Identifying regions for conservation of sloth bears through occupancy
modelling in north-eastern Karnataka, India
Sayantan Das1, Saurav Dutta2, Sharmi Sen3, Jijumon A. S.2, S. Babu4, Honnavalli Nagraj Kumara4,
and Mewa Singh1,5,6
1
Biopsychology Laboratory and Institution of Excellence, University of Mysore, Manasagangotri, Mysore-570006,
Karnataka, India
2
Indian Institute of Science Education and Research-Kolkata, Mohanpur Campus, Post: BCKV Main Office, Mohanpur,
Nadia-741252, West Bengal, India
3
Indian Institute of Science Education and Research-Mohali, Knowledge City, Sector 81, SAS Nagar, Manauli-140306,
Punjab, India
4
Sálim Ali Centre for Ornithology and Natural History, Anaikatty, Coimbatore-641108, Tamil Nadu, India
5
Organismal Biology Unit, Jawaharlal Nehru Centre for Advanced Scientific Research, Bangalore, Karnataka, India
Abstract: In the absence of information on species in decline with contracting ranges,
management should emphasize remaining populations and protection of their habitats.
Threatened by anthropogenic pressure including habitat degradation and loss, sloth bears
(Melursus ursinus) in India have become limited in range, habitat, and population size. We
identified ecological and anthropogenic determinants of occurrence within an occupancy
framework to evaluate habitat suitability of non-protected regions (with sloth bears) in
northeastern Karnataka, India. We employed a systematic sampling methodology to yield
presence–absence data to examine a priori hypotheses of determinants that affected occupancy.
These covariates were broadly classified as habitat or anthropogenic factors. Mean number of
termite mounds and trees positively influenced sloth bear occupancy, and grazing pressure
expounded by mean number of livestock dung affected it negatively. Also, mean percentage of
shrub coverage had no impact on bear inhabitance. The best fitting model further predicted
habitats in Bukkasagara, Agoli, and Benakal reserved forests to have 38%, 75%, and 88%,
respectively, of their sampled grid cells with high occupancies (.0.70) albeit little or no legal
protection. We recommend a conservation strategy that includes protection of vegetation standstructure, maintenance of soil moisture, and enrichment of habitat for the long-term welfare of
this species.
Key words: conservation, distribution, ecology, India, Karnataka, Melursus ursinus, occupancy modelling,
sloth bears
DOI: 10.2192/URSUS-D-14-00008.1
Ursus 25(2):111–120 (2014)
Conservation and management policies and practices directed at animal species should ideally be
developed from complete information on demographic characteristics, behavioral patterns, lifehistory traits, and habitat use (Margules and Pressey
2000). Habitat use by a species, and the factors that
influence habitat use, should in turn be evaluated
using rigorous statistical and modelling techniques.
With recent advances in geo-spatial tools and
modelling techniques, predicting species occurrence,
distribution, and abundance has become common
(e.g., Scott et al. 2001, Pearce and Boyce 2006, Best
et al. 2007, Krishna et al. 2008, Boyd and Foody
2011, Nandy et al. 2012). However, occurrence or
abundance of species might not be a veritable output
of the relationship between a species and its habitat,
because occurrence and abundance can be an
imperfect surrogate for demographic performance
(Van Horne 1983, Hobbs and Hanley 1990, Tyre
et al. 2001). Therefore, patterns of occupancy along
with occupancy correlates can only determine
habitat suitability and its capacity to support
populations. This functional link between a species’
occupancy and the environmental factors influencing
it is critical for understanding habitat processes that
can be used to guide management (Franklin et al.
6
email: [email protected]
111
112
CONSERVATION OF SLOTH BEARS N Das et al.
2000, Breininger and Carter 2003). This exercise is
even more important for generalist species that occur
primarily outside protected areas and are involved in
conflicts with humans, including human casualties.
The sloth bear (Melursus ursinus) in India serves as
a model species under these criteria. Because of
resource extraction, habitat fragmentation, habitat
loss, and high conflict across their range in drier
regions in central (Johnsingh 1986, Servheen 1990,
Chauhan et al. 1999, Rajpurohit and Krausman
2000, Bargali 2004), western, and southeastern
(Krishna Raju et al. 1987) India, Garshelis et al.
(1999b:234–235, 2008) appealed for a concerted
action to ‘‘expand and update’’ distribution maps
of the species in ‘‘relation to forest cover and
boundaries of protected areas.’’ Such an action is
imperative ‘‘to delineate discrete population units’’
coupled with information on ‘‘land use and land
conditions’’ in regions outside the protected area
network to assess their potential for supporting
viable populations (Garshelis et al. 2008).
The sloth bear is endemic to the Indian subcontinent (Erdbrink 1953, Sathyakumar et al. 2012),
with a historical distribution from the foothills of the
Himalayas in northern India to the dry slopes of the
Western Ghats in the south (Bargali et al. 2004).
However, sloth bear populations are currently
limited to 5 regions in India: northern, northeastern,
central, southeastern, and southwestern populations
(Garshelis at al. 1999b, Johnsingh 2003, Yoganand
et al. 2006, Sathyakumar et al. 2012). This drastic
range contraction has rendered the species Vulnerable to Extinction (IUCN 2013) and led to its
inclusion in Schedule I of the Indian Wildlife
(Protection) Act as amended in 2003 (GOI 1972,
2003). Currently, only about 10% of the species’
current distribution in India contains high-quality
habitat (Yoganand et al. 2006). Techniques used to
assess sloth bear–habitat relationships included assessment of home-range and habitat-selection analysis using radiotelemetry (Joshi et al. 1995, 1999;
Ratnayeke et al. 2007), den-site examination, and
indirect bear signs such as droppings, tracks, and claw
marks (Akhtar et al. 2004, Chauhan et al. 2004).
