conservation Association. 2008. Minding the Gap: Is Wildlife Research
Sufficiently Funded in Alaska’s National Parks. National Parks Conservation
Association, Alaska Regional Office.
National Parks Conservation Association. 2006. Who’s Counting? How Insufficient
Support for Science is Hindering National Park Wildlife Management in Alaska.
National Parks Conservation Association, Alaska Regional Office.
National Park Service. “Glacier Bay 2010 Fact Sheet.” [Brochure] 2010.
National Park Service. “Glacier Bay 2012 Fact Sheet.” [Brochure] 2012.
National Park Service. Date Unknown. Nps.gov. Retrieved Nov. 3, 2013, from
http://www.nps.gov/glba/naturescience/research.htm.
56
Nichols, James D., Hines, James E., Pollock, Kenneth H. 1982. Effects of Permanent
Trap Response in Capture Probability on Jolly-Seber Capture-Recapture Model
Estimates. North Carolina State University Press.
Noss, R.F., Quigley, HB, Hornocker, MG et al. 1996. Conservation Biology and
Carnivore Conservation in the Rocky Mountains. Conservation Biology 10:949963.
Otis, D. L., Burnham, K. P, White, G. C. Anderson, D.R. 1978. Statistical Inference from
Capture Data on Closed Animal Populations. Wildlife Monographs, 62:1-135.
Paetkau, D. 2012. Project g1207 Glacier Bay 2012. Wildlife Genetics International
Project Report.
Paetkau, D. 2003. An Empirical Exploration of Data Quality in DNA Based Population
Inventories. Molecular Ecology, 12:1375-1387.
Peacock, Elizabeth, Titus, Kimberly, Garshelis, David L., Peacock, Mary M., Mirowlaw
Kuc. 2011. Mark-Recapture using Tetracycline and Genetics Reveal Record-High
Bear Density. The Journal of Wildlife Management, 75:1513-1520.
Pelton, M.R. 1982. Black Bear. Wild Animals of North America: Biology, Management,
and Economics. Johns Hopkins University Press, Baltimore and London.
Poole, K.G., Mowat, G. Fear, D.A. 2001. DNA-Based Population Estimate for Grizzly
Bears Ursus Arctos in Northeast British Columbia, Canada. Wildlife Biology,
7:105-115.
Powel, Roger A., Zimmerman, John Wayne, Seaman, David Erran. 1997. Ecology and
Behavior of North American Black Bears. T.J. Press, Padstow, Cornwall,
England.
Sawaya, Michael A., Stetz, Jeffery B., Clevenger, Anthony P., Gibeau, Michael L.,
Kalinowski, Steven T. 2012. Estimating Grizzly and Black Bear Population
Abundance and Trend in Banff National Park Using Noninvasive Genetic
Sampling. Plos One, Vol. 7, Issue 5.
Scott, Ryan. 2009. Unit 1 Brown Bear Management Report. Brown Bear Management
Report of Survey and Inventory Activities 1 July 2006-30 June 2008. Alaska
Department of Fish and Game. Juneau, Alaska.
Seber, G. A. 1965. A Note on the Multiple Recapture Census. Biometrica 52:249-259.
Sherwood, Morgan. 1979. Specious Speciation in the Political History of the Alaskan
Brown Bear. The Western Historical Quarterly, 10(1):49-60.
57
Shoen, J.W. and Beier, L.R. 1990. Brown Bear Habitat Preferences and Brown Bear
Logging and Mining Relationships in Southeast Alaska. Final Report. Alaska
Department of Fish and Game. Federal Aid in Wildlife Restoration Project W-221, Study 4.22. Juneau, Alaska.
Stetz, Jeffrey Brain. 2008. Using Noninvasive Genetic Sampling to Assess and Monitor
Grizzly Bear Population Status in the Northern Continental Divide Ecosystem,
Montana. Graduate Thesis. The University of Montana, Missoula: USA.
Stetz, Jeffrey B., Kendall, Katherine C., Servheen, Christopher. 2010. Evaluation of Bear
Rub Surveys to Monitor Grizzly Bear Population Trends. The Journal of Wildlife
Management, 74(4):860-870.
Taberlet, P., Mattock, H., Dubois-Paganon, C., Bouvet, J. 1993. Sexing Free-Ranging
Brown Bears Ursus Arctos Using Hairs Found in the Field. Molecular Ecology
2:399-403.
Taberlet, P. and Luikart, G. 1999. Noninvasive Genetic Sampling and Individual
Identification. Biological Journal of the Linnean Society 68:41-55.
Taberlet, P., Waits, Lisette P., Leukart, Gordon. 2009. Noninvasive Genetic Sampling:
Look Before You Leap. TREE 14(8):323-327.
Titus, K., Beier, L. 1993. Population and Habitat Quality of Brown Bears on Admiralty
and Chichagof Islands. Alaska Department of Fish and Game. Federal Aid in
Wildlife Restoration. Research Progress Report. Project W-24-1. 40pp.
Thornton, Thomas F. 1992. Subsistence Use of Brown Bear in Southeast Alaska.
Technical Paper No 214. Alaska Department of Fish and Game Division of
Subsistence. Juneau, Alaska.
U.S. Census Bureau. 2010. Gustavus, Alaska. Censusviewer.com. Retrieved 13 April,
2013.
Waits, Juliann L. and Leberg, Paul L. 2000. Biases Associated with Population
Estimation Using Molecular Tagging. Animal Conservation 3(3):191-199.
White, G. 1998. Lecture on Open Population Capture-Recapture Models. Colorado State
University.
White, K.S., Barten, N.L., Crouse, J.A. 2006. Ecology of Moose on the Gustavus
Forelands: Population Eruption, Nutritional Limitation, and Conservation
Implications. Proceedings of the fourth Glacier Bay Science Symposium, Oct. 26
– 28, 2004. U.S. Geological Survey Scientific Investigations Report 2007 – 5047.
P. 25 – 28.
58
Wegan, Michael T., Curtis, Paul D., Rainbolt, Raymond E., Gardner, Beth. 2012.
Temporal Sampling Frame Selection in DNA-Based Capture-Mark-Recapture
Investigations. Ursus 23(1):42-51.
Wilder, James M. 2003. Quantifying Bear Populations and Bear-Human Conflicts using
Non-Invasive Genetic Sampling in the Kennicott Valley of Wrangel-St. Alias
National Park and Preserve, Alaska. Graduate Thesis. University of Idaho: USA.
Williams, B. K., Nichols, J. D., and Conroy, M.J. 2002. Analysis and Management of
Animal Populations. Academic Press, New York.
Woods, John G., Paetkau, David, Lewis, Lewis, Dave, McLellan, Bruce N., Proctor,
Michael, Strobeck, Curtis. 1999. Genetic Tagging of Free-Ranging Black and
Brown Bears. Wildlife Society Bulletin, 27(3):616-627.
59
Pinjuv_K2013.pdf