Using distance sampling with camera traps to estimate the density of group-living and solitary mountain ungulates

Page created by Duane Contreras
 
CONTINUE READING
Using distance sampling with camera traps to estimate the density of group-living and solitary mountain ungulates
Using distance sampling with camera traps to
            estimate the density of group-living and solitary
            mountain ungulates
                                                                   RANJANA PAL, TAPAJIT BHATTACHARYA, QAMAR QURESHI
                                                                     S T E P H E N T . B U C K L A N D and S A M B A N D A M S A T H Y A K U M A R

            Abstract Throughout the Himalaya, mountain ungulates                                           Introduction
            are threatened by hunting for meat and body parts, habitat
            loss, and competition with livestock. Accurate population
            estimates are important for conservation management but                                        U       ngulates are an integral component of Himalayan
                                                                                                                   mammalian fauna and play an essential role in shaping
                                                                                                           ecosystems by influencing vegetation structure (McNaughton,
            most of the available methods to estimate ungulate densities
            are difficult to implement in mountainous terrain. Here,                                       ; Bagchi & Ritchie, ) and as primary prey for large
            we tested the efficacy of the recent extension of the point                                    predators (Bagchi & Mishra, ; Sathyakumar et al., a).
            transect method, using camera traps for estimating density                                     Population estimates are important for effective conservation
            of two mountain ungulates: the group-living Himalayan                                          management (Singh & Milner-Gulland, ; Suryawanshi
            blue sheep or bharal Pseudois nayaur and the solitary                                          et al., ). Methods to estimate animal abundance include
            Himalayan musk deer Moschus leucogaster. We deployed                                           distance sampling (Buckland et al., ), track count (Sulk-
            camera traps in – for the bharal (summer:  loca-                                     ava & Liukko, ), dung count (Laing et al., ), the
            tions; winter: ) in the trans-Himalayan region (,–                                       abundance induced heterogeneity model (Royle & Nichols,
            , m) and in – for the musk deer (summer:                                           ), repeated count (Royle, ) and the double observer
             locations; winter: ) in subalpine habitats (,–                                        method (Forsyth & Hickling, ; Suryawanshi et al., ;
            , m) in the Upper Bhagirathi basin, Uttarakhand,                                           Suryawanshi et al., ). In mountains, however, rugged
            India. Using distance sampling with camera traps, we esti-                                     and steep terrain, inaccessibility and harsh weather conditions
            mated the bharal population to be . ± SE . individuals/                                   make these techniques less effective (Singh & Milner-Gulland,
            km (CV = .) in summer and . ± SE . individuals/                                       ).
            km (CV = .) in winter. For musk deer, the estimated                                            As a consequence, several studies on mountain ungulates
            density was . ± SE . individuals/km (CV = .) in                                        have used an indirect index of abundance (e.g. Schaller et al.,
            summer and . ± SE . individuals/km (CV = .) in                                        ; Sathyakumar, ; Bagchi & Mishra, ; McCarthy
            winter. The high variability in these estimates is probably                                    et al., ; Suryawanshi et al., ) as an alternative to ab-
            a result of the topography of the landscape and the biology                                    solute abundance. However, these estimates are less reliable
            of the species. We discuss the potential application of dis-                                   and highly dependent on the assumption of constant de-
            tance sampling with camera traps to estimate the density                                       tection probability throughout the survey period (Yoccoz
            of mountain ungulates in remote and rugged terrain, and                                        et al., ). In addition, small population sizes, cryptic and
            the limitations of this method.                                                                elusive behaviour, and patchy distribution of Himalayan
                                                                                                           ungulates limit the number of observations that can be
            Keywords Bharal, camera trapping, density estimates,                                           made for a given survey effort (Singh & Milner-Gulland,
            musk deer, point transect method, subalpine, trans-                                            ). Forest-dwelling mountain ungulates may have activity
            Himalaya, Upper Bhagirathi basin                                                               peaks at night (Cavallini, ; Bhattacharya et al., a)
            Supplementary material for this article is available at                                        and are rarely detected during day-time surveys.
            doi.org/./SX                                                                  Distance sampling is one of the most popular methods
                                                                                                           for assessing the density of large herbivores in tropical
                                                                                                           forests (Buckland et al., ). However, meeting the un-
            SAMBANDAM SATHYAKUMAR (Corresponding author,         orcid.org/0000-0003-                      derlying assumptions of this method in the mountains is
            2027-4706), RANJANA PAL ( orcid.org/0000-0002-6011-104X) and QAMAR                             difficult (Corlatti et al., ), which can lead to underesti-
            QURESHI Wildlife Institute of India, Chandrabani, Dehradun, Uttarakhand
            248001, India. E-mail ssk@wii.gov.in                                                           mation of population sizes. In the mountains, non-random
            TAPAJIT BHATTACHARYA ( orcid.org/0000-0002-1154-4033) Durgapur Government
                                                                                                           locations of non-linear transects, inaccurate measurements
            College, Durgapur, India                                                                       of sighting distance and angle, and elusive behaviour of tar-
            STEPHEN T. BUCKLAND ( orcid.org/0000-0002-9939-709X) The Centre for                            get species violate the assumptions underlying conventional
            Research into Ecological and Environmental Modelling, University of St                         distance sampling (O’Neill, ; Singh & Milner-Gulland,
            Andrews, St Andrews, UK
                                                                                                           ). Furthermore, the structure of mountainous terrain
            Received  January . Revision requested  May .                                      can hamper animal detectability, as animals hidden behind
            Accepted  July .