These studies have thus revealed only qualitative
habitat characteristics essential for the species.
Sloth bears appear to use different habitats to
meet their life requisites (e.g., foraging, cover;
Wrangham and Rubenstein 1986). Therefore, ecological studies should characterize each habitat to
develop a thorough understanding of requirements
of the species with respect to feeding, hiding, and
other requisites. This information is crucial to
identify suitable habitats for conservation and
development of management plans. Our objective
was to estimate occupancy of sloth bears and assess
the potential effects of environmental and anthropogenic determinants such that we can establish
potential suitability of habitats and demarcate them
for long-term conservation of the species.
Study area
We conducted the study in 6 sites in the revenue
districts of Hospet and Koppal, specifically Daroji
Wildlife Sanctuary (WLS; 15u169N, 76u369E; 5,083 ha),
Bukkasagara Reserve Forest (RF; 15u199N, 76u339E;
3,169 ha), Daroji RF (15u149N, 76u409E; 465 ha),
Toranagallu RF (15u129N, 76u419E; 547 ha), Agoli
RF (15u289N, 76u239E; 2,796 ha), and Benakal RF
(15u269N, 76u239E; 3,864 ha; Fig. 1). We selected these
areas based on their close proximity to each other
and unpublished reports of sloth bear presence and
incidences of conflicts with humans. Only Daroji WLS
and a non-demarcated 2,685.50 ha of Bukkasagara
RF is ‘‘legally protected’’ for wildlife, whereas the
others are reserved forests located close to villages and
are open to resource extraction, including timber
harvest, sand-mining, and livestock grazing (Kiran
2011). Daroji WLS is separated from Bukkasagara
RF by a road and village settlements along its
northern boundary but is contiguous with Daroji
RF and Toranagallu RF along its southeastern
border. Six kilometers northwest of Bukkasagara RF
is Agoli RF and Benakal RF, which are divided by a
state highway. The habitat is largely rocky and
bouldery; dry deciduous scrub and southern thorn
forests (Champion and Seth 1968) are the only forest
types, with Acacia spp. as the dominant species. The
vegetation in this region exists in a degraded state,
though afforestation programs have ensured prominent vegetation in certain pockets. The study area has a
semi-arid climate characterized by hot summers (34u–
45uC) during April–June and low rainfall (571.92 mm)
from June to November.
Methods
Survey design
We determined sampling units by overlaying the
study sites with 2-km2 grid cells as per the Forest
Department’s survey estimates of the existence of 1
Ursus 25(2):111–120 (2014)
CONSERVATION OF SLOTH BEARS N Das et al.
113
Fig. 1. The location of the study sites, Daroji Wildlife Sanctuary, Bukkasagara Reserve Forest (RF), Daroji RF,
Toranagallu RF, Agoli RF, and Benakal RF, and pictorial depiction of survey design for estimation of sloth bear
occupancy are shown. Map showing (a) the location of study region in India, (b) the six study sites with their
approximate geographical spacing overlaid with a 2-km2 grid, and (c) sampling units (strips) represented as
‘‘boxes’’ along the diagonals of each grid cell. The study was carried out in two phases, firstly during June–
July 2011 and secondly, June–July 2012.
bear/1.5 km2 in Daroji WLS (Kiran 2011). Diagonal
of a square grid cell was designated as ‘‘1’’ if it ran
from the direction of north-west to the south-east
and ‘‘2’’ if otherwise. We inscribed spatial replicates
in the form of strips of 500-m length and 20-m width
along the 2 diagonals of the square grid at intervals
of 150 m (Fig. 1). To avoid duplication of data, we
ascribed the covariates to the third strip of diagonal
1 when strip 3 of diagonal 1 coincided with strip 3 of
diagonal 2. Limitation of manpower and logistics,
and crepuscular activity cycle of the species, justified
the usage of spatial replicates instead of temporal
replicates in a single-season survey (Krishna et al.
2008, Saracco et al. 2011). Thus, Daroji WLS,
Bukkasagara RF, Daroji RF, Toranagallu RF,
Agoli RF, and Benakal RF bore 44, 30, 9, 9, 37,
Ursus 25(2):111–120 (2014)
and 30 grid cells, respectively (Fig. 1 and Table 1).
Additionally, we selected grid cells with a minimum
spatial replicate of 3 for modelling because the
detection probability of indirect signs of bear was
‘‘low’’ otherwise.
Identification of covariates
We selected measurable habitat covariates representative of food, vegetation structure, and covariates indicative of past and present anthropogenic
disturbances. Sloth bears are myrmecophagous (Pocock 1933, Erdbrink 1953, Gopal 1991, Sacco and
Valkenburgh 2006), predating especially on ants
(Formicidae) and termites (Isoptera; Laurie and
Seidensticker 1977, Gopal 1991, Joshi et al. 1997)
along with other animal matter (Bargali et al. 2004)
114
CONSERVATION OF SLOTH BEARS N Das et al.
