Content uploaded by Tom S Smith
Author content
All content in this area was uploaded by Tom S Smith on Apr 05, 2019
Content may be subject to copyright.
Original Article
Human–Bear Conflict in Alaska: 1880–2015
TOM S. SMITH,
1
Department of Plant and Wildlife Sciences, Brigham Young University, 5050 LSB, Provo, UT 84602, USA
STEPHEN HERRERO, Environmental Science Program, Faculty of Environmental Design, University of Calgary, Calgary, AB T2N 1N4, Canada
ABSTRACT We present an analysis of human–bear (Ursus spp.) conflicts that occurred in Alaska, USA,
from 1880 to 2015. We collected 682 human–bear conflicts, consisting of 61,226 data entries, from various
sources available to us. We found that human–bear attacks are rare events, averaging 2.6/year across the study
period, though increasing to 7.6/year in the current decade. Grizzly bears (U. arctos) dominated conflicts
(88%), followed by black bears (U. americanus; 11%), and lastly polar bears (U. maritimus; 1%). Although
grizzly bear family groups are often involved in conflicts (32% of all attacks), single grizzlies are involved more
than any other cohort (45%). Human–bear conflicts occurred during every month of the year and the majority
occurred during daytime when people were most active (82%). Human group size was a significant factor in
bear conflicts: the larger the group (2 persons), the less likely to be involved in a confrontation. Habitat
visibility also contributed to conflict, the poorer the visibility the more likely bears were to engage with
people, presumably because of an inability to detect them until very close. When domestic dogs intervened in
attacks, they terminated them nearly half of the time (47.5%). However, in 12.5% of cases, dogs appeared to
have initiated the conflict. When involved, rescuers terminated maulings in 90.3% of cases, but were
themselves mauled 9.7% of the time. We offer these, and other, insights derived from this work that will
inform wildlife biologists’ bear safety training and public outreach. Ó2018 The Wildlife Society.
KEY WORDS Alaska, bear attacks, black bear, brown bear, grizzly bear, human–bear conflict, polar bears, Ursus
americanus,Ursus arctos, Ursus maritimus.
Throughout North America, human–bear conflict periodi-
cally results in serious, and sometimes fatal, injuries to both
humans and bears (Herrero 2002). Conflicts between people
and bears include negative interactions that are aggressive,
defensive, or nuisance in nature (Gore et al. 2006). A number
of studies have investigated human–bear conflict in North
America (Herrero 1970; Middaugh 1987; Miller and
Chihuly 1987; Herrero and Higgins 1998, 1999, 2003;
Herrero and Fleck 1990; Miller and Tutterow 1999;
Gunther et al. 2004; Herrero et al. 2011; Wilder et al.
2017). A few authors have specifically addressed human–bear
conflict in Alaska, USA (Middaugh 1987; Miller and
Chihuly 1987; Miller and Tutterow 1999; Suring and Del
Frate 2002; Smith et al. 2008, 2012). Among these,
Middaugh (1987) analyzed Alaska bear attacks from 1900
to 1985. Miller and Tutterow (1999) reported that brown
bear (synonymous with grizzly bear) attacks resulted in 2.75
injuries and 0.42 deaths/year in Alaska from 1986 to 1996.
Our work here expands on that of previous authors,
including 135 years of human–bear conflict in Alaska.
Most of Alaska is home to polar (Ursus maritimus), black (U.
americanus), and grizzly (U. arctos) bears. All 3 have been
involved in human–bear conflicts that can be minimized by
understanding circumstances associated with such events.
Without this knowledge, fear and ignorance may lead to
conflict that is otherwise avoidable. A willingness of humans to
tolerate apical predators, in spite of the risk of injury they
represent, is key to bear conservation (L€oe and R€oskaft 2004).
It is critical that managers provide bear safety messages that are
based on the best information possible, not on supposition and
conventional wisdom. Accurate data regarding the risk of bear
conflict generate public support for bear management and
conservation (L€oe and R€oskaft 2004, Gore et al. 2006).
Our research objectives were to locate, categorize, and
analyze all available records of human–bear conflict in Alaska,
and synthesize implications for management, including how
bear conflict may be avoided. We focused on both attacks
(injury) and incidents (the person was not injured) because
both provide insight into bear behavior and the most effective
means for dealing with aggressive bears. We discuss the process
used for reviewing and deciding which records to include in
these types of analyses. To construct this data set, we included
variables associated with these events, including specific
aspects of the bear(s) and human(s) involved, temporal and
spatial characteristics of the attack–incident location, and all
other information we believed would foster a better
understanding of the nature of these events.
STUDY AREA
Alaska is located in the northwestern portion of North
America and occupies an area of 1,530,699 km
2
. The human
Received: 9 December 2016; Accepted: 29 December 2017
1
E-mail: tom_smith@byu.edu
Wildlife Society Bulletin; DOI: 10.1002/wsb.870
Smith and Herrero Alaska Human–Bear Conflict 1
population in Alaska was estimated to be 33,426 in 1880, but
grew to 739,828 by 2016. We obtained population estimates
for each decade from 1880 to present from Alaska censuses
(http://live.laborstats.alaska.gov/pop/). The grizzly bear
ranges throughout the state, except for a few islands and
wetlands in western Alaska. The most recent published
estimate of grizzly bears in Alaska was 31,700 (Miller 1993).
Black bears occur in most forested areas of Alaska. Formal
population estimates do not exist for black bears, but the
Alaska Department of Fish and Game states that there were
approximately 100,000 black bears in Alaska (Alaska
Department of Fish and Game 2017). Polar bears are
marine mammals that occasionally venture onto land
(Amstrup 2003). In Alaska, polar bears from both the
Chukchi and Southern Beaufort Seas subpopulations
occasionally range up to 80 km inland, primarily for maternal
denning. Recent estimates of polar bear were about 3,800
(Amstrup 2003).
METHODS
For clarity, we use human–bear conflict definitions consistent
with Smith et al. (2005) and Hopkins et al. (2010). A
“human–bear interaction” (also known as an “encounter”),
occurs when a person and bear are mutually aware of each
other (Smith et al. 2005). Bears may react with seeming
indifference, by leaving the area or by approaching the person
(s). “Human–bear conflict” occurs when a bear has exhibited
stress-related or curious behavior such that a person took
evasive action, a bear made physical contact with a person,
exhibited predatory behavior, or was intentionally harmed or
killed (not including legal harvests) by a person. Human–
bear conflict includes both bear attacks and human–bear
incidents. A “bear attack“ involves intentional contact by a
bear resulting in human injury (Smith et al. 2005). A
“human–bear incident” (or simply “bear incident”) involves
“human–bear conflict” but did not lead to a person that was
physically harmed, yet the person was at significant risk of
injury. By the term “significant risk of injury,” we mean that
the bear could have injured (or even killed) the person,
though the outcome was otherwise. In all such cases, the
person and bear were in close proximity (generally within
<10 m), and sometimes made contact. We included bear
incidents in our database and analyses because the only
difference between an incident and attack was physical
injury; we can learn important aspects of human–bear
conflict from them. Hence, in this paper, we report on both
bear attacks (injury events) and bear incidents (noninjury
events).
No state or federal agency is responsible for maintaining
records of injuries or deaths from bear attacks or bear
incidents in Alaska (Miller and Chihuly 1987). Therefore,
we collected bear incidents from available state and federal
records, news media, books, online computer searches, and
word of mouth. The state of Alaska maintains records of
defense of life or property (DLP) involving the out-of-season
taking of bears when they are damaging property or
threatening persons. We accessed those records up through
1990, but we were not granted access to more recent records,
presumably to protect the identities of persons involved.
Upon request, a number of federal agencies (National Park
Service, U.S. Fish and Wildlife Service, Bureau of Land
Management, U.S. Geological Survey, U.S. Forest Service)
provided access to records of human–bear conflict having
occurred within their respective jurisdictions. We carefully
reviewed these records and included them in the database as
appropriate. By the term “as appropriate,” we mean to say
that some records did not include risk of injury to the person
(s) involved (e.g., a National Park that included a record of a
bear damaging a structure). We did not include events of that
nature. Some agencies maintained very detailed accounts,
including multiple interviews of the person(s) involved.
Others, however, provided records lacking essential details
(i.e., time of day, no. of persons involved, firearm type, etc.),
but we included them because they provided data that
promoted a better understanding of the behaviors by bears
and humans that likely played a role in human–bear conflict.
We searched many online databases of regional news
sources for archived records of human–bear conflict. All
online searches included the term “bear” along with the
following words in various combinations: attack, confronta-
tion, incident, encounter, mauling, and injury. News sources
included both newspapers and television news reports. We
accessed newspapers—such as the Fairbanks Daily News
Miner, the Anchorage Daily News, the Juneau Empire, and
the Chilkat Valley News—over the internet. We also
accessed archived television news reports, such as those from
local television stations in Anchorage, Alaska. Several of
these sources reported on the same conflict and provided
additional details that increased our understanding of a given
confrontation. We drew from all sources to gather as much
detail as possible.
In search of additional human–bear conflicts, as well as
supplementary details regarding records already collected, we
carefully reviewed a number of books devoted to human–bear
conflicts. For example, L. Kaniut published 4 books (Kaniut
1983, 1989, 1997, 2001) that chronicled human–bear
conflict in Alaska. Other works, such as Etling’s (1997)
double volume Bear Attacks: Classic Tales of Dangerous North
American Bears provided additional confrontations from
Alaska. We contacted several of these authors for additional
details and clarification as needed. We occasionally
interviewed persons regarding bear confrontations in which
they were involved so we could gather additional information
or clarify existing reports. For example, we telephoned and
emailed a number of federal and state employees who had
been involved in bear conflict. These interviews provided
valuable insight regarding these events. Occasionally, victims
of bear attacks contacted us with details their encounter,
often providing additional details.
There is no means of independently verifying the accuracy
of human–bear conflict, particularly those instances that
occurred many years previous; therefore, we included all
records that fit our selection criteria. However, we did
exclude those encounters where the person(s) involved was
not at significant risk of injury. For example, Alaska State law
allows for the legal killing of bears out of season, or without a
2 Wildlife Society Bulletin 9999()
harvest permit, when a person’s life or property (DLP) is at stake
(https://www.adfg.alaska.gov/static/license/otherlicense/pdfs/
dlp.pdf). When reviewing Alaska’s records, we found that the
majority of DLP records involved little to no risk to the person
involved (e.g., shot the offending bear from a vehicle, cabin
window, etc.), so we did not include them in analysis. However,
in some DLP records, the person involved was at risk of the bear
injuring them, so we included those records. Some speculate that
more records are needed for a more accurate analysis of human–
bear conflict in Alaska, but we have no reason to believe that the
basicinsightsprovidedbythisanalysiswouldchangein
significant ways with their addition.
We have exercised caution when interpreting each record of
conflict, and are aware that our sample may be unintention-
ally biased. We do believe, nonetheless, that we have not
excluded many records because of the sensational nature of
bear attacks and intense interest in bear–human conflict
shared by wildlife professionals and the public alike.