            This is an Open Access article, distributed under the terms of the Creative Commons Attribution licence (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted re-use,
Downloadeddistribution,
            from https://www.cambridge.org/core.     IP address:
                        and reproduction in any medium,  provided 94.175.106.18, onis31properly
                                                                  the original work      May 2021   at 09:26:36, subject to the Cambridge Core terms of use, available at
                                                                                                cited.
https://www.cambridge.org/core/terms. https://doi.org/10.1017/S003060532000071X
            Oryx, Page 1 of 9 © The Author(s), 2021. Published by Cambridge University Presson behalf of Fauna & Flora International doi:10.1017/S003060532000071X
Using distance sampling with camera traps to estimate the density of group-living and solitary mountain ungulates
2         R. Pal et al.

                                                                                                                                        FIG. 1 Study area in the
                                                                                                                                        trans-Himalayan region and
                                                                                                                                        subalpine region of the Upper
                                                                                                                                        Bhagirathi basin, with the locations
                                                                                                                                        of camera traps used for estimating
                                                                                                                                        densities of the bharal Pseudois
                                                                                                                                        nayaur, and Himalayan musk deer
                                                                                                                                        Moschus leucogaster, respectively.
                                                                                                                                        The inset map shows the location of
                                                                                                                                        the Bhagirathi basin in Uttarakhand
                                                                                                                                        State, Western Himalaya, India.

             rocks or in valleys could remain undetected, irrespective of                            blue sheep or bharal Pseudois nayaur and the Himalayan musk
             their distance from the observer.                                                       deer Moschus leucogaster, both of which are affected by
                 Camera traps are an efficient tool for detecting elusive                            anthropogenic impacts (Mishra et al., ; Bhattacharya
             and rare species in remote habitats (Burton et al., ;                               & Sathyakumar, ). The bharal is a social species of
             Rovero & Zimmermann, ), and extending the point                                     the Caprinae subfamily. It is associated with alpine and
             transect method to accommodate data from camera traps                                   steppe mountain pastures, and subalpine slopes devoid
             could help solve some of the issues related to the violation                            of tree cover (,–, m; Prater, ; Sathyakumar
             of the assumptions underlying classic distance sampling                                 & Bhatnagar, ). The species is categorized as Least
             (Howe et al., ). Distance sampling with camera traps                                Concern on the IUCN Red List and listed in Schedule I of
             has recently been tested for estimating the populations                                 the Indian Wildlife (Protection) Act, . In contrast, the
             of Maxwell’s duiker Philantomba maxwellii (Howe et al.,                                 Himalayan musk deer is solitary and sedentary, remain-
             ) and the western chimpanzee Pan troglodytes versus                                 ing within a defined home range throughout the year. The
             (Cappelle et al., ) in Côte d’Ivoire. However, the efficacy                         musk deer, a primitive deer-like ruminant, is a member of
             of this technique in mountainous terrain has yet to be tested.                          the family Moschidae. In the Indian Himalayan region, the
                 Mountain ungulates are threatened by hunting for meat                               southern side of the Greater Himalaya, it is restricted to areas
             and body parts (Sathyakumar et al., a,b), habitat loss                              between , m and the treeline (Green, ; Sathyakumar
             (Namgail et al., ; Kittur et al., ) and competition                             et al., b). It is categorized as Endangered on the IUCN
             with livestock (Mishra et al., ; Bhattacharya et al., b)                        Red List (Timmins & Duckworth, ) and listed in the
             throughout the Himalaya. Here, we focused on the Himalayan                              Indian Wildlife (Protection) Act, , in Schedule I.

                                   Oryx, Page 2 of 9 © The Author(s), 2021. Published by Cambridge University Presson behalf of Fauna & Flora International doi:10.1017/S003060532000071X
Downloaded from https://www.cambridge.org/core. IP address: 94.175.106.18, on 31 May 2021 at 09:26:36, subject to the Cambridge Core terms of use, available at
https://www.cambridge.org/core/terms. https://doi.org/10.1017/S003060532000071X
Using distance sampling with camera traps to estimate the density of group-living and solitary mountain ungulates
Distance sampling with camera traps         3

                                                                                                                                          PLATE 1 The study was conducted
                                                                                                                                          in the trans-Himalayan part
                                                                                                                                          (Nilang valley) of Gangotri
                                                                                                                                          National Park characterized by
                                                                                                                                          dry alpine scrub vegetation, broken
                                                                                                                                          terrain, deep gorges, high gradient
                                                                                                                                          slopes, and narrow valleys (a), and
                                                                                                                                          in the subalpine portion of the Park
                                                                                                                                          and Uttarkashi Forest Division (b)
                                                                                                                                          within Uttarakhand State, India.

