Content uploaded by Petra Kaczensky
All content in this area was uploaded by Petra Kaczensky on Jun 20, 2020
Content may be subject to copyright.
Human inﬂuence on the choice of winter dens by European
brown bears in Slovenia
, Felix Knauer
, Petra Kaczensky
Wildlife Research and Management Unit, Technische Universit€atM€unchen, Linderhof 2, D-82488 Ettal, Germany
Alterra Green World Research, Landscape Centre, Team Ecological Modelling and Monitoring, P.O. Box 47, 6700 AA Wageningen, The Netherlands
Institute of Wildlife Biology and Game Management at the Agricultural University of Vienna, Peter Jordan Straße 76, A-1190 Vienna, Austria
International Takhi Group – Research, Salzburg Zoo, A-5081 Anif, Austria
Received 15 February 2003; received in revised form 16 July 2003; accepted 16 July 2003
The Slovenian brown bear (Ursus arctos ) population is the only viable population in Central and Western Europe, and it coexists
with humans in a multi-use landscape. Bears are most vulnerable to human disturbances during denning. To assess the inﬂuence of
humans on the choice of winter dens by bears we compared availability and use of caves suitable for denning in central Slovenia.
Surprisingly, all direct measures of human inﬂuence showed no or only a small eﬀect on the use of the caves by bears. We found that
the landscape type (big dolines, canyons, river valleys, and karst plateau) was the most important variable. The less accessible a
landscape type is, the more it is used. The probability that a cave in a big doline is used is about 200 times higher than on the karst
plateau. Furthermore bears preferred long caves with small entrances away from villages. Bears did not use any cave <540 m
distance to the next village. In total, three quarters of all caves were predicted correctly by our model of being used or unused. For
conservation and human safety reasons, human activity should be banned from steep ravines and large karst dolines in winter.
Ó2004 Elsevier Ltd. All rights reserved.
Keywords: Brown bear; Human–bear co-existence; Human inﬂuence; Denning behaviour; Multi-use landscape
In Central and Western Europe, brown bears (Ursus
arctos) have been extirpated in most of the areas and the
remaining populations have been reduced to a few small,
isolated and often non-viable populations. Today, only
the bear population of the Alps is slowly expanding, due
to re-introductions and natural dispersal from the large
bear population of the Dinaric Mountains (Fig. 1;
Swenson et al., 2000; Kaczensky and Knauer, 2001),
which is the only natural source for the recovery of the
bears in the Alps. The northern tip of the Dinaric
population is located in Slovenia, where bears and
people coexist in a multi-use landscape.
In the densely settled areas of Central Europe human
activity has become an important factor inﬂuencing the
habitat choice of brown bears (Kusak and Huber, 1998;
Kaczensky, 2000b; Knauer, 2000). During the denning
period especially, brown bears are very sensitive to hu-
man disturbance (Swenson et al., 1997; Linnell et al.,
2000) and hence suitable winter dens are an important
component of brown bear habitat. Brown bears use a
wide variety of sites as winter dens depending on abiotic
factors such as climate and geomorphology, and biotic
factors such as sex, age, reproductive status and indi-
vidual preferences (Craighead and Craighead, 1972;
Vroom et al., 1980). Winter dens shelter bears from
unfavourable weather conditions and help to save vital
energy (Folk et al., 1976; Nelson et al., 1983). Even more
importantly, the den provides protection from predation
and human disturbance, especially for reproducing fe-
males (Servheen, 1983; Schoen et al., 1987; MacHutchon
et al., 1993). Surprise encounters by humans with
Corresponding author. Tel.: +31-317-477889; fax: +31-317-424988.
E-mail address: firstname.lastname@example.org (F. Knauer).
0006-3207/$ - see front matter Ó2004 Elsevier Ltd. All rights reserved.
Biological Conservation 119 (2004) 129–136
denning bears not only disturb them, but may also result
in serious human injury (Kaczensky, 2000a; B. Krze,
personal communication), a major concern in the den-
sely populated bear areas in Europe.