Nonetheless, we believe that this large sample size has
allowed us to identify important variables associated with
human–bear conflict that will promote human safety and
bear conservation though conflict avoidance. We do caution
the reader, however, to be aware of potential biases in
conclusions based on these data of unproven authenticity,
accuracy, and perhaps possible inconsistencies of reporting
over time.
Human–bear conflicts included in this analysis, involved
1 bears, 1 persons, and the person(s) was either injured
(i.e., an attack) or at significant risk of injury (i.e., an
incident). Our database contained up to 141 variables for
each conflict, including the date and time of conflict,
location, number of persons, bear species, bear cohort,
human injury, success of firearms or bear spray, etc. We made
an effort to restrict analyses and conclusions to those aspects
of these data we felt confident were accurate and unbiased.
Importantly, no single record provided information regard-
ing every variable we assessed, so in analyses and reporting
our results, the number of incidents that contributed
information rarely added up to the total number reported.
For example, few victims reported how long incidents lasted
in minutes; however, some victims did provide a time
duration, and when we analyzed those incidents to provide
an assessment of duration, the number of contributing
confrontations did not add up to the total number of
incidents in the database.
We recorded both the size of groups in bear country (e.g.,
size of a hunting party), as well as the size of the group
involved in the bear encounter. For example, a party of 4
individuals may have been on a hunting trip, but only 1 of
them encountered a bear while hunting solo. Similarly,
when 3 persons were hiking on a trail but not together, the
encounter group size was 1, not 3, when encountering a
bear.
We assigned a person’s activity to 1 of 2 categories: primary
and secondary. Primary activity referred to the reason that a
person(s) was in bear country (i.e., hunting, camping,
logging, etc.), whereas the secondary activity referred to what
the person(s) was doing at the time of the encounter (e.g.,
hiking, hunting, jogging, etc.). After reviewing each event we
ascribed probable cause using the following categories:
1. Wounded—the bear had been shot and was attempting to
evade pursuers when the conflict occurred.
2. Carcass defense—the bear was guarding–feeding upon a
carcass when the person(s) came upon it.
3. Surprise encounter—the human–bear encounter was
abrupt and the human and bear were surprised. Such
encounters generally occurred at close range (<50 m). A
person rounding the bend in a trail and suddenly
encountering a mother and cubs would be an example
of a surprise encounter.
4. Curiosity—the bear appears to have been attracted to a
person, their camp, or property. There is no surprise
element and curiosity appears to have been the bear’s
motivation.
5. Provoked—the person approached the bear, which
triggered a defensive–aggressive reaction. An example
would be a person moving closer for a photograph.
6. Predation—was identified by a series of behaviors—
searching, following and testing, attacking (capturing),
killing, sometimes dragging a person, sometimes burying,
and often feeding upon a person. Vocalizing and stress
behaviors by the bear were usually absent (Herrero and
Higgins 2003).
7. Potential predation—the bear stalked or moved steadily
toward people and was persistent in this approach.
We used linear regression to quantify the relationship
between human population size in Alaska and number of
human–bear conflicts per decade. We obtained Alaska’s
human population statistics from https://www.census.gov/
dmd/www/resapport/states/alaska.pdf.
We used Wallace’s “rule of nines” to subdivide the human
body into representative surficial proportions (Evers et al.
2010). According to Wallace, the surface area of an average
human body can be subdivided as follows: head–neck (9%),
back (18%), chest (18%), arms (9% each), perineum (1%),
and legs (18% each). Bear-inflicted injuries were tallied by
region of the body (i.e., head, back, arms, chest, perineum,
and legs), then compared with expected rates per region had
attack sites been randomly distributed. We determined
differences between observed and expected attack-site
frequencies using chi-square analysis.
In an attempt to determine whether the bear or the
human initiated a given conflict, we qualitatively evaluated
each incident by reviewing whether or not the person(s)
involved adhered to conventional wisdom for bear
avoidance (e.g., making noise while hiking) and deterrence
(e.g., carrying bear spray, a firearm or some other defense).
As we reviewed the details of each incident, we scored
persons a þ1 for each bear safe practice they followed and a
1 for those activities that predispose one to bear
problems. We classified the resulting summed value as
follows: initiator indeterminable, bear was likely the
initiator, bear was the initiator, person was likely the
initiator, and the person was the initiator.
Smith and Herrero Alaska Human–Bear Conflict 3
We qualitatively ranked visibility of the land cover in which
conflicts occurred. Those rated “poor” were either heavily
forested, dense shrub lands, or rough terrain with short sight
distances. Land cover rated “fair” had trees–shrub cover, but
also open areas providing improved visibility, as compared
with those rated “poor.” Land cover rated “good” was
typically very open, such as beaches, alpine meadows, and
fields.
We used the chi-square goodness-of-fit test to determine
whether observed and expected data were independent of
each other, such as whether party group size and bear
encounter group size were different or not (Dytham 2003).
We used the z-test to compare the proportions from 2
independent groups to determine if they were different. We
also used the Pearson product-moment correlation coeffi-
cient (r) as a measure of the strength of the linear relationship
between two variables. We determined significance at
a¼0.05 level for all statistical tests.
RESULTS
We collected 682 human–bear conflicts, consisting of 61,226
data entries, from the various sources available to us in Alaska
(Fig. 1). Using our defined human-at-risk criteria, we
reviewed 650 DLP incidents on file at Alaska Department of
Fish and Game and included 59 (9%) in our database. These
conflicts spanned 135 years (1880–2015), and affected
groups comprised 1,447 persons, although only 1,141 (79%)
of them were directly involved in the confrontation. Of these
conflicts, 326 were noncontact incidents and 352 were
attacks, for average annual occurrences of 2.4 and 2.6/year,
respectively. These conflicts involved 698 bears with 131
dependent offspring, including 543 grizzly bears (78%), 92
black bears (13%), 13 polar bears (2%), and 34 bears of
unreported species (5%).
Injuries
Three-hundred forty six bear-inflicted injuries included 62
fatalities, 59 persons severely injured, 71 persons moderately
injured, 86 persons slightly injured, and 68 persons injured
without the specific degree reported. In 336 incidents (336 of
682 ¼49%), no one was injured but we included these in the
analysis because persons involved were deemed to have been
at significant risk of injury. We found that grizzly bears
(n¼289) inflicted 83.5% of all injuries; black bears (n¼32)
9.2%; polar bears (n¼4 conflicts) 1.2%, and unknown
species of bears 6.1% (n¼21). Of the 62 bear-inflicted
fatalities, grizzly bears accounted for 79.0% (n¼49), black
bears 8.1% (n¼5), polar bears 3.2% (n¼2), and unknown
species of bear 9.7% (n¼6; Table 1).
Specific locations of bear-inflicted injuries on victims’
bodies were reported in 312 of 682 incidents (46%; Fig. 2).
Locations of victims’ injuries differed from those expected
under a nonselective, random scenario (x
22
¼94.46,
P<0.001). The head–neck region of victims was attacked
4.5 times more than expected, while other regions of the
body were involved roughly half as much as expected, other
than the arms, where injuries did not vary from those
expected.
Temporal Nature of Bear Conflicts
Alaska human–bear conflicts steadily increased from the
earliest recorded in 1880 until 2015, with an annual average
of 2.6 attacks/year across the entire study period, 4.8/year in
the past 50 years, and 7.6 attacks in the past decade (Fig. 3).
Alaska population growth and the steady increase in bear
attacks between 1880 and 2015 were found to be highly, and
positively, correlated (r¼0.93, P<0.001).
Bear attacks (n¼300) occurred during every month of the
year in Alaska. Most (145/300; 48%) occurred during the
summer months (Jun–Aug), followed by autumn (Sep–Nov;
78/300; 26%), with equal numbers (38/300; 13%) in winter
(Dec–Feb), and spring (Mar–May). The time of day was
reported for 266 human–bear conflicts, including 141
incidents and 125 attacks (Fig. 4). These conflicts were
not randomly distributed throughout the day (x
223
¼140.40,
P<0.001), with the most incidents occurring between
1000–1200 (23%) and 1700–1900 (20%), with attacks most
frequently occurring from 1400 to 1900 (36%). Few
incidents occurred from midnight through 1000 (18%);
Figure 1. Spatial distribution of human–bear confrontations in Alaska,
USA, 1880–2015.
Table 1. Bear-inflicted injuries caused by grizzly, black, and polar bears in
Alaska, USA, 1880–2015.
Injury level
Unknown
species Polar Black Grizzly
Row
total
Extent
unknown
4 0 10 47 61
No injury 12 7 58 242 319
Slight injury 4 1 11 66 82
Moderate
injury
21 1 61 65
Severe injury 2 0 4 45 51
Fatality 6 2 5 47 60
Grand total 30 11 89 508 638
4 Wildlife Society Bulletin 9999()
whereas, a third of all attacks took place during those early
morning hours.
Estimates of duration were provided for many human–bear
conflicts (attacks: n¼254; incidents: n¼303; Fig. 5). Most
attacks lasted <3 min (n¼137, 54%), and half of all
incidents were <3 min as well (n¼152, 50%). Approxi-
mately 85% of attacks (216 of 254) and 75% of incidents (227
of 303) were over in <10 min. Very few of either category
(7% or 10%, respectively) extended beyond 30 min, though
some lasted more than 1 hr (3% and 5% each).
Profile of Bears Involved
Polar bears conflicts have been so rare (n¼12) that few
conclusions can be made from these data (Table 2). However,
a key difference between black and grizzly bears involved in
conflicts is apparent in the involvement of family groups:
black bear mothers and dependent young rarely engaged with
humans (16% of all black bear conflicts), whereas grizzly bear
family groups were the second only to single bears (32% of all
grizzly bear conflicts; Table 2). Focusing specifically on
attacks data, black bear family groups constituted only 11%
of attacks, whereas grizzly bear family groups accounted for
37% of all attacks (Table 2).
In 426 conflicts (62%), persons reported the bear’s activity
when first observed. For example, upon rounding a bend in a
trail a person may have encountered a bear digging roots,
fishing, or simply walking the trail toward them (Fig. 6). In a
majority of instances (55%), bears were engaged in natural
activities (e.g., resting, walking, foraging, etc.) when first
encountered. In the remaining observations (45%), bears
were already engaged with the person when the person
became aware of them (e.g., charging, stalking, etc.). Bears
involved in conflicts with people (n¼577) had the following
outcomes: bears suffering no injury (n¼351, 61%), bears
wounded to some degree (n¼24, 4%), bears severely injured
(n¼10, 2%), bears that died from human-inflicted injuries
(n¼186, 32%), or died because of management action
(n¼6, 1%).
Profile of Humans Involved
The majority of persons involved in bear conflicts (n¼669)
were adult men (n¼553, 83%), with less than one-fifth as
many women (n¼98, 14%). Few children (n¼24, 4%) were
directly involved in bear conflicts and they were accompanied
by adults in all cases.
Data for 648 groups were considered, except for 7 groups
that were >7 persons (Fig. 7). If bear encounters were
independent of group size, we would expect conflicts to
reflect the proportion of group sizes reported. However, the
observed distribution varied from these expectations, with
single persons involved more than expected and groups 2
involved less than expected (x
26
¼145.21, P<0.001; Fig. 7).
The primary activity (n¼421) persons were most often
engaged in when encountering bears was hunting (n¼190,
45%), followed by hiking (n¼119, 28%; Fig. 8). However,
when confronted by a bear, the most common secondary
activity (n¼376) people were engaged in was hiking
(n¼221, 59%), followed by hunting (n¼59, 16%). This
reversal is due to the fact that many hunters were hiking, not
hunting, at the time they engaged with bears. Persons
conducting research, fishing, and hunting were less often
involved in bear incidents than expected, whereas those
preparing fish and game, as well as hiking, were involved
more often than expected (x
29
¼327.90, P<0.001).
Generally, bears involved in incidents left the area on their
own (n¼420, 72.8%), but were sometimes driven off by
rescuers or dogs. In 17.9% of incidents (n¼124), a person
came to the rescue of the person(s) being mauled. Rescuers
were successful in terminating the mauling 90.3% of the
time, whereas 9.7% (n¼12) of them were also attacked. In
5.9% (n¼40) of incidents, domestic dogs (Canis familiaris)
intervened. Dogs defending persons were successful in
terminating the mauling 47.5% (n¼19) of the time. In 5
instances (12.5%) the dog was likely responsible for inciting
an attack, either by bringing a bear back to its owners (n¼4)
or barking, thus attracting the bear (n¼1).
Bear Avoidance and Deterrence Evaluations
We subjectively evaluated each conflict and ascribed probable
causes to those with sufficient information (n¼596; Fig. 9).
Human-initiated incidents, including surprise encounters
(n¼239), invading bears’ personal space (n¼62), wounding
bears while hunting (n¼27), provoking the bear (n¼7), and
stumbling upon a bear defending a carcass (n¼15),
accounted for the majority of human–bear interactions
(350 of 596 conflicts; 59%). Bear-initiated incidents,
Figure 2. Bodily distribution of bear-inflicted injuries on 313 persons in
Alaska, USA, 1880–2015.
Figure 3. Distribution of human–bear conflicts (incidents þattacks) from
1880 to 2015, Alaska, USA.
Smith and Herrero Alaska Human–Bear Conflict 5
including curiosity (n¼197) and predatory (n¼49),
accounted for 41% of conflicts (246 of 596 incidents).
Examining conflicts in yet greater detail, we evaluated
individuals’ actions with respect to conventional wisdom
pertaining to bear avoidance and deterrence. This analysis
allowed us to assess whether the bear or human had initiated
the conflict (n¼577; Fig. 10). In 24% of these records
(n¼138), we could not determine whether the bear or
person(s) initiated the interaction (“indeterminable”). Bears
were deemed initiator 30% of the time (n¼175 of 577
records), and humans 46% of the time (n¼264 of 577
records). Rankings of land cover in which human–bear
conflicts occurred (n¼403) indicated that the poorer the
visibility, the more likely conflicts were to occur, with the
attack-to-incident ratio highest in the poor visibility areas
(n¼234, 58%; Fig. 11).
The role of firearms and bear spray in resolving human–
bear conflicts has been previously addressed (Herrero and
Higgins 1998; Smith et al. 2008, 2012). As of 2015, 75
instances of bear spray use were recorded, of which 70
(93.3%) were successful in altering bears’ aggressive behavior,
whereas 5 (6.7%) were not. However, of the 197 persons
involved in these 75 encounters, only 4 received slight
injuries (2.0%), all inflicted by grizzly bears. We reviewed
328 conflicts involving firearms and found that firearm
success varied by type, with long guns 75% successful (144 of
193 incidents) and handguns 81% (35 of 43 incidents)
successful in defending the user against aggressive bears,
Figure 4. Temporal distribution (hr of day) of human–bear conflicts in Alaska, USA (n¼266), 1880–2015.
Figure 5. Duration (min) of attacks and conflicts of human–bear conflicts in
Alaska (n¼557), 1880–2015.
Table 2. Demographic composition of bears involved in conflicts with
humans in Alaska, USA, 1880–2015.
Bear species Bear cohort Attacks (n) Incidents (n)
Polar Unknown or single 2 4
Female with young 0 2
Adult females 0 1
Adult males 2 1
Black Unknown or single 29 37
Female with young 4 11
Adult females 2 4
Adult males 0 5
Grizzly Unknown or single 148 123
Female with young 107 64
Adult females 9 22
Adult males 29 40
Total 332 314
6 Wildlife Society Bulletin 9999()
although the differences between handgun and long gun
success were not significant (z¼1.07, P¼0.29). Reasons for
firearms failure included not enough time to react, shots
missed the bear, wounded bear, mechanical failure (i.e.,
short-stroked or mechanism jammed), and reluctance to
shoot (which gave the bear time to make contact). Only long
guns were used on polar bears with 100% kills (n¼6), but
one person was fatally mauled before the bear died. Long
guns were the predominant firearm used on both black (73%)
and grizzly bears (82%). Given that nearly 50% of all
encounters occurred within <10 m, it is not surprising that
firearms can be difficult to bring into play in many bear
encounters (Smith et al. 2012).
DISCUSSION
Our research not only confirms many widely believed tenets
of bear safety, but also provides a number of unique insights.
The risk of aggressive bear encounters in Alaska is very low,
as elsewhere in North America. Undoubtedly, countless
interactions between people and bears occur without incident
(Herrero 2002). Nonetheless, conflict does occur and our
research provides findings that can help to reduce it.
We found a strong, positive correlation between the
increase of human–bear conflicts and human population
growth in Alaska. Herrero et al. (2011) reported a similar
relationship between black bear-inflicted fatalities in North
America and the size of the associated human population.
This relationship between population size and bear attack
frequency has also been reported for other bear species, such
as the sloth bear (Melursus ursinus) in India (Sharp et al.
2017). This suggests that the more people work and recreate
in bear country, the more likely conflict will occur. However,
alternative theories explaining the increase in bear conflict
exist. For example, Shelton (1998) interpreted a rise in bear
conflict in British Columbia, Canada, to be due to wildlife
management policies that supported the release, rather than
euthanizing, of problem bears back into the wild. Alaska
rarely relocates bears; therefore, we believe the steady
increase in human-bear conflict over time is largely due to an
increase in the number of Alaskans and more people
spending time in bear habitat, rather than problem bears
being released into the wild instead of killed.
The number of people in parties involved in bear
confrontations suggests that smaller groups (1 or 2 persons)
are more likely to be attacked than are larger parties. Our
analysis of encounter group size and overall group size
showed that small groups (<2 persons) were much more
likely than expected to be involved in bear conflicts.
Underlying reasons for this relationship may include 1)
Figure 6. Bear activity prior to human–bear conflict in Alaska, USA
(n¼430), 1880–2015.
Figure 7. Size of groups of people involved in human–bear conflicts in
Alaska, USA (n¼638), 1880–2015.
Figure 8. The primary and secondary activities of persons involved in
human–bear conflicts in Alaska, USA, 1880–2015.
Figure 9. The probable cause for human–bear conflicts in Alaska, USA
(n¼592), 1880–2015.
Smith and Herrero Alaska Human–Bear Conflict 7
bears are less likely to be surprised by larger groups because
they are noisier and therefore easier to detect and avoid, and
2) larger groups represent a greater counter-threat to a bear.
Regardless, hiking in groups of 2 persons appears to
provide greater safety when in bear country, but only if
groups remain together rather than dispersed. A group of 5
persons hiking hundreds of meters apart is actually 5 groups
of 1 person each. Importantly, we have no records of 2
persons grouping together and standing their ground when
faced with an aggressive bear and being injured. Unfortu-
nately, bear encounters are generally sudden, chaotic, and
result in some degree of panic with persons running for
safety. In those incidents, bears often pursue fleeing single
persons and attack.
The majority of persons involved in bear attacks were adult
males (83%), with adult females comprising 14% and
children 4%. Children were never involved with bears when
unaccompanied by adults. We do not know the exposure rate
of bears to the different age–sex classes of people in Alaska,
so we cannot tell if bears are more likely to attack a specific
age–sex group of humans. The user groups most often
involved in bear confrontations were hikers, hunters, and
anglers. Middaugh (1987) reported a significant shift in the
activities that persons were engaged in when attacked for the
periods of 1900–1979 to 1980–1985. Earlier in the 20th
century, hunters (49%) were more often engaged in bear
attacks than any other user group (e.g., hikers comprised
13%). After 1980, however, the hunter user group had
declined to 15% of all incidents, while hiking had increased
to 35%. However, Middaugh (1987) had relatively small
sample sizes (79 attacks before 1980 and 26 after 1980).
When we compared our data from before and after 1980, we
found that hunters were involved in 37% of all attacks prior
to 1979 (n¼77) and had declined to 27% (n¼124) after
1979. Before 1979, hiking was involved in 18% of all attacks
(n¼38), but this increased only slightly after 1979 (n¼105)
to 23%. Although our numbers differ, trends are the same as
noted by Middaugh (1987), showing that the percentage of
attacks involving hunters has declined while percentage of
those involving outdoor recreationists has increased. For the
same time period (before and after 1980), joggers involved in
bear attacks increased 19-fold (from 1 incident to 19),
bicyclists 5-fold (from 1 incident to 5), and researchers 7-fold
(from 5 incidents to 33 after 1980). None of these joggers or
bicyclists were carrying a bear deterrent and we believe that
contributed to the outcome.
In our data, no person carrying bear spray was killed, and
98% of persons involved suffered no bear-inflicted injuries.
Those injured (n¼4), received only minor injuries. This
finding is consistent with previous studies of the effectiveness
of bear spray (Herrero and Higgins 1998, Smith et al. 2008).
Persons carrying firearms fared worse (76% success rate
overall; Smith et al. 2012). Attempting to dispatch a
charging bear in heavy cover over uneven ground while under
extreme duress is undoubtedly difficult. Hence, we suggest
only those proficient with firearms in extreme conditions as
present in a sudden bear encounter should rely on them for
protection.
There has been much written about the different
temperaments of North America’s bear species. Herrero
(1972, 2002) speculated that the grizzly bear was most often
involved in injurious human–bear conflicts because of its
evolutionary history. On the sparsely treed tundra of North
America where grizzlies evolved, the best defense may have
been having a good offense: attack first and evaluate later.
Alternately, forest-dwelling black bears have opportunities
to avoid conflict not afforded grizzlies by either climbing a
tree or disappearing into the underbrush. Consequently,
most human–black bear interactions end with the bear
seeking refuge in cover. The paucity of black bear attacks in
135 years of history attest to this strategy. Polar bears, though
occasionally touted as stalkers and killers of man, have failed
largely to live up to that moniker (Cramond 1986, Fleck and
Herrero 1989, Wilder et al. 2017). Polar bears are primarily
dependent on marine mammals, such as ringed seals (Pusa
hispida) for sustenance (Amstrup 2003), and humans are not
a part of bears’ search image for food. Anything as large and
equipped with claw and fang as a polar bear potentially is very
dangerous, but that is not to say that they have an innate
inclination to see humans as food and our data confirm that.
Our findings agree with the conclusions that grizzly bears are
Figure 10. An assessment of responsibility for human–bear conflict in
Alaska, USA (n¼577), 1880–2015.
Figure 11. Relationship between habitat visibility and human–bear
conflicts, Alaska, USA (n¼403), 1880–2015.
8 Wildlife Society Bulletin 9999()
by far the most dangerous of North American bears, with
black bears a distant second, and polar bears the least likely to
engage with humans. Although all 3 species are potentially
dangerous, the innate tendencies of these bears toward
human intrusion indicate very different survival strategies as
evidenced by our data.
From these analyses, we summarize here a number of
insights regarding safety in bear country that will be
beneficial for managers to incorporate in the bear safety
trainings as well as safety messaging to the public:
1. Human–bear conflicts are rare events (2.6/yr overall),
and recent increases (up to 7.6/yr in the most recent
decade) closely track population increases. This suggests
that rather than bears becoming more aggressive over
time, increases in human–bear conflict are largely due to
more people using bear habitat.
2. Our data suggest that grizzly bears are much more likely
to engage in conflict with humans than are black bears,
and the least likely bear to engage in conflict is the polar
bear.
3. When bears attack, they focus on the victims’ head–
neck region 4.5 times more often than would be
expected if the attack site was a random choice. Once
physically attacked, protection of the head and neck is
critical. The defensive positions described by Herrero
(2002) are recommended to protect the head and
neck.
4. Bear conflicts have occurred during every month of the
year in Alaska. Therefore, it is unwise to enter bear
country without a deterrent, or to be careless with food at
any time.
5. Bear conflicts occur mostly during the day (82%) when
people are hiking through bear habitat, not in camp or at
night.
6. Bear attacks are brief, with most lasting <3 min.
7. Grizzly bear family groups are often involved in conflicts
(32% of all attacks), but single bears are involved more
than any other cohort (45%).
8. When in bear country, group size plays an important role
in avoiding bear conflicts: soloists were involved in bear
conflicts significantly more than expected, whereas
groups 2 were involved significantly less than expected.
9. We deemed the majority of bear conflicts as avoidable,
with 60% likely due to people’s inappropriate actions in
bear country.
10. Bear conflicts were most common in poor visibility areas.
When poor visibility areas cannot be avoided, people
should group together and make noise to avoid
surprising bears.
11. Bear spray was highly effective in Alaska, with 98% of
persons using spray avoiding any injury.
12. Firearms were effective 76% of the time when used as
bear deterrents. Only skilled firearms users should rely
primarily on firearms for bear protection.
13. Rescuers made a difference in the outcome of bear attack
victims. When rescuers came to the aid of bear attack
victims, the mauling ended 91% of the time. However,
8.8% of rescuers also suffered injuries by the attacking
bear (12 of 136 instances).
14. Domestic dogs helped terminate maulings nearly half of
the time (47.5%). However, in 12.5% of cases, dogs
appeared to have initiated the attack.
This data set represents a sampling of human–bear conflict
in Alaska. Consequently, we have attempted to limit
conclusions to those aspects of these data that would be
relatively unaffected by the addition of more records. For
example, the distribution of conflicts by time of day or year
would not likely be affected by the addition of more reports
because there is no reason to believe a bias exists in this
respect. Unquestionably, many incidents go unreported for a
variety of reasons. It is believed that many human–bear
interactions resolve peaceably, are not newsworthy, and
therefore, underreported. This includes times when persons
successfully dispatch a bear with a firearm. Such conflicts are
not reported by those who desire to avoid attention for any
number of reasons. That said, although these data are
incomplete, we believe that many useful insights can be
gained from our analysis of the data. Just as provoking a bear
attack for scientific investigation is unethical, so is doing
nothing with available data. It is our hope that insights
gained from this effort will further human safety and bear
conservation.
MANAGEMENT IMPLICATIONS
An important management strategy to reduce bear attacks is
to inform people how to avoid and manage aggressive
encounters with bears. Strategies such as carrying deterrents
like bear spray, traveling in groups of 2 people, and being
alert for bears and bear sign are well-known and supported by
our results. Given the convenience and effectiveness of bear
spray, as well as the efficacy of firearms in the hands of skilled
persons, entering bear country with no deterrent is unwise.
Nonetheless, the vast majority of persons in our database
(n¼855) had no bear deterrent on them, and when faced
with an aggressive or predatory bear, had few options. Hence,
people run, climb trees and play dead, all with poor results,
when they should have been readying a proven deterrent and
standing their ground. We have presented a number of
insights regarding safety in bear country, information that
managers can incorporate into their respective bear safety
messages to help the public and colleagues recreate and work
safely in bear country. It is our hope that efforts such as this
will not only enhance human safety in bear country but also
promote bear conservation through conflict avoidance.
ACKNOWLEDGMENTS
We thank the many persons who provided incidents for this
analysis. We also thank many biological technicians and
assistants who aided in gathering incidents and inputting
them into the database. We thank those who provided
additional insights regarding specific bear encounters. We
are grateful for the constructive comments of Associate
Editor D. Haukos of the Wildlife Society Bulletin and for
valuable input from several anonymous reviewers.
Smith and Herrero Alaska Human–Bear Conflict 9
LITERATURE CITED
Alaska Department of Fish and Game. 2017. Black bear species profile.
http://www.adfg.alaska.gov/index.cfm?adfg¼blackbear.main. Accessed
16 Jun 2017.
Amstrup, S. C. 2003. Polar bear. Pages 587–610 in G. A. Feldhamer, B. C.
Thompson, and J. A. Chapman editors. Wild mammals of North
America: biology, management, and conservation. Second edition. Johns
Hopkins University Press, Baltimore, Maryland, USA.
Cramond, M. 1986. Of bears and man. University of Oklahoma, Norman,
USA.
Dytham, C. 2003. Choosing and using statistics: a biologist’s guide. Second
edition. Blackwell, Malden, Maine, USA.
Etling, K. 1997. Bear attacks: classic talks of dangerous North American
bears. Safari, Long Beach, California, USA.
Evers, L. H., D. Bhavsar, and P. Mailander. 2010. The biology of burn
injury. Experimental Dermatology 19:777–783.
Fleck, S., and S. Herrero. 1989. Polar bear conflicts with humans. Pages
201–202 in M. Bromley, editor. Bear–people conflicts: proceedings of a
symposium on management strategies. Northwest Territories Department
of Renewable Resources, Yellowknife, Canada.
Gore, M. L., B. A. Knuth, P. D. Curtis, and J. E. Shanahan. 2006.
Education programs for reducing American black bear-human conflict:
indicators of success? Ursus 17:75–80.
Gunther, K. A., M. A. Haroldson, K. Frey, S. L. Cain, J. Copeland, and
C. C. Schwartz. 2004. Grizzly bear–human conflicts in the Greater
Yellowstone ecosystem, 1992–2000. Ursus 27:10–22.
Herrero, S. 1970. Human injury inflicted by grizzly bears. Science
170:593–598.
Herrero, S. 1972. Aspects of evolution and adaptation in American black
bears (Ursus americanus Pallas) and brown and grizzly bears (U. arctos
Linne.) of North America. Bears: Their Biology and Management
2:221–231.
Herrero, S. 2002. Bear attacks: their causes and avoidance. Second edition.
Lyons and Burford, New York, New York, USA.
Herrero, S., and S. Fleck. 1990. Injury to people inflicted by black, grizzly or
polar bears: recent trends and new insights. Bears: Their Biology and
Management 8:25–32.
Herrero, S., and A. Higgins. 1998. Field use of capsicum spray as a bear
deterrent. Ursus 10:533–537.
Herrero, S., and A. Higgins. 1999. Human injuries inflicted by bears in
British Columbia: 1960–97. Ursus 11:209–218.
Herrero, S., and A. Higgins. 2003. Human injuries inflicted by bears in
Alberta: 1960–98. Ursus 14:44–54.
Herrero, S., A. Higgins, J. E. Cardoza, L. I. Hajduk, and T. S. Smith. 2011.
Fatal attacks by American black bear on people: 1900–2009. Journal of
Wildlife Management 75:596–603.
Hopkins, J. B., S. Herrero, R. T. Shideler, K. A. Gunther, C. C. Schwartz,
and S. T. Kalinowski. 2010. A proposed lexicon of terms and concepts for
human-bear management in North America. Ursus 21:154–168.
Kaniut, L. 1983. Alaska bear tales. Alaska Northwest Books, Seattle,
Washington, USA.
Kaniut, L. 1989. More Alaska bear tales. Alaska Northwest Books, Seattle,
Washington, USA.
Kaniut, L. 1997. Some bears kill. Safari, Long Beach, California, USA.
Kaniut, L. 2001. Bear tales for the ages: from Alaska and beyond. Paper
Talk. Anchorage, Alaska, USA.
L€oe, J. and E. R€oskaft. 2004. Large carnivores and human safety: a review.
AMBIO: A Journal of the Human Environment 33:283–288.
Middaugh, J. P. 1987. Human injury from bear attacks in Alaska, 1900–
1985. Alaska Medicine 29:121–126.
Miller, S. D. 1993. Brown bears in Alaska: a statewide management report. Alaska
Department of Fish and Game Wildlife Techniques Bulletin 11, Juneau, USA.
Miller, S. D., and M. A. Chihuly. 1987. Characteristics of nonsport brown
bear deaths in Alaska. International Conference on Bear Research and
Management 7:51–58.
Miller, S. D., and V. L. Tutterow. 1999. Characteristics of nonsport
mortalities to brown and black bears and human injuries from bears in
Alaska. Ursus 11:239–252.
Sharp, T. R., S. Swaminathan, A. S. Arun, T. S. Smith, K. Satyanarayan,
and G. Seshamani. 2017. Sloth bear attack behavior and a behavioral
approach to safety. Final report to the International Association for Bear
Research and Management. https://bearbiology.com/wp-content/
uploads/2017/11/Sharp-2017_Sloth-Bear-Attack-Behavior-and-a-
Behavioral-Approach-to-Safety.pdf. Accessed 8 Mar 2018.
Shelton, J. G. 1998. Bear attacks: the deadly truth. Pogany Productions,
Hagensborg, British Columbia, Canada.
Smith T. S., S. Herrero, and T. D. DeBruyn. 2005. Alaskan brown bears,
humans, and habituation. Ursus 16:1–10.
Smith, T. S., S. Herrero, T. D. DeBruyn, and J. M. Wilder. 2008. Efficacy of
bear deterrentspray in Alaska. Journal of WildlifeManagement 72:640–645.
Smith, T. S., S. Herrero, C. S. Layton, R. Larsen, and K. R. Johnson. 2012.
Efficacy of firearms for bear deterrence in Alaska. Journal of Wildlife
Management 76:1021–1027.
Suring, L. H., and G. Del Frate. 2002. Spatial analysis of locations of brown
bears killed in defense of life or property on the Kenai Peninsula, Alaska.
Ursus 13:237–245.
Wilder, J. M., D. Vongraven, T. Attwood, B. Hansen, A. Jessen, A.
Kochnev, G. York, R. Vallendaer, D. Hedman, and M. Gibbons. 2017.
Polar bear attacks on humans: implications of a changing climate. Wildlife
Society Bulletin 41:537–547.
Associate Editor: Haukos.
10 Wildlife Society Bulletin 9999()