               Here, we tested the efficacy of the extension of the                                series, Cuddeback, De Pere, USA) to capture the bharal at
           distance sampling method to accommodate camera-trap                                      locations during summer (May–September ; ,
           data for estimating the density of the group-living bharal                              trap nights) and  locations during winter (October –
           in the trans-Himalayan region (,–, m) and the                                   January ; , trap nights) in the trans-Himalayan re-
           solitary Himalayan musk deer in the subalpine region                                    gion of Gangotri National Park (Fig. ). For the Himalayan
           (,–, m) of the Upper Bhagirathi basin. We exam-                                 musk deer, we set up camera traps in subalpine habitat, at
           ined the field applicability and possible limitations of this                            locations during summer (May–September ; ,
           method for estimating the density of these two ungulates                                trap nights) and  locations during winter (October –
           in mountainous terrain.                                                                 January ;  trap nights). Camera traps were mounted
                                                                                                   – cm above the ground, and programmed to trigger
           Study area                                                                              immediately and record an image followed by a -s video
                                                                                                   when movement was detected.
           We carried out this study in the trans-Himalayan part                                      Distance analysis requires calculating distance of the tar-
           (Nilang valley) of Gangotri National Park, in the subalpine                             get species from the observer, in our case from the camera.
           portion of the Park and Uttarkashi Forest Division within                               To estimate the distance of photo-captured individuals from
           Uttarakhand State, India (Fig , Plate ). Nilang valley is                             the camera trap, we calibrated image measurements against
           characterized by broken terrain, deep gorges, steep slopes                              actual measurements during camera installation. For this,
           (. °) and narrow valleys (Bhardwaj et al., ). The study                           we took measurements using a calibration pole of known
           area does not have permanent human settlements, but the                                 height at known distances from the camera, in the centre
           alpine and subalpine zones are seasonal grazing ground                                  and along both sides of the camera’s field of view. This cali-
           for livestock from lower parts of the Bhagirathi basin.                                 bration was done for a total of  camera traps, and we consid-
           Tourists also use the area in summer (June–September).                                  ered the measurements taken at these locations representative
           Nilang valley forms the international boundary with Tibet                               for others with similar topography and field of view.
           and is controlled by military personnel. There is a network
           of snow-fed tributaries of the Jadh Ganga, which drains the
                                                                                                   Data analysis
           area to meet the Bhagirathi River. We surveyed areas of dry
           alpine scrub vegetation at ,–, m for the bharal, and                            Test for sampling bias To test whether we had sampled all
           subalpine habitats dominated by Betula utilis, Pinus wallichi-                          elevations and topographic features according to their avail-
           ana, Quercus semecarpifolia and Cedrus deodara, at ,–                               ability in the landscape, we compared elevation, ruggedness,
           , m, for the Himalayan musk deer.                                                   slope and aspect of camera-trap locations and  randomly
                                                                                                   generated points using a non-parametric Mann–Whitney
           Methods                                                                                 U test for scale variables (elevation, slope, ruggedness) and
                                                                                                   Bonferroni confidence intervals for the categorical variable
           Data collection                                                                         (aspect). We resampled elevation data from the Shuttle
           We divided the study area into grid cells of  ×  km. In each                          Radar Topography Mission, at  km resolution (Jarvis et al.,
           grid cell, we generated  random points using a sampling                               ). We calculated slope, ruggedness and aspect informa-
           tool in ArcGIS . (Esri, Redlands, USA), and selected                                 tion using spatial analyst from the Digital Elevation Model
           accessible points for the placement of camera traps. Some                               in ArcGIS.
           randomly generated locations were inaccessible because of
           precipitous terrain or the presence of seasonal pastoral                                Availability for detection With camera traps, we can only
           nomad camps. We deployed camera traps (Cuddeback blue                                   estimate the density of populations that are available for

           Oryx, Page 3 of 9 © The Author(s), 2021. Published by Cambridge University Presson behalf of Fauna & Flora International doi:10.1017/S003060532000071X
Downloaded from https://www.cambridge.org/core. IP address: 94.175.106.18, on 31 May 2021 at 09:26:36, subject to the Cambridge Core terms of use, available at
https://www.cambridge.org/core/terms. https://doi.org/10.1017/S003060532000071X
Using distance sampling with camera traps to estimate the density of group-living and solitary mountain ungulates
4         R. Pal et al.