In North America, Tietje and Ruﬀ (1980), Johnson
and Pelton (1981), and Schoen et al. (1987) have shown
that bears choose remote and undisturbed areas for
denning, far from sites of human activity. In most areas
of Europe, human population density in brown bear
range is much higher than in North America (Mattson,
1989). However, bears in Europe have had more time to
adapt to human disturbance and persecution. Several
studies in Europe have tried to assess requirements of
critical denning habitat by describing known winter dens
and by measuring the distances of dens to human in-
frastructure (Huber and Roth, 1997; Groﬀ et al., 1998).
However, describing the choice of winter dens only,
based on the description of used winter dens without
knowing the availability of suitable dens, may cause
In Slovenia a high-density brown bear population (1
bear per 10 km2) coexists with people in a multi-use
landscape. Bears seem to tolerate predictable human
activity at close range, at least in summer time (Kac-
zensky, 2000b). Very little data on the location of winter
dens was available, but bears were generally believed to
den only in remote locations. In south central Slovenia,
Podlogar (1997) found that dens were at least 1 km away
from houses, 70 m from roads and 40 m from log-
skidding trails. In southern Slovenia, Potocnik and Kos
(1998) found no den closer than 40 m from a forest road.
But again, no information about the availability of caves
suitable for denning was available.
The objective of this study was to quantify factors
inﬂuencing the choice of winter dens by brown bears by
comparing the availability and use of suitable caves in
central Slovenia. Such data would help to identify crit-
ical denning habitat and adapt land use practices in
order to minimise the disturbance of bears and the
chance of surprise encounters of humans with denning
bears. The results would be applied to assess habitat
suitability for brown bear recovery in the eastern Alps
and for feasibility studies in similar areas.
2.1. Study area
The study area comprises 200 km2and is located in
central Slovenia, about 30 km south-west of the capital
Ljubljana. The relief shows typical karst scenery, with
steep canyons, caves and shallow soils. Elevation ranges
from 300 to 1200 m, annual precipitation averages 1500
mm and annual temperature 7–8 °C (Republic of
Slovenia, 1997). In most winters there is a permanent
snow cover for a few months.
The study area can be divided into four major land-
scape types (Fig. 2).
(I) The karst plateau. This covers 73 km2and is strewn
with rocks, small dolines (up to 10 m deep and 50
m in diameter) and caves. Because of the shallow
soil and the irregularity of the terrain the only pos-
sible land-use is forestry. Forest road density is 1.8
km/km2and allows easy access for people. Other
human activities on the karst plateau are hunting
and recreation (biking, hiking, and mushroom
picking). There are no permanent habitations on
the karst plateau.
(II) Seven big dolines, resulting from collapsed large
caves, are situated on the karst plateau. They are
mainly circular, up to 100 m deep, 300 m wide
and with steep rock faces along the walls. Many
caves can be found, mainly from old subterranean
river channels. Because of their inaccessibility
there is hardly any human use despite the fact that
roads are nearby. All these big dolines are situated
on the karst plateau.
(III) There are two canyon systems, the larger one
covering about 24 km2and the smaller one 8
km2. The slopes are steep with multiple rock
faces containing caves. The area is used for for-
estry, hunting and recreation. Parts of the valley
of the bigger canyon are rather densely settled,
Fig. 1. Bear distribution in Europe west of Russia. The shaded areas
show the actual distribution range, the dark dots re-introduced pop-
ulations, and the dotted black line the border of Slovenia.
130 W. Petram et al. / Biological Conservation 119 (2004) 129–136
and the sounds from human activities are audible
in the whole canyon. In the smaller canyon there
are no settlements, but it is frequently used for
recreation, hunting and forestry. Forest road
density is 1.0 km/km2.
(IV) The river valley area is the largest part of the study
area and covers 93 km2. This area is dominated by
rolling hills and small valleys with creeks and small
rivers. Many small villages occur in a mosaic of ag-
ricultural land and forests, dissected by roads.
Caves are only located in the river valleys, which
are characterised by rather gentle slopes with few
rock faces and are more accessible than the can-
yons and are therefore regularly frequented by
hunters, foresters and recreationalists. Forest road
density is about 1.0 km/km2.
Bear habitat consists of mixed, uneven aged forests of
Abieto-Fagetum-Dinaricum, which is dominated by
beech (Fagus sylvatica) and ﬁr (Abies alba), intermingled
with varying amounts of spruce (Picea abies), sycamore
(Acer pseudoplatanus) and elm (Ulmus sp.); forest cov-
erage averages >80%. The forest is intensively managed,
but only selective cutting practices (single tree extrac-
tion) are allowed. Human population density averages
42 inhabitants per km2, but is highly variable within the
study area. Most people live in villages. Single houses
are very rare. Other large and medium sized predators
present in the study area are wolves (Canis lupus),
Eurasian lynx (Lynx lynx), Eurasian wild cat (Felis
sylvestris), red fox (Vulpes vulpes), badger (Meles meles)
and occasionally golden jackal (C. aureus). The ungulate
population consists of red deer (Cervus elaphus), roe
Fig. 2. The four landscape types of the study area and the location of used and suitable but unused caves.
W. Petram et al. / Biological Conservation 119 (2004) 129–136 131
deer (Capreolus capreolus), chamois (Rupicapra rupica-
pra) and wild boar (Sus scrofa).
2.2. Data collection
We conducted all ﬁeldwork between May and Octo-
ber 1998, which is outside the denning period. We sys-
tematically searched for caves in 193 evenly distributed
plots, 250 250 m in size. These plots were located at
the south-western corner of the standard km grid on
1:25,000 topographic maps more or less covering land-
scape types I, III and IV, and together amounted to 12
km2. To increase our sample size, we additionally sear-
ched in all big dolines and tried to search through all
larger rocky outcrops in canyons and river valleys. The
total area searched was about 19.5 km2or 10% of the
Use of caves as winter dens by brown bears was
evaluated by the presence of a nest, an excavation and/
or additional indications such as bear scratches and
hairs. Caves without bear signs were classiﬁed as unused
caves, but these were classiﬁed as suitable if character-
istics were within the range of caves used as winter dens.
We measured a variety of variables on three diﬀerent
spatial scales to characterise used and unused caves: the
cave itself, immediately around the cave and the general
For the ﬁrst we noted whether the cave was a natural
cave, a cleft, or a hollow under boulders and the di-
mension of the entrance, tunnel and chamber. We also
assessed the relative humidity of the cave, measured the
dimensions of the (potential) nest, and whether or not
light reached it. The nest could contain vegetation
material such as leaves and roots, but often only a
depression was excavated in the ground.
Around the cave we assessed the visibility and ac-
cessibility of the cave entrance, and mapped the aspect
of the cave entrance as well as the vegetation in a plot
20 20 m around the entrance. The vegetation was
measured as percentage cover of the ground, shrub and
tree layers. We additionally mapped the proportion of
deciduous and coniferous trees.
Finally in an area of 250 250 m around the entrance
we mapped average steepness and aspect of the slope, as
well as the average distance to settlements (more than
ﬁve inhabitants), highways, railways, paved roads and
forest roads. In addition we described the vegetation in
the same way as mentioned above.
Before we analysed the factors inﬂuencing the choice
of dens, we conducted a threshold analysis in order to
distinguish suitable unused from unsuitable unused
caves. As threshold values we used the minimal width of
the nest found in the 46 used dens, the height from the
middle of the nest to the cave ceiling and the minimum
and maximum volume of the chamber (Table 1).
To ﬁnd out the most important factors determining
use or non-use of suitable dens by brown bears we used
a multiple logistic regression model (Fahrmeir et al.,
1996). After a pre-selection of the variables from an
ecological point of view we conducted a backward var-
iable selection procedure. Additionally we analysed the
background of the resulting model by bivariate statis-
tics. In multiple comparisons we conducted a Bonfe-
roni–Holm correction (Bonferroni, 1936; Holm, 1977;
Holm, 1979, in Quednau, 1992).
Out of 115 caves in total we found 107 suitable caves
of which 46 had been used and 61 had not been used by
bears, but were suitable for denning, according to the
threshold analysis (Table 1).
Of the used caves, 80.5% were natural caves, 17.5%
were clefts and 2% of the winter dens were located under
big boulders. 75.5% of the unused caves were natural
caves, 21.5% clefts and 3% were under boulders. In the
following we call all these structures ‘caves’. There was
no diﬀerence between the type of used and unused caves
The density of all caves in the four landscapes types
was very diﬀerent and in decreasing order of frequency
showed the following trend: big dolines (23 caves/
km2)>canyons (7.5 caves/km2)>karst plateau (5 caves/
km2)>river valley area (1.2 caves/km2).
Table 2 gives an overview of the values for the vari-
ables, which we considered to be most important from
ecological reasons. There were no signiﬁcant diﬀerences
between used and unused caves (t-test).
3.1. Variables inﬂuencing den choice
Three quarters (75.7%) of all caves were correctly de-
termined by our logistic regression model (Table 3). The
landscape type was the strongest predictor for the use of a
cave. Caves in big dolines had the highest chance of being
used followed by caves in canyons, in the river valleys area
and least likely on the karst plateau. The probability that
Ranges for classifying non-used caves as potential dens
Variable Used cave Non-used
Minimum volume of the chamber (m3) 0.50 <0.45
Maximum volume of the chamber (m3) 9.9 >10
Minimum width of nest (cm) 40 <40
Minimum height above the nest (cm) 60 <50
132 W. Petram et al. / Biological Conservation 119 (2004) 129–136
Comparison of used and unused caves of all landscape types
Variable Used caves (n¼46) Unused cave (n¼61)
Mean SD Minimum Maximum Mean SD Minimum Maximum
Median entrance height (cm) 59 37 20 255 69 48 24 350
Length of cave (cm) 394 239 100 960 292 173 120 1080
Steepness of slope (°) 37 12 0 58 29 21 0 90
Distance to villages (m) 1818 753 536 3841 1788 1077 273 4411
Distance to forest roads (m) 291 282 32 1029 270 264 5 1082
Logistic regression model, calculated with the dependent variable – used caves (n¼46) and non-used caves (n¼61) – and a set of independent
variables characterizing the caves. The band Pvalues for Big dolines, Canyons, and River valley are measured against Karst plateau caves as the
standard. The term ‘diﬀ. )2*log likelihood’ gives an estimate of the relative importance of the four main variables (*, P<¼0:05; **, P<0:01; ***,
Variable bSE exp (b)PDiﬀ. )2log likelihood
Landscape type 36.722
Big dolines 5.3570 1.2051 212.0862
Canyons 4.6285 1.3363 102.3614
River valley 4.0827 1.5596 59.3073
Median height of entrance )0.0281 0.0077 0.9723 18.121
Cave length 0.0054 0.0018 1.0055 12.185
Distance to village 0.0009 0.0004 1.0009 4.520
Constant )5.8166 1.7093
Bivariate comparison of used and non-used caves within landscape types (U-test, P<0:05)
Variable Used caves Non-used caves
Mean SD nMean SD nP
Median entrance height (cm)
Big dolines 72 44 25 96 82 14 >0.1
Canyons 43 14 15 68 34 20
River valley 36 9 3 66 18 7
Karst plateau 58 27 3 54 23 20 >0.1
Length of cave (cm)
Big dolines 335 247 25 292 242 14 >0.1
Canyons 275 176 15 258 124 20 >0.1
River valley 450 223 3 208 58 7 >0.1
Karst plateau 726 135 3 356 173 20
Steepness of slope (°)
Big dolines 38 11 25 43 16 14 >0.1
Canyons 40 8 15 38 17 20 >0.1
River valley 40 10 3 39 5 7 >0.1
Karst plateau 9 9 3 7 10 20 >0.1
Distance to villages (m)
Big dolines 2211 284 25 1693 901 14 >0.1
Canyons 1089 488 15 1172 623 20 >0.1
River valley 900 178 3 822 242 7 >0.1
Karst plateau 3104 657 3 2811 922 20 >0.1
Distance to forest roads (m)
Big dolines 170 96 25 155 74 14 >0.1
Canyons 487 362 15 565 332 20 >0.1
River valley 148 81 3 222 187 7 >0.1
Karst plateau 148 100 3 137 87 20 >0.1
W. Petram et al. / Biological Conservation 119 (2004) 129–136 133
bears use a cave in a big doline was roughly 200 times
higher than on the karst plateau. The second most im-
portant variable predicting cave use was the entrance
height – bears preferred smaller entrances. The third im-
portant variable was cave length – bears preferred longer
caves. The least important, but still signiﬁcant variable
was the distance to villages – bears preferred caves farther
away from villages.
3.2. Bivariate comparison
In the canyons and the river valley area bears pre-
ferred lower cave entrances, and on the karst plateau
they showed signiﬁcant preference for longer caves, but
otherwise they showed no such preferences for any cave
types (Table 4). The number of caves close to villages
was too small to make a meaningful statistical com-
parison within or among landscape types. However, no
cave closer than 540 m to any village was used as a
winter den (Fig. 3).
We also found no used cave closer than 50 m to a forest
road, but the sample size of suitable caves was small
(n¼5). Beyond 50 m, forest roads did not seem to have an
inﬂuence on the use of caves as winter dens in any of the
landscape types. Humidity and vegetation also showed no
signiﬁcant diﬀerences among and within the diﬀerent
landscape types and all caves were located in forest. In
general, access to caves in the big dolines and canyons was
more diﬃcult, but within these landscape types, bears also
used caves that were easily accessible.
We are able to quantitatively determine the inﬂuence
of human infrastructure and activity on the choice of
winter dens by brown bears. To our knowledge, this is
the ﬁrst study of den availability and use by brown
bears. In Slovenia, bears mostly use caves for hiber-
nating. This favoured our approach as the availability is
measurable. According to our radiotracking data (own
unpublished data), bears only use the caves for hiber-
nating, but not as day beds in summer. Otherwise it
would have been diﬃcult to separate the use in summer
from denning in winter by bear signs. We limited our
study area to about the size of the home range of one
male or about three female bears. Because of this we can
assume the same bear density over the whole study area.
In Alaska, Schoen et al. (1987) estimated the avail-
ability of suitable denning habitat from landscape vari-
ables, but without mapping potential dens. For
American black bears (U. americanus), Johnson and
Pelton (1981) also used an availability/use approach,
showing a signiﬁcant preference of tree dens over
ground dens. Most other studies only described used
dens and tried to draw conclusions on the inﬂuence of
human disturbance (Swenson et al., 1997) or human
infrastructure (Huber and Roth, 1997; Groﬀ et al.,
1998). But, because the availability of suitable dens was
not measured, the interpretation of these results is dif-
ﬁcult and the human impact might be overestimated.
Our ﬁndings show a quite low impact of human in-
frastructure, but we should point out that most suitable
dens were in steep and rocky areas. We conclude that
bears generally preferred landscape types where the
chances of encounters with humans are lowest. During
the denning period in winter, almost no people enter big
dolines or canyons, hence areas with a lot of caves re-
ceive very little or no disturbance. This means that the
real eﬀect of human inﬂuence can be higher, but is in-
cluded in the variable landscape type. If this is so, then
the avoidance of humans acts on a large scale. This is in
accordance with results from North America, where
bears also prefer remote areas (Tietje and Ruﬀ, 1980;
Johnson and Pelton, 1981; Schoen et al., 1987; McLellan
and Shackleton, 1989) and retreat distances are much
larger than in Slovenia.
In accessible areas bears used long caves and/or caves
with small entrances only, whereas in the inaccessible
dolines, far from villages, almost all suitable caves were
used. From this study, we cannot derive predictions
about ﬂat areas without caves or comparable structures.
It might be possible for bears to tolerate a higher rate of
disturbance, but in our study area they had access to
some very secure dens.
In Slovenia, forestry and hunting activities do not
seem to have a negative impact on the brown bear
population. Nevertheless, we know of three cases where
our radio-collared bears were disturbed by humans
while denning (own unpublished data) – two reproduc-
tive females and one subadult male. Whereas the females
did not change dens, the male left but came back after a
week to resume denning. Although the disturbance did
not seem to have caused much harm, data from Sweden
show that the reproductive success of females that are
500 1500 2500 3500 4500
distance in m
Fig. 3. Frequency of caves at diﬀerent distances from villages. 500 m
seem to be a critical distance to villages. We had a total of nine suitable
caves within this distance to villages in the canyons, dolines and river
valleys, but none of them was used (H0: random, contingency tables,
Fisher test: P¼0:010).
134 W. Petram et al. / Biological Conservation 119 (2004) 129–136
disturbed in a den is signiﬁcantly lower than that of
undisturbed bears (Swenson et al., 1997). In addition,
there have been several close encounters of people and
bears because people accidentally approached bears in
their winter dens (I. Kos, personal communication, B.
Krze, personal communication, S. Stock, personal
communication). In three such cases the person was
severely injured by the bear.
5. Implications for conservation
Our results can be used to evaluate the suitability of
denning habitat in comparable mountainous and alpine
areas. The following predictions can be derived from
Bears generally select inaccessible areas that provide a
low potential of human disturbance. Areas that allow
easy access (e.g. ﬂat areas) and receive high human use
are unsuitable. In such areas bears have much higher
safety requirements concerning cave dimensions and
distances to human infrastructure.
In less accessible regions brown bears are less sus-
ceptible to human disturbance and less rigorous in their
den choice. However, to reduce the disturbance poten-
tial for the bears and the risk of surprise encounters
between people and bears, forestry and hunting activi-
ties (especially drive hunts for wild boar) should be
minimised along steep slopes, in big dolines and near
rocky outcrops during the denning period.
For the eastern Alps we predict that, due to the relief,
suitable areas for denning are probably not a limiting
factor. However, in a region with few natural and suitable
caves, bears have to hibernate in open nests or have to
excavate dens themselves. In such a situation, the minimal
distance a bear feels safe from human disturbance might
be longer and any disturbance more critical.
We like to thank Eloy Revilla, Ren
ee de Mutsert, and
an anonymous referee for helpful comments and Wolf-
oder, Klaus Horstmann, Hartmut Gossow and
Blaz Krze for supervising and support in fund raising.
uchenhoﬀ from the Statistical Advice Oﬃce at
the University of Munich and his student Thomas Nitter
provided valuable support for the statistical analysis.
Prof. Sustersic of the Geological Institute at the Uni-
versity of Ljubljana, Milan Podlogar of the Slovenian
Forest Service, and Ivan Kos and Hubert Potocnik of the
Biological Institute at the University of Lubljana helped
with their knowledge about bears, winter dens and ge-
ology in Slovenia. Special thanks also to the many people
that helped with the rather demanding ﬁeld work, espe-
cially: Christine Große, Mateja Blazic, Klemen Sustersic,
Verena Prantner, Evi Tschunko, Falk Petram, Heike
artner, Christian Leidenberger, Ilka Reinhardt,
Marcus Regelmann, Sonja Sinnmayer, Monika Kern,
Daniel Steuer, Philipp Fersterer, Gutrun Batek, Milo
and Knut Renault. Funding for this research was pro-
vided by the Austrian Science Foundation (FWF project:
P 11529-BIO), the Donors Association for the Promo-
tion of Science and Humanities in Germany, and the
Slovenian Hunters Association.
Bonferroni, C.E., 1936. Theoria statistica classi a calcolo delle
probabilita. Pubbliccazione Ricerca Internazionale Superiore Sci-
enza Economica Commerciale Firenze 8, 1–62.
Craighead, F.C., Craighead, J.J., 1972. Grizzly bear prehibernation
and denning activities as determined by radio tracking. Wildlife
Monographs 32, 35.
Fahrmeir, L., Hamrele, A., Tutz. G., 1996. Multivariate statistische
Verfahren, second ed., 247 pp.
Folk, G.E., Larson, A., Folk, M.A. 1976. Physiology of hibernating
bears. In: International Conference on Bear Research and Man-
agement, vol. 3, pp. 373–380.
Groﬀ, C., Caliari, A., Dorigatti, E., Gozzi, A., 1998. Selection on
denning caves by brown bears in Trentino, Italy. Ursus 10,
Holm, S., 1977. Sequentially rejective multiple test procedures.
Statistical research report 1977-1. University of Umea, Sweden.
Holm, S., 1979. A simple sequentially rejective multiple test proce-
dures. Scandinavian Journal of Statistics 6, 65–70.
Huber, D., Roth, H.U., 1997. Denning of brown bears in Croatia.
International Conference on Bear Research and Management 9 (2),
Johnson, K.G., Pelton, M.R., 1981. Selection and availability of dens
for black bears in Tennessee. Journal of Wildlife Management 45,
Kaczensky, P., 2000a. Bear politics in Slovenia. International Bear
News 9 (2), 9.
Kaczensky, P., 2000b. Co-existence of brown bears and men in
Slovenia. Dissertation am Fachgebiet f€
ur Wildbiologie und Wild-
tiermanagement, Department f€
Okosystem- und Landschafts-
management der Technischen Universit€
Germany. Available from http//tumb1.biblio.tu-muenchen.de/
Kaczensky, P., Knauer, F., 2001. Wiederkehr des Braunb€
aren in die
Alpen – Erfahrungen mit einem anspruchsvollen Großr€
age zur Jagd- und Wildforschung, 26,67–75.
Knauer, F., 2000. Dispersal and expansion patterns of brown bears in
the Eastern Alps. Dissertation am Fachgebiet f€
ur Wildbiologie und
Wildtiermanagement, Department f€
Okosystem- und Lands-
chaftsmanagement der Technischen Universit€
Germany. Available from http,//tumb1.biblio.tu-muenchen.de/
Kusak, J., Huber, D., 1998. Brown bear habitat quality in Gorski
Kotar, Croatia. Ursus, 10,281–291.
Linnell, J.D., Swenson, J.E., Andersen, R., Barbu, P., Barnes, B., 2000.
How vulnerable are denning bears to disturbance? Wildlife Society
Bulletin 28 (2), 400–413.
MacHutchon, A.G., Himmer, S., Bryden, C.A., 1993. Khutzeymaten
Valley grizzly bear study. Ministry of Forest and Wildlife, Habitat
Research Report R-25. Victoria, B.C., Canada.
Mattson, D.J., 1989. Human impacts on bear habitat use. In:
International Conference on Bear Research and Management,
vol. 8, pp. 33–56.
W. Petram et al. / Biological Conservation 119 (2004) 129–136 135
McLellan, B.N., Shackleton, D.M., 1989. Immediate reactions of
grizzly bears to human activities. Wildlife Society Bulletin 17, 269–
Nelson, R.A., Folk, G.E., Jr., Pfeiﬀer, E.W., Craighead, J.J., Jonkel,
C.J., Steiger, D.L., 1983. Behavior, biochemistry, hibernation in
black, grizzly, polar bears. In: International Conference on Bear
Research and Management, vol. 5, pp. 284–290.
Podlogar, M., 1997. Man and bear in the area of Krim and Mokrec.
Potocnik, H., Kos, I., 1998. Preliminary data of brown bear den
characteristics and its occupation in Kocevje Forests, (Slovenia).
In: Euro-American Mammal Congress, Santiago de Compostela,
Spain, 19–24 July, p. 266 (443).
Quednau, H.D., 1992. Das Testen statistischer Hypothesenfamilien.
Acta Sanmarinensia 2.5, ISBN 83-85033-07-1.
Republic of Slovenia. 1997. Statistical Yearbook 1997. Statistical
Oﬃce of the Republic of Slovenia, Ljubljana, Slovenia.
Schoen, W.J., Beier, L.R., Lentfer, J.W., Johnson, L.J., 1987. Denning
ecology of brown bears on Admirality and Chichagof islands,
Alaska. In: International Conference on Bear Research and
Management, vol. 7, pp. 293–304.
Servheen, K.C., 1983. Grizzly bear dens and denning activity in the
Mission and Rattlesnake Mountains, Montana. In: International
Conference on Bear Research and Management, vol. 5, pp. 201–207.
Swenson, J.E., Sandegren, F., Brunberg, S., Wabakken, P., 1997.
Winter den abandonment by brown bears (Ursus arctos), causes
and consequences. Wildlife Biology 3, 35–38.
Swenson, J.E., Gerstl, N., Dahle, B., Zedrosser, A., 2000. Draft Action
Plan for Conservation of the Brown Bear in Europe. Large
Carnivore Initiative for Europe. Available from http//www.large-
Tietje, W.D., Ruﬀ, R.L., 1980. Denning behavior of black bears in
boreal forest of Alberta. Journal of Wildlife Management 44 (4),
Vroom, G.W., Herrero, S., Ogilvie, R.T., 1980. The ecology of winter
den sites of grizzly bears in Banﬀ National Park, Alberta. In:
International Conference on Bear Research and Management, vol.
4, pp. 321–330.
136 W. Petram et al. / Biological Conservation 119 (2004) 129–136