Table 1. Outline of grid-cell–level information included in the analysis and details of sloth bear droppings and/
or tracks recovered from each survey site of Daroji Wildlife Sanctuary (WLS), Bukkasagara Reserve Forest
(RF), Daroji RF, and Toranagallu RF (India) during June–July 2011, and Agoli RF and Benakal RF (India) during
June–July 2012.
Study area
Daroji WLS
Bukkasagara RF
Daroji RF
Toranagallu RF
Agoli RF
Benakal RF
Percent of grid cells included
in occupancy analysis (%)
Total no. of
grid cells
No. of grid cells with animal
signs
66
43
00
00
65
57
44
30
09
09
37
30
18
08
00
00
21
17
and berries (Gopal 1991). Therefore, food-resource
determinants included number of termite mounds
and feeding trees, and habitat factors included
number of trees and percentage of shrub coverage.
We visually estimated shrub coverage based on the
proportion of area covered by shrubs in a single
strip. Record of feeding (fruiting or non-fruiting)
trees was combined with record of non-feeding trees
to form the covariate ‘‘trees’’ (with diam at breast
ht [DBH] .5 cm), which represented both food
resources and habitat characteristics. Anthropogenic
pressure in the form of grazing and timber harvest
was also quantified. We quantified grazing by
calculating the number of livestock (i.e., cattle and
sheep) -dung present, and extraction of firewood
through enumeration of stumps and loppings
(chopped branches and twigs) from trees of DBH
.5 cm, within the strips. This resulted in the
identification of 5 site covariates presumed to
influence occupancy of sloth bears. We used shrubcoverage, the only covariate that could influence
detection of indirect signs, to model detection
probability.
Field methods
We conducted single surveys in Daroji WLS,
Bukkasagara RF, Daroji RF, and Toranagallu RF
during June–July 2011 and in Agoli RF and Benakal
RF during June–July 2012. We selected cells in grids
for sampling that included .25% of the cell area
because grid cells at the periphery usually enclosed
smaller area. In cases where undisturbed habitat
occurred beyond forest boundaries within a grid cell,
we extended the sampling effort to strips within the
grid cell but beyond the study sites. However,
sampling was limited by inaccessibility to strips
owing to their location on steep ridges or on
sprawling boulders. Such strips were treated as
missing data points or observations and were
accommodated within the model likelihood framework proposed by Mackenzie et al. (2002).
The sloth bear is largely nocturnal (Laurie and
Seidensticker 1977, Joshi et al. 1995, Gopalaswamy
2006) or crepuscular throughout the study regions
because the vegetation cover is sparse and daytime
temperature remains high throughout the year
(Akhtar et al. 2004). Therefore, presence of the
species had to be established by indirect means (i.e.,
through droppings or tracks; Garshelis et al. 1999a).
A minimum of 2 observers surveyed strips on foot
using a hand-held Global Positioning System device
(Garmin eTrex30 and Garmin HCx Vista, Kansas
City, Kansas, USA) and recorded bear droppings or
tracks, number of termite mounds (Te), percentage
shrub coverage (SC), number of trees (T), number of
stumps and/or logs (SL), and number of livestock
dung (LD) within each strip. Droppings of bears
were identified by their characteristic cylindrical
shape with tapering ends and ascertained through
close examination of content, which remained dominated by either insects or seeds (Joshi et al. 1997,
Sreekumar and Balakrishnan 2002, Bargali et al.
2004, Ramesh et al. 2009). Sloth bear tracks were
identified based on its characteristic specular print
with the larger toe pointing outward and prominent
claw marks (Karanth et al. 2002). Consequently, a
site (grid cell) was deemed occupied if a bear
dropping or track was detected in .1 of the 8 spatial
replicates (naı̈ve estimate of occupancy). These
detection–non-detection data in each strip formed
the basis for construction of detection histories for
each grid cell. For strips that could not be sampled,
we treated the replicate as a missing observation
(MacKenzie et al. 2002). Overall, 60% of the grid
cells were sampled. We first averaged each covariate
over 8 strips and secondly, z-transformed it to
Ursus 25(2):111–120 (2014)
CONSERVATION OF SLOTH BEARS N Das et al.
Table 2. Summary of model-selection procedure for
factors affecting detection probability of bear
droppings when data from the study regions of
Daroji Wildlife Sanctuary, Bukkasagara Reserve
Forest (RF), Agoli RF, and Benakal RFs (India) were
cumulatively evaluated. The study regions were
surveyed in two phases, firstly from June to July
2011 and secondly from June to July 2012.
Modela
y(.),p(SC)
y(.),p(.)
b
p^
AICcc
wid
Ke
b1f
0.456
0.465
709.79
714.57
0.916
0.084
3
2
0.25
0.00
a
SC—shrub coverage of the plot.
b^
p—estimated species detection probability.
c
AICc—AIC corrected for small-sample bias.
d
wi—AICc model wt.
e
K—no. of parameters estimated by the model.
f
b1—beta co-efficient of the sampling covariate.
rescale and normalize the data, prior to occupancy
analysis.
Occupancy estimation
We constructed detection histories and estimated
^ ), and
the probability that a grid cell was occupied (y
the detection probability ( ^p), using maximum likelihood functions (MacKenzie et al. 2002) in Program
PRESENCE (ver. 3.0; MacKenzie et al. 2006). Data
from all the study regions were pooled into a single
occupancy analysis because of the small area of each
study region (,5,100 ha). We employed logistic
models with logit link and binomial error to evaluate
the effect of covariates on model parameters, occu^ ) and detection probability ( ^p). Based
pancy (y
on current knowledge of sloth bear ecology, we
speculated that covariates indexing food and vegetation structure would positively affect occupancy,
whereas covariates indexing anthropogenic pressures
would negatively affect it. We used a step-wise
approach to first model effects of covariates on
detection (p), and then modelled y. Among the
covariates we considered, only undergrowth–shrubcoverage could affect the detection of bear droppings
or tracks and was therefore used to model p[y(.),
p(SC)] (Table 2). Both spatial and numerical pair-wise
correlation analyses using Pearson correlation analysis did not identify autocorrelation (21 , r , 1, p .
0.05) among covariates, and thus all model selections
were uncorrelated. Subsequently, we formulated a
candidate set of 8 a priori models (involving 1 of the 5
covariates) to investigate the influence of covariates
on occurrence. Selection of models, succeeded by
computation of model weights and averaging of
Ursus 25(2):111–120 (2014)
115
parameters, were executed as per the framework of
Burnham and Anderson (1998). We tabulated models
in ascending order of their relative quality based on
Akaike Information Criterion values adjusted for
sample size (AICc; Burnham and Anderson 1998). We
report the estimates of both detection probability and
the occupancy as mean 6 standard error.
Results
We did not detect sloth bear sign at Daroji RF and
Toranagallu RF and excluded these sites from
analyses. Therefore, only Daroji WLS, Bukkasagara
RF, Agoli RF, and Benakal RF were considered,
which contributed 29, 13, 24, and 17 grid cells,
respectively, to the occupancy model. In the 83 grid
cells where sloth bear sign was observed, the number
of detections varied from 1/strip to 27/strip. The
highest number of cumulative bear signs per grid cell
were recorded from Benakal RF (8.4 signs/cell),
closely followed by Agoli RF (6.35 signs/cell) while
Daroji WLS (1.59 signs/cell) and Bukkasagara RF
(0.67 signs/cell) had lower recovery of animal signs
per grid cell (Table 1).
The estimated average detection probability ( ^p) of
bears in a grid cell was 0.47 6 0.003. Contrary to
expectation, detection probability was positively
influenced by shrub coverage (b1 5 0.25; Table 2);
however, this being a logical improbability, we
ignored it. As a result, detection probability p was
held constant over all spatial replicates (plots) in
subsequent models. The model-averaged occupancy
^ 5 0.79 6 0.007. The number of
estimate was y
termite mounds (Te) and trees (T) together [y(T +
Te)] had the greatest predictive power of site
occupancy (Table 3). However, each of these covariates performed poorly when modelled individually.
The second-best model also included percentage
shrub cover; however, the DAICc value (1.65)
suggests that percentage shrub cover did not have
a substantial effect on occupancy. The covariates,
trees (b1 5 3.17, CI 5 1.93–4.41) and termite
mounds (b1 5 2.65, CI 5 1.27–4.02) had the greatest
effect on determining bear occupancy (TSwi 5 0.994,
TeSwi 5 0.985). Among the anthropogenic determinants, sloth bear occurrence was negatively correlated to livestock dung (b1 5 20.53, CI 5 20.98 to
20.08) but positively correlated to stumps and
loppings (b1 5 0.73, CI 5 0.54–1.41) albeit at low
relative values of importance (LDSwi 5 0.244, SLSwi
5 0.244), signifying a weak effect on occupancy.
116
CONSERVATION OF SLOTH BEARS N Das et al.
Table 3. Summary of model-selection procedure for all the factors and/or covariates influencing occupancy of
sloth bear when the study regions of Daroji Wildlife Sanctuary, Bukkasagara Reserve Forest (RF), Agoli RF,
and Benakal RFs (India) were cumulatively assessed. The study regions were surveyed in two phases, firstly
from June to July 2011, and secondly from June to July 2012.
Modela
T + Te
T + Te + SC
T + Te + LD + SL
T + Te + SC + LD + SL
T
Te
SC
LD + SL
y(.),p(.)
^
^b
y
(95% CI)
c
p^
0.782
0.782
0.780
0.781
0.786
0.790
0.799
0.791
0.798
0.730–0.832
0.731–0.834
0.727–0.833
0.726–0.836
0.747–0.826
0.752–0.827
0.759–0.839
0.762–0.819
0.689–0.876
0.471
0.471
0.472
0.471
0.469
0.468
0.463
0.467
0.465
d
(SE)
DAICce
wif
Kg
0.024
0.024
0.024
0.024
0.024
0.024
0.024
0.024
0.025
0.00
1.65
2.31
3.64
7.22
9.27
11.71
17.82
22.74
0.512
0.224
0.161
0.083
0.014
0.005
0.002
0.000
0.000
4
5
6
7
3
3
3
4
2
a
T—Trees; Te—Termite; SC—Shrub coverage; LD—Livestock dung; SL—Stumps and loppings.
y—estimated occupancy parameter corrected for ‘detection probability’.
p^—estimated species detection probability.
d
SE—associated standard error of occupancy estimate.
e
AICc—AIC corrected for small-sample bias; DAICc—difference in AICc values between each model and the model with the lowest
AICc.
f
wi—AICc model wt.
g
K—no. of parameters estimated by the model.
b^
c^
Overall, 4.8%, 28.9%, and 66.27% of the sampled
^ 5
grid cells (83 grid cells) were classified as low (y
^
^ 5
0.01–0.30), medium (y 5 0.31–0.70), and high (y
0.71–1.00), respectively, based on their occupancy
estimates (Fig. 2). The study region of Benakal RF
had the greatest percentage (88%) of sampled grid
cells designated as highly occupied, followed by
Agoli RF (75%), Daroji WLS (59%), and finally
Bukkasagara RF (39%).
Discussion
We provide the first occupancy estimate of sloth
bears in their dry range in Hospet and Koppal
districts of Karnataka, India. Sloth bears occurred
over a large expanse of sites in reserved forests
outside the protected region, emphasizing the need
for their protection and uninterrupted management.
We identified portions of Agoli RF, Benakal RF and
Bukkasagara RF as highly used and occupied by the
species because these consisted of a greater proportion of grid cells with high degree of occupancy than
were found in Daroji WLS. Despite earlier evidences, bears were absent in Daroji RF and Toranagallu
RF due to possible extirpation; Daroji RF is
partially submerged by Daroji reservoir (owing to
heavy sewage discharge by industries), and largescale industries (e.g., steel factories) operate adjoining to these 2 regions. We additionally propose
the usage of ‘‘occupancy’’ as an index to evaluate
effectiveness of bear management strategies across
sites and times.
Occupancy
Information on recent range, abundance, and
distribution of sloth bears was determined through
the compilation of previous literatures, unpublished
information, and state-wide questionnaires of field
managers (Sathyakumar et al. 2012). The National
Bear Conservation and Welfare Action Plan (2012)
projected the extent of the national distribution at
400,000 km2. Furthermore, almost 50% of subsisting
populations reside outside protected areas (Yoganand et al. 2006) and these populations have been
steadily declining (Garshelis et al. 1999b). Estimate
available under an occupancy framework designed
to specifically study tigers (Panthera tigris), reported
a figure of 16,582 km2 of sloth bear distribution in
Karnataka in 2010 (a declination of 19% from 2006;
Jhala et al. 2011). The only attempt to generate
occupancy-level information of sloth bears in India
^5
by Ramesh et al. (2012) yielded an occupancy of y
0.83 6 0.01SE in Mudumalai Tiger Reserve that was
beyond the moderate range of 0.20–0.80 obtained in
site-occupancy studies (Linkie et al. 2007, Mortelliti
and Boitani 2008, Sarmento et al. 2011). This
anomalous estimate was based on a detection
probability of p 5 0.23 6 0.07SE, lower than the
threshold of 0.30 recommended for an unbiased
estimate of occupancy (for no. of sites 5 40, Spatial–
Ursus 25(2):111–120 (2014)
CONSERVATION OF SLOTH BEARS N Das et al.
117
Fig. 2. Estimated sloth bear occupancy data generated for each grid, which were extracted from the bestfitting model, [y(T + Te)] when all the study sites were cumulatively modelled. Un-sampled grid cells
comprised those that were partially sampled or not sampled. Daroji Wildlife Sanctuary, Bukkasagara Reserve
Forest (RF), Daroji RF, and Toranagallu RF (India) were surveyed during June–July 2011; and Agoli RF and
Benakal RF were sampled during June–July 2012.
Temporal replicate . 5) as outlined by Mackenzie
et al. (2003). This was primarily due to larger landscape and heterogeneous vegetation composition.
Previous efforts to evaluate sloth bear distribution
in India have also ignored the issue of detectability. In
this study, bear sign was recorded in 83 out of 139
sampled grid cells, generating a naı̈ve occupancy of
0.60. Through incorporation of imperfect detection of
animals into occupancy models, the estimated proportion of area occupied by sloth bears substantially
increased to 0.79, a 24% increase over the naı̈ve
estimate.
Covariates influencing detection probability
and occurrence
The estimated detection probability of sloth bear
droppings was 0.47, which is considerably high for
Ursus 25(2):111–120 (2014)
this low-density species (e.g., Yoganand et al. 2006).
This is possibly due to either intensive search by
observers, and/or bears preferably defecating in
available open spaces. Alternatively, unlike habitat
of the southwestern population of bears, these
patches are isolated and are smaller (,321 km2;
Ramesh et al. 2012) in size, which could have led to
high detection probability. An occupancy study in
Mudumalai (Ramesh et al. 2012) showed weak
influences of termite mounds (AICtermites . AICnull)
and fruiting trees (AICfruiting trees . AICnull) in the
occupancy models, undermining their importance.
This can be explained by diverse availability of food
resources in Mudumalai (Ramesh et al. 2009). On
the contrary, the current study regions remained
resource-limited during the surveyed season, resulting in enhanced dependence on the only available
118
CONSERVATION OF SLOTH BEARS N Das et al.
resources—termite mounds and possibly, feeding
trees. Therefore, the only covariates that co-affected
inhabitation–occupancy by the species, as predicted
by the a priori hypothesis were (1) Tree and (2)
Termite mounds. Trees, although not distinguished
between feeding and non-feeding, could be a strong
indicator of both hiding shelters and food resources
in general. Conclusively, bears used only those areas
of the forests that were secure in terms of food
resources and safety. This finding mirrors reports of
sloth bears using patches of Lantana shrub in Panna
National Park in India (Yoganand et al. 2013) and
areas with thick vegetation cover in Royal Chitwan
National Park in Nepal (Laurie and Seidensticker
1977). In congruence to our a priori hypothesis
and the findings of Ratnayeke et al. (2007), grazing
pressure had a negative influence on occupancy, but
lopping influenced it positively (though weakly).
Usage of lopped sites by bears is explained by the
severe limitation of space because most of these
habitats are small and exist as disconnected islands.
However, occupancy analysis was limited by our
inability to include or record ‘‘presence–distance
from den–cave’’ as a possible covariate; we feel that
foraging explorations and hence site selection–
settlement depend on proximity to den site, as
predicted by Multiple Central Place Foraging
Theory (Chapman et al. 1989) and/or the energymaximizing strategy hypothesis (Hall 1962, Schoener
1971). Similar studies on sloth bears in the future
should explore the role of presence–absence of ‘‘den
site’’ on site occupancy.
Recommendation for conservation of
the species
Management intervention for the species’ protection and conservation should include the following:
Firstly, Agoli RF, Benakal RF, and Bukkasagara RF
require immediate management plans that ensure
their strict protection. Secondly, removal of trees,
including the ones not fed on by bears, should be
curtailed immediately. Thirdly, because the sloth bear
is the umbrella species in the region, habitats in
Bukkasagara RF, Agoli RF, and Benakal RF could
be enriched through planting of native tree species
(fruit-bearing or otherwise), which can serve as both
hiding and/or escape areas and food resources.
Fourthly, moisture content of the soil is to be
maintained within a range that facilitates nesting
by termites (Wood 1988). Finally, only long-term
monitoring of the population, stochasticity–dynamics
of the factors identified, and public education–
awareness can help secure the future of the species.
Acknowledgments
We thank the Karnataka Forest Department for
according permission and assistance during fieldwork. We acknowledge the support of Mr. D.
Sharma, Chief Wildlife Warden, Karnataka. Our
gratitude also to Mr. G. Hanumanthaiyya, Range
Forest Officer, Gangawati Circle, for accommodations and logistic support. Mr. Maulabhussain,
Forest Guard, Gangawati Circle, is gratefully
acknowledged for his support and assistance. SD,
SD, SS and JAS thank the Indian Academy of
Sciences, Bangalore, for their fellowship awards.
Literature cited
AKHTAR, N., H.S. BARGALI, AND N.P.S. CHAUHAN. 2004.
Sloth bear habitat use in disturbed and unprotected
areas of Madhya Pradesh, India. Ursus 15:203–211.
BARGALI, H.S. 2004. The ecology of the problematic sloth
bear (Melurusus ursinus) and mitigation of human–bear
conflict in Bilaspur forest division, Madhya Pradesh.
Thesis, Wildlife Institute of India, Dehradun, India.
———, N. AKHTAR, AND N.P.S. CHAUHAN. 2004. Feeding
ecology of sloth bears in a disturbed area in central
India. Ursus 15:212–217.
BEST, B.D., P.N. HALPIN, E. FUJIOKA, A.J. READ, S.S.
QIAN, L.J. HAZEN, AND R.S. SCHICK. 2007. Geospatial
web services within a scientific workflow: Predicting
marine mammal habitats in a dynamic environment.
Ecological Informatics 2(3):210–223.
BOYD, D.S., AND G.M. FOODY. 2011. An overview of recent
remote sensing and GIS based research in ecological
informatics. Ecological Informatics 6(1):25–36.
BREININGER, D.R., AND G.M. CARTER. 2003. Territory
quality transitions and source–sink dynamics in a
Florida scrub-jay population. Ecological Application
13:516–529.
BURNHAM, K.P., AND D.R. ANDERSON. 1998. Model
selection and inference: A practical information-theoretic approach. Springer-Verlag, New York, New York,
USA.
CHAMPION, S.H., AND S.K. SETH. 1968. A revised survey of
the forest types of India. Manager of Publications,
Government of India, New Delhi, India.
CHAPMAN, C.A., L.J. CHAPMAN, AND R.L. MCLAUGHLIN.
1989. Multiple central place foraging by spider monkeys: Travel consequences of using many sleeping sites.
Oecologia 79:506–511.
CHAUHAN, N.P.S., H.S. BARGALI, AND N. AKHTAR. 1999.
Human–sloth bear conflicts in the state of Madhya
Ursus 25(2):111–120 (2014)
CONSERVATION OF SLOTH BEARS N Das et al.
Pradesh, India. Presented in 12th international conference on bear research and management. International
Association for Bear Research and Management, 13–18
October 1999, Poiana Brasov, Romania.
———, ———, AND ———. 2004. Activity patterns of
sloth bear in fragmented and disturbed areas of
Bilaspur Forest Division, Chattisgarh, India. Presented
in 15th international conference on bear research and
management. International Association for Bear Research and Management, 8–13 February 2004, San
Diego, California, USA.
ERDBRINK, D.P. 1953. A review of fossil and recent bears of
the Old World. Volume 2. Deventer, Jan de Lange,
Netherlands.
FRANKLIN, A.B., D.R. ANDERSON, R.J. GUTIÉRREZ, AND
K.P. BURNHAM. 2000. Climate, habitat quality, and
fitness in northern spotted owl populations in northwestern California. Ecological Monographs 70:539–590.
GARSHELIS, D.L., A.R. JOSHI, AND J.L.D. SMITH. 1999a.
Estimating density and relative abundance of sloth
bears. Ursus 11:87–98.
———, ———, ———, AND C.G. RICE. 1999b. Sloth bear
conservation action plan (Melursus ursinus). Pages 225–
240 in C. Servheen, S. Herrero, and B. Peyton, editors.
Bears: Status survey and conservation action plan.
International Union for Conservation of Nature-The
World Conservation Union, Gland, Switzerland and
Cambridge, England, UK.
———, S. RATNAYEKE, AND N.P.S. CHAUHAN (INTERNATIONAL UNION FOR CONSERVATION OF NATURE SPECIES
SURVIVAL COMMISSION BEAR SPECIALIST GROUP). 2008.
Melursus ursinus. In International Union for Conservation of Nature 2013. International Union for
Conservation of Nature red list of threatened species.
Version 2013.2. www.iucnredlist.org. Accessed 11 Jun
2014.
GOPAL, R. 1991. Ethological observations on the sloth bear
(Melursus ursinus). Indian Forester 117:915–920.
GOPALASWAMY, A.M. 2006. Estimating sloth bear abundance from repeated presence–absence data in Nagarahole–Bandipur National Parks, India. Thesis, University of Florida, Gainesville, Florida, USA.
GOVERNMENT OF INDIA [GOI]. 1972. The wildlife (protection act). Natraj Publ., Dehradun, India.
———. 2003. Page 218 in The wildlife (protection act).
Natraj Publ, Dehradun, India.
HALL, K.R.L. 1962. Numerical data, maintenance activities and locomotion of the wild chacma baboon, Papio
ursinus. Journal of Zoology 139:181–220.
HOBBS, N.T., AND T.A. HANLEY. 1990. Habitat evaluation:
Do use/availability data reflect carrying capacity?
Journal of Wildlife Management 54:515–522.
INTERNATIONAL UNION FOR CONSERVATION OF NATURE
[IUCN]. 2013. IUCN red list of threatened species.
Version 2013.2. www.iucnredlist.org. Accessed 6 Jun 2014.
Ursus 25(2):111–120 (2014)
119
JHALA Y.V., Q. QURESHI, R. GOPAL, AND P.R. SINHA,
EDITORS, Page 302 in Status of the tigers, co-predators,
and prey in India, 2010. National Tiger Conservation
Authority, Government of India, New Delhi, India;
and Wildlife Institute of India, Dehradun, India.
JOHNSINGH, A.J.T. 1986. Diversity and conservation of
carnivorous mammals in India. Proceedings of the
Indian Academy of Sciences (Animal Sciences) Supplement:73–89.
———. 2003. Bear conservation in India. Journal of
Bombay the Natural History Society 100:190–201.
JOSHI, A.R., D.L. GARSHELIS, AND J.L.D. SMITH. 1995.
Home ranges of sloth bears in Nepal: Implications
for conservation. Journal of Wildlife Management 59:
204–214.
———, ———, AND ———. 1997. Seasonal and habitatrelated diets of sloth bears in Nepal. Journal of
Mammalogy 78:584–597.
———, J.L.D. SMITH, AND D.L. GARSHELIS. 1999. Sociobiology of the myrmecophagus sloth bear in Nepal.
Canadian Journal of Zoology 77:1690–1704.
KARANTH, K.U., N.S. KUMAR, AND R.S. CHUNDAWAT.
2002. Field surveys: Assessing spatial distributions of
tigers and prey. Appendix 1.1. Page 167 in K.U.
Karanth and J.D. Nichols, editors. Monitoring tigers
and their prey—A manual for researchers, managers
and conservationists in tropical Asia. Centre for
Wildlife Studies, Bangalore, India.
KIRAN, M.N. 2011. Management plan for Daroji Sloth
Bear Sanctuary. Karnataka Forest Department, Kamalapura, India.
KRISHNA, Y.C., J. KRISHNASWAMY, AND N.S. KUMAR. 2008.
Habitat factors affecting site occupancy and relative
abundance of four-horned antelope. Journal of Zoology 276:63–70.
KRISHNA RAJU, K.S.R., A.V.R.G. KRISHNA MURTHY, C.
SUBBA REDDI, N.A.V. PRASAD REDDY, R. LOKARANJAN,
AND K.J.N.G. SHANKAR. 1987. Status of wildlife and
habitat conservation in Andhra Pradesh. Journal of
Bombay Natural History Society 84:605–619.
LAURIE, A., AND J. SEIDENSTICKER. 1977. Behavioural
ecology of the sloth bear (Melursus ursinus). Journal
of Zoology 182:187–204.
LINKIE, M., Y. DINATA, A. NUGROHO, AND I.A. HAIDIR.
2007. Estimating occupancy of a data deficient mammalian species living in tropical rainforests: Sun bears
in the Kerinci Seblat region, Sumatra. Biological
Conservation 137:20–27.
MACKENZIE, D.I., J.D. NICHOLS, J.E. HINES, M.G.
KNUTSON, AND A.B. FRANKLIN. 2003. Estimating
site occupancy, colonisation and local extinction
when a species is detected imperfectly. Ecology 83:
2248–2255.
———, ———, G.B. LACHMAN, S. DROEGE, J.A. ROYLE,
AND C.A. LANGTIMM. 2002. Estimating site occupancy
120
CONSERVATION OF SLOTH BEARS N Das et al.
rates when detection probabilities are less than one.
Ecology 83:2248–2255.
———, ———, J.A. ROYLE, K.H. POLLOCK, L.L. BAILEY,
AND J.E. HINES. 2006. Occupancy estimation and
modeling: Inferring patterns and dynamics of species
occurrence. Elsevier, New York, New York, USA.
MARGULES, C.R., AND R.L. PRESSEY. 2000. Systematic
conservation planning. Nature 405:243–253.
MORTELLITI, A., AND L. BOITANI. 2008. Evaluation of scentstation surveys to monitor the distribution of three
European carnivore species (Martes foina, Meles,
Vulpes vulpes) in a fragmented landscape. Mammalian
Biology 73:287–292.
NANDY, S., S.P.S. KUSHWAHA, AND P. GAUR. 2012.
Identification of swamp deer (Cervus duvauceli Cuvier)
potential habitat in Jhilmil Jheel Conservation Reserve,
Uttarakhand, India using multi-criteria analysis. Environmental Management 49:902–914.
PEARCE, J.L., AND M.S. BOYCE. 2006. Modelling distribution and abundance with presence-only data. Journal of
Applied Ecology 43:405–412.
POCOCK, R.I. 1933. The black and brown bears of Europe
and Asia. Part II. The sloth bear (Melursus), the
Himalayan black bear (Selenarctos) and the Malayan
bear (Helarctos). Journal of Bombay Natural History
Society 36:101–138.
RAJPUROHIT, K.S., AND P.R. KRAUSMAN. 2000. Human–
sloth bears conflicts in Madhya Pradesh, India. Wildlife
Society Bulletin 28:393–399.
RAMESH, T., R. KALLE, K. SANKAR, AND Q. QURESHI. 2012.
Factors affecting habitat patch use by sloth bears in
Mudumalai Tiger Reserve, Western Ghats, India.
Ursus 23:78–85.
———, K. SANKAR, AND Q. QURESHI. 2009. Additional notes
on the diet of sloth bear (Melursus ursinus) in Mudumalai
Tiger Reserve as shown by scat analysis. Journal of
Bombay Natural History Society 106:204–206.
RATNAYEKE, S., V.F.T. MANEN, AND U.K.G.K. PADMALAL.
2007. Home-range and habitat use of sloth bears (M.
ursinus) in Wasgomuwa National Park, Sri Lanka.
Ursus 13:272–284.
SACCO, T., AND V.B. VALKENBURGH. 2006. Ecomorphological indicators of feeding behaviour in the bears
(Canivora:Ursidae). Journal of Zoology 263:41–54.
SARACCO, J.F., R.B. SIEGEL, AND R.L. WILKERSON. 2011.
Occupancy modeling of black-backed woodpeckers on
burned Sierra Nevada forests. Ecosphere 2:art31.
SARMENTO, P.B., J. CRUZ, C. EIRA, AND C. FONSECA. 2011.
Modelling the occupancy of sympatric carnivorans in a
Mediterranean ecosystem. European Journal of Wildlife Research 57:119–131.
SATHYAKUMAR, S., R. KAUL, N.V.K. ASHRAF, A. MOOKERJEE, AND V. MENON. 2012. National bear conservation
and welfare action plan. Ministry of Environment and
Forests, Wildlife Institute of India and Wildlife Trust of
India, Delhi, India.
SCHOENER, T.W. 1971. Theory of feeding strategies.
Annual Review of Ecological Systems 2:369–404.
SCOTT, J.M., P.J. HEGLUND, M.L. MORRISON, J.B. HAUFLER, M.G. RAPHAEL, W.A. WALL, AND F.B. SAMSON.
2001. Introduction. Pages 1–5 in J.M. Scott, P.J.
Heglund, M.L. Morrison, J.B. Haufler, M.G. Raphael,
W.A. Wall, and F.B. Samson, editors. Predicting
species occurrences: Issues of accuracy and scale. Island
Press, Washington, DC, USA.
SERVHEEN, C. 1990. The status and conservation of the
bears of the world. International Conference on Bear
Research and Management Monograph 2.
SREEKUMAR, P.G., AND M. BALAKRISHNAN. 2002. Seed
dispersal by the sloth bear (Melursus ursinus) in South
India. Biotropica 34:474–477.
TYRE, A.J., H.P. POSSINGHAM, AND D.B. LINDENMAYER.
2001. Matching observed pattern with model process:
Can territory occupancy provide information about
life history parameters. Ecological Application 11:
1722–1737.
VAN HORNE, B. 1983. Density as a misleading indicator of
habitat quality. Journal of Wildlife Management 47:
893–901.
WOOD, T.G. 1988. Termites and the soil environment.
Biology and Fertility of Soils 6:228–236.
WRANGHAM, R.W., AND D.I. RUBENSTEIN. 1986. Social
evolution in birds and mammals: Ecological aspects of
social evolution. Princeton University Press, Princeton,
New Jersey, USA.
YOGANAND, K., C.G. RICE, AND A.J.T. JOHNSINGH. 2006. Is
the sloth bear in India secure? A preliminary report on
distribution, threats and conservation requirements.
Journal of Bombay Natural History Society 103:
172–181.
———, ———, AND ———. 2013. Sloth bear.
Pages 439–456 in A.J.T. Johnsingh and N. Manjrekar,
editors. Mammals of South Asia. Volume I. Universities Press (India) Private Ltd., Himayatnagar, Hyderabad, India.
Received: 2 February 2014
Accepted: 1 September 2014
Associate Editor: S. Sambandam
Ursus 25(2):111–120 (2014)