                                                                                                                                FIG. 2 Kernel density estimates of daily
                                                                                                                                activity pattern of the bharal and the musk
                                                                                                                                deer in summer and winter in the Upper
                                                                                                                                Bhagirathi basin.

             detection (Howe et al., ). If the population surveyed is                            to correct the naïve density estimate by dividing it by propor-
             not available for detection during the data collection pe-                              tion of time active, using Distance . (Thomas et al., ).
             riod selected for analysis, temporal sampling effort is over-
             estimated, and as a result, density could be underestimated
             (Cappelle et al., ). To avoid this bias, either the sam-                            Density estimation Distance sampling with camera traps
             pling period should be defined as the time during which                                 requires calculating the distance between the animal and the
             the entire population was available for detection (peak ac-                             camera at snapshot moments to ensure that animal movement
             tivity period) or the proportion of time when animals were                              does not bias the distribution of detection distances (Howe
             available for detection should be included as a parameter in                            et al., ). We thus defined a finite set of snapshot moments
             the model (Howe et al., ). In our study, the bharal was                             ( s apart) within the sampling period (as suggested in Howe
             active during .–., without a marked peak in activity                             et al., ). For each snapshot moment when the species
             (Fig. ). The Himalayan musk deer was active at night                                   was captured, we estimated the radial distance between each
             (.–.) in summer and during the day (.–.)                                  animal and the camera trap, using a regression equation
             in winter (Fig. ). We used the active period of each species                           developed from the field calibration. In this equation, the de-
             as the sampling period for the analysis. We corrected for                               pendent variable was the ratio of the actual height of an indi-
             the bias caused by animals being unavailable for detection                              vidual to its height in the photograph, and the explanatory
             by calculating the mean proportion of animals that were                                 variable was the distance at which the individual was photo-
             active during the period selected for analysis and incor-                               captured (see Supplementary Material  for details). We ob-
             porating this proportion in the density estimates. For                                  tained information on actual heights for different age and
             example, for the bharal we first plotted the number of                                  sex classes of the bharal by comparing the camera-trap photos
             independent captures (i.e. at least a -minute interval                                of the species with the height of the calibration pole height. We
             between subsequent captures) to visualize the activity pat-                             identified eight, , two and  comparable photographs of
             tern of the species (Fig. ). We assumed that if all animals                            adult males, adult females, subadults and fawns, respectively.
             were active throughout the day, then the curve would be a                               We calculated the mean height as . ± SE . cm (adult
             flat line between . and .. On the other hand, if all                             male), . ± SE . cm (adult female) and . ± SE . cm
             the animals were active around . (at the highest point                              (subadult) and . ± SE . cm (fawn). For adult Himalayan
             of the curve), then this flat line will coincide with the curve                         musk deer we used a mean height of  cm (Sathyakumar
             at .. We calculated both the areas under the imaginary                              et al., b) to estimate their distance from the camera.
             flat line and under the actual activity curve shown in Fig. .                            Density was estimated following the equation for
             We then calculated the mean proportion of animals that                                  camera-trap point transects (Howe et al., ):
             are active between . and . by dividing the propor-
             tion of the area under the actual activity curve by the area                                                                  
                                                                                                                                           K
                                                                                                                                               nk
             under the imaginary flat line, where animal activity reaches                                                              k=1                     1
                                                                                                                              D̂ =                         ×
             a peak. The estimated mean proportion of animals that are                                                                   
                                                                                                                                         K                     A
             active during the period selected for analysis was .                                                                 pw2          ek P̂k
                                                                                                                                            k=1
             in summer and . in winter for the bharal. For the
             Himalayan musk deer it was . in summer and . in                                    where nk is the number of observations of animals at a point
             winter. We used the proportion of time animals are active                               k (camera-trap location), ek is the temporal effort, and P̂k is

                                   Oryx, Page 4 of 9 © The Author(s), 2021. Published by Cambridge University Presson behalf of Fauna & Flora International doi:10.1017/S003060532000071X
Downloaded from https://www.cambridge.org/core. IP address: 94.175.106.18, on 31 May 2021 at 09:26:36, subject to the Cambridge Core terms of use, available at
https://www.cambridge.org/core/terms. https://doi.org/10.1017/S003060532000071X
Distance sampling with camera traps   5

           the estimated probability of obtaining an image of an animal                            points (Supplementary Fig. ). Encounter rates were highly
           that is within θ degrees (angle covered by the camera’s field                           variable amongst locations and did not show any spatial
           of view), K is the total number of camera-trap locations and                            autocorrelation for the bharal (Moran’s I P: . in summer,
           w (truncation distance) in front of the camera at a snapshot                            . in winter) or the musk deer (Moran’s I P: . in summer,
           of the moment. The effort at a point k was measured as                                  . in winter).
           ek = θ Tk/ πt where θ/π describes the fraction of a circle
           covered by a camera, Tk is the period of camera deployment
           (in seconds), and t is the unit of time used to determine a                             Density estimates
           finite set of snapshot moments within Tk (also in seconds).                             The bharal was photo-captured by  out of  camera
           We defined the period of camera deployment as the time the                              traps deployed in summer, and  out of  in winter.
           target species was expected to be active during the sampling                            We obtained , snapshots in  videos in summer
           period. For the bharal, this was a -hour period per day                               and  snapshots in  videos in winter. In summer,
           (.–.) in both seasons and for the Himalyan musk                                  one of the cameras contributed a large number of captures
           deer a -hour period per day (.–.) in summer                                    (c. % of the total dataset). This particular camera was
           and a -hour period per day (.–.) in winter. /A                                placed on a steep slope with cliffs on both sides, and
           is the availability correction factor.                                                  close (– m) to an intensively used bharal trail along a
               Seven camera traps malfunctioned because of technical
                                                                                                   stream. Consequently, a large number of observations by
           errors and were not included in the final analysis. For the
                                                                                                   this camera were within – m as most of the bharals fol-
           analysis in Distance, we modelled the detection from using
                                                                                                   lowed the path to move up or down the slope. Because of
           the same functions as Howe et al. (): half normal with ,
                                                                                                   this bias, the initially estimated density of . ± SE .
            or  Hermite polynomial adjustment terms; hazard rate
                                                                                                   individuals/km had a high CV (.). We removed
           with , , or  cosine adjustments; uniform with  or  cosine
                                                                                                   this camera from the final analysis to get an estimate
           adjustments. As model selection methods based on Akaike’s
                                                                                                   with reduced bias. Amongst the summer captures, we
           information criterion (AIC) tend to favour overly complex
                                                                                                   found an excess of distances close to the camera (Fig. ).
           models because of overdispersion in the data, we selected
                                                                                                   The hazard-rate model is more sensitive than the half-
           models using a recently proposed two-step procedure
                                                                                                   normal model to this excess, resulting in an implausible
           (Howe et al., ): () Firstly, the best model is selected
                                                                                                   rapid fall-off in the detection probability. Therefore, we
           on the basis of AIC adjusted for overdispersion (QAIC)
                                                                                                   used the second-best model (Table ), the half-normal
           within each key function, where the overdispersion param-
                                                                                                   model, for estimating bharal density in summer (. ± SE
           eter (Ĉ) is calculated from the ratio between the χ statistics
                                                                                                   . individuals/km, CV = .). In winter, the best model
           of the most parameterized model for each key function and
                                                                                                   was the hazard-rate model, and the second-best half-
           its degrees of freedom (χ/df). () Secondly, the best model
                                                                                                   normal model resulted in the same density estimates
           is selected with the smallest values of the χ goodness-of-fit
                                                                                                   (. ± SE . individuals/km, CV = .; Table ).
           statistic divided by its degrees of freedom (across QAIC-
                                                                                                       Himalayan musk deer were captured by  out of 
           selected models, one from each key function). We used
                                                                                                   cameras in summer and  out of  cameras in winter. We
           the point transect distance sampling method in Distance
                                                                                                   obtained  snapshots in  videos in summer and 
           (Thomas et al., ) for all analyses.
                                                                                                   snapshots in  videos in winter. Himalayan musk deer
                                                                                                   data did not show the heterogeneity in capture probabil-
           Results                                                                                 ities amongst cameras that we observed for the bharal,
                                                                                                   nor any evidence of bias in terms of distances (Fig. ). The
           Sampling bias test                                                                      best model was the hazard-rate model with cosine adjust-
                                                                                                   ment in both seasons (Table ), and estimated density was
           In case of the Himalayan musk deer, the elevation, rugged-                              . ± SE . individuals/km (CV = .) in summer and .
           ness and slope of sampled camera-trap locations were not                                 ± SE . individuals/km (CV = .) in winter (Table ).
           biased: the mean values of sampled locations were not sig-
           nificantly different from the mean elevation, ruggedness
           and slope of  random points in both seasons (P . .                                Discussion
           in each case, Mann–Whitney U test; Supplementary
           Fig. ). Bonferroni confidence intervals indicated no par-                              Our estimates of bharal density in summer (. ± SE .
           ticular aspect category was preferred for sampling in winter                            individuals/km) and winter (. ± SE . individuals/km)
           or summer (Supplementary Fig. ). Similarly, in the case of                             were similar. Estimates of bharal densities from three differ-
           the bharal, the ruggedness, elevation and slope of the sam-                             ent locations in Spiti, the nearest trans-Himalayan landscape
           pled camera locations were not different from the mean                                  (using standardized double observer method; Suryawanshi
           ruggedness, elevation, slope and aspect of the random                                   et al., ) were ., . and ./km. Estimates of bharal

           Oryx, Page 5 of 9 © The Author(s), 2021. Published by Cambridge University Presson behalf of Fauna & Flora International doi:10.1017/S003060532000071X
Downloaded from https://www.cambridge.org/core. IP address: 94.175.106.18, on 31 May 2021 at 09:26:36, subject to the Cambridge Core terms of use, available at
https://www.cambridge.org/core/terms. https://doi.org/10.1017/S003060532000071X
6         R. Pal et al.

                                                                                                                                    FIG. 3 Detection probability and
                                                                                                                                    probability density for the models
                                                                                                                                    selected for estimating density. The bars
                                                                                                                                    show the data distribution, and the line
                                                                                                                                    represents the model fit. The heights of
                                                                                                                                    the bars are scaled so that they cover the
                                                                                                                                    same total area as the area under the
                                                                                                                                    line, to show how well the detection
                                                                                                                                    function fits the data.

             densities in our study are low compared to those of Spiti can                           coverage of the entire survey area in a short period, a require-
             be expected because of differences in habitat type and topog-                           ment that could not be fulfilled in our study because parts of
             raphy. The Spiti landscape comprises vast trans-Himalayan                               the study area were inaccessible and visual coverage was
             meadows (Biotic Province B; Rodgers et al., ), whereas                             insufficient (Plate ).
             our study area consists primarily of narrow valleys and gorges                             The density of Himalayan musk deer was higher in sum-
             with rough terrain and barren slopes (Biotic Province C;                               mer (. ± SE . individuals/km) than winter (. ± SE .
             Kumar et al., ; Plate ). The differences of bharal density                         individuals/km). The analysis of seasonal habitat use in the
             estimates and mean group sizes (. in Nilang valley vs .                            study area also showed a trend of decline in captures at high
             in Kibber, Spiti) between these two areas may thus be a result                          elevations (Pal et al., ) in winter, possibly because musk
             of differences in habitat quality. In addition, Nilang valley                           deer migrate to lower altitudes during periods of heavy
             is affected by anthropogenic disturbances such as hunting                               snowfall. Similar seasonal movements were also observed
             (Bhardwaj et al., ), livestock grazing (Chandola, ;                             in other areas (Anwar & Minhas, ; Dendup & Lham,
             RP pers. obs., ) and presence of free-ranging dogs (Pal                             ). Other studies using the silent drive count meth-
             et al., ). The differences could also be caused by different                        od in Kedarnath Wildlife Sanctuary estimated musk deer
             survey techniques. We were unable to estimate bharal density                            density to be . ± SE . individuals/km in –
             using the point count distance method because of insufficient                           (Sathyakumar, ), . individuals/km in –
             observations. The double observer method used to estimate                               (Sathyakumar & Malik, ) and . individuals/km in
             bharal density in trans-Himalayan habitat requires visual                                (S. Sathyakumar, unpubl. data). These studies may have

                                   Oryx, Page 6 of 9 © The Author(s), 2021. Published by Cambridge University Presson behalf of Fauna & Flora International doi:10.1017/S003060532000071X
Downloaded from https://www.cambridge.org/core. IP address: 94.175.106.18, on 31 May 2021 at 09:26:36, subject to the Cambridge Core terms of use, available at
https://www.cambridge.org/core/terms. https://doi.org/10.1017/S003060532000071X
Distance sampling with camera traps    7

           TABLE 1 Details of the top three models used to estimate the densities of the bharal Pseudois nayaur and the Himalayan musk deer Moschus
           leucogaster in summer and winter in the Upper Bhagirathi basin, Uttarakhand, India, showing key functions (defining parametric shapes
           for the detection function), adjustment types (to allow for departures from the parametric shape), the number of adjustment terms selected
           (order), overdispersion factor (Ĉ), Akaike’s information criterion adjusted for overdispersion (QAIC), and density estimates with standard
           error (SE) and coefficient of variance (CV).

           Key function                            Adjustment type                       Order              Ĉ                  QAIC                  Estimate ± SE   CV
           Bharal (summer)
           Hazard                                                                        0                    2.93             1,083.87              1.61 ± 0.60     0.38
           Half normal                             Hermite polynomial                    2                   29.23               118.05              0.51 ± 0.10     0.31
           Uniform                                 Cosine                                1                  123.98                33.06              0.16 ± 0.05     0.30
           Bharal (winter)
           Hazard                                  Cosine                                1                    5.51               626.89              0.64 ± 0.20     0.37
           Half normal                             Hermite polynomial                    2                    8.71               400.63              0.64 ± 0.20     0.37
           Uniform                                 Cosine                                2                   31.24               118.95              0.35 ± 0.10     0.36
           Musk deer (summer)
           Hazard                                  Cosine                                1                      6.30             287.50              0.42 ± 0.10     0.34
           Half normal                                                                   0                      8.12             381.97              0.29 ± 0.10     0.34
           Uniform                                 Cosine                                1                      9.63             279.80              0.26 ± 0.10     0.34
           Musk deer (winter)
           Hazard                                  Cosine                                1                      4.28             172.03              0.10 ± 0.05     0.48
           Half normal                                                                   0                      4.54             119.78              0.10 ± 0.05     0.47
           Uniform                                 Cosine                                1                      4.67             142.58              0.08 ± 0.03     0.46

           overestimated musk deer densities as the drive count meth-                              individuals recorded at shorter distances. However, such
           od is known for overestimating the density of animals                                   incidents were relatively rare in our study (six occasions).
           (Takeshita et al., ). In addition, they were carried out                            Herd behaviour also affects captures, as bharals tend to
           in a small portion (c. . km) of a protected area; small                              follow the first individual when moving together. Because
           study areas combined with a bias towards good habitat qual-                             we analysed individual distances from the camera, this can
           ity can result in highly overestimated densities (Suryawanshi                           cause heaping in the distances recorded (Fig. ).
           et al., ).                                                                              Distance sampling with camera traps requires setting the
               Our density estimates are associated with high coefficients                         cameras in burst or video mode. Our effort to implement
           of variation. This high variability is probably caused by land-                         this method in the Greater Himalayan alpine habitats failed
           scape topography and species biology. The fit of the model                              because cameras were continuously triggered by grass
           for the solitary Himalayan musk deer was better than for                                movements in the field of view (RP, pers. obs., ). We
           the group-living bharal. Here, we discuss some of the issues                            had to discard data from four camera traps in this study
           we faced using sampling with camera traps, and make sugges-                             for the same reason. Mounting cameras higher off the
           tions as to how these can be addressed in future studies.                               ground could help minimize this problem. In addition,
               For the bharal, the main problem that caused bias in the                            the imprecise (high CV) estimates suggest that more sam-
           distances at which individuals were captured was the inad-                              pling locations are required to improve precision (Howe
           equate camera view because of slopes. The ruggedness of the                             et al., ; Cappelle et al., ).
           landscape also influences the approach angle and the dis-                                   The ability of camera sensors to detect moving animals
           tance covered by the cameras: those on hilltops or at the base                          may vary depending on camera type and placement, tem-
           of a hill covered distances of – m, whereas cameras on                              perature, and humidity (Hofmeester et al., ). Different
           hill slopes covered distances of – m (depending on the                               camera models can be tested at a site to assess the ability
           slope). Topographic variability probably also influenced                                to detect animals. There could be inconsistencies between
           detection probability and the estimated angle of the cam-                               the theoretical and actual angle of view θ, which can lead
           era view. Future studies in similar landscapes could use                                to biased estimates. This can result in underestimates if
           statistical tests to examine the effects of these parameters                            sensors are less sensitive to movements near the edges of the
           more thoroughly.                                                                        camera’s field of view (i.e. the effective angle can be smaller
               Another issue encountered with the group-living bharal                              than the assumed angle). This can be addressed with field
           was that animals grazing close to the camera blocked the                                tests to estimate the effective angle θ, which can then be ac-
           view of animals that were further away. This can make it                                counted for in the analysis. Imprecise measures of distance
           impossible to calculate the distance from the camera for                                should not be an issue if they are appropriately binned in
           individuals in the background, leading to a bias towards                                distances for the analysis (Buckland et al., ). However,

           Oryx, Page 7 of 9 © The Author(s), 2021. Published by Cambridge University Presson behalf of Fauna & Flora International doi:10.1017/S003060532000071X
Downloaded from https://www.cambridge.org/core. IP address: 94.175.106.18, on 31 May 2021 at 09:26:36, subject to the Cambridge Core terms of use, available at
https://www.cambridge.org/core/terms. https://doi.org/10.1017/S003060532000071X
8         R. Pal et al.

             imprecise estimates of the target species’ availability for                             Author contributions Conception of study: SS, RP, TB, QQ; data
             detection can cause erroneous estimates. Care should thus                               collection: RP; data analysis: RP, STB, TB; writing: all authors; revi-
                                                                                                     sions: QQ, STB, SS.
             be taken in selecting the appropriate time period of animal
             activity and availability for detection (Howe et al., ).                            Conflicts of interest None.
                Distance sampling with camera traps has paved the way
             for a new analytical approach to estimate the abundance of                              Ethical standards This work was carried out with permission from
             both group-living and solitary mountain ungulates in rug-                               Uttarakhand Forest Department (Letter no. 836/5-6) and abided by
             ged and inaccessible terrains of the Himalaya. It can to                                the Oryx guidelines on ethical standards.
             some extent overcome the logistic constraints associated
             with rugged terrain and harsh weather that affect other
             methods such as point counts and transects sampling.                                    References
             These traditional methods are difficult to implement effec-
                                                                                                     ANWAR, M. & M I N H A S , R.A. () Distribution and population status
             tively in a high-altitude, rugged and remote landscape be-                                 of Himalayan musk deer (Moschus chrysogaster) in the Machiara
             cause they require cover of most vantage points in a single                                National Park, AJ and K. Pakistan Journal of Zoology, , –.
             day or within a defined, short period of time (Singh &                                  B A G C H I , S. & M I S H R A , C. () Living with large carnivores:
             Milner-Gulland, ). In addition, the low number of de-                                  predation on livestock by the snow leopard (Uncia uncia).
             tections often limits the conventional analytical process.                                 Journal of Zoology, , –.
                                                                                                     B A G C H I , S. & R I T C H I E , M.E. () Herbivore effects on above- and
             Indirect observations such as dung counts can be useful                                    belowground plant production and soil nitrogen availability in the
             but require estimation of the decay rate, which is often dif-                              trans-Himalayan shrub-steppes. Oecologia, , –.
             ficult to obtain (Buckland, ; Kuehl et al., ). Distance                         B H A R D WA J , M., U N I YA L , V.P. & S A N YA L , A.K. () Estimating
             sampling with camera traps can work for longer periods in                                  relative abundance and habitat use of Himalayan blue sheep
             the field and may help to overcome the challenges presented                                Pseudois Nayaur in Gangotri National Park. Galemys: Boletín
                                                                                                        Informativo de la Sociedad Española para la Conservación y Estudio
             by low numbers of detections and observer bias (Cappelle                                   de los Mamíferos, , –.
             et al., ). An important advantage of camera traps over                              B H AT TA C H A R YA , T. & S A T H Y A K U M A R , S. () Natural resource use
             conventional distance sampling is that they are better suited                              by humans and response of wild ungulates. Mountain Research and
             to monitor solitary, elusive and nocturnal species such as the                             Development, , –.
             Himalayan musk deer. Camera traps have been extensively                                 B H AT TA C H A R YA , T., B A S H I R , T., P O U DYA L , K., S AT H YA K U M A R , S. &
                                                                                                        S A H A , G.K. (a) Distribution, occupancy and activity patterns
             used to survey the snow leopard Panthera uncia. A slight
                                                                                                        of goral (Nemorhaedus goral) and serow (Capricornis thar) in
             modification in the sampling design (modified camera                                       Khangchendzonga Biosphere Reserve, Sikkim, India. Mammal
             placement) could help gain information on its main prey                                    Study, , –.
             species, including the bharal, the ibex Capra sibirica, the                             B H AT TA C H A R YA , T., K I T T U R , S., S AT H YA K U M A R , S. & R AWAT , G.S.
             argali Ovis ammon and the musk deer.                                                       (b) Diet overlap between wild ungulates and domestic livestock
                                                                                                        in the Greater Himalaya: implications for management of grazing
                Despite these advantages, there are limitations to the use of
                                                                                                        practices. Proceedings of the Zoological Society, , –.
             camera traps, including the high cost of the cameras and the                            B U C K L A N D , S.T., A N D E R S O N , D.R., B U R N H A M , K.P., L A A K E , J.L.,
             extensive time required to process photographs and videos.                                 B O R C H E R S , D.L. & T H O M A S , L. () Introduction to Distance
             Substantial numbers of camera traps would be required to                                   Sampling: Estimating Abundance of Biological Populations.
             improve the precision of density estimates derived from                                    Oxford University Press, Oxford, UK.
             distance sampling with camera traps (Cappelle et al., ).                            B U C K L A N D , S.T., R E X S T A D , E.A., M A R Q U E S , T.A. & O E D E KO V E N , C.S.
                                                                                                        () Distance Sampling: Methods and Applications. Springer
             Despite the high initial cost, we believe this approach could                              International Publishing, Cham, Switzerland.
             help improve abundance estimations for both group living                                B U R T O N , A.C., N E I L S O N , E., M O R E I R A , D., L A D L E , A., S T E E N W E G , R.,
             and solitary mountain ungulates in rough, mountainous ter-                                 F I S H E R , J.T. et al. () Wildlife camera trapping: a review and
             rain where conventional techniques cannot be implemented.                                  recommendations for linking surveys to ecological processes.
                                                                                                        Journal of Applied Ecology, , –.
             Acknowledgements This work is part of a project initiated un-                           C A P P E L L E , N., D E S P R É S -E I N S P E N N E R , M., H O W E , E.J., B O E S C H , C. &
             der the National Mission for Sustaining the Himalayan Ecosystem                            K Ü H L , H.S. () Validating camera trap distance sampling for
             (NMSHE) Programme funded by the Department of Science and                                  chimpanzees. American Journal of Primatology, , e.
             Technology, Government of India (grant no.: DST/SPLICE/CCP/                             C AVA L L I N I , P. () Survey of the goral Nemorhaedus goral
             NMSHE/TF-2/WII/2014[G]). The Miriam Rothschild Travel Bursary                              (Hardwicke) in Himachal Pradesh. Journal of Bombay Natural
             Programme provided funding for a 4-week internship for R. Pal with S.                      History Society, , –.
             T. Buckland at St Andrews University, UK. We thank the Director and                     C H A N D O L A , S. () Vegetational inventory of cold desert habitat of
             Dean of the Wildlife Institute of India for their guidance and support;                    Nilang area of Jadh Ganga catchment (Uttarkashi) in Garhwal
             D.V.S. Khati, Principal Chief Conservator of Forests and Chief                             Himalaya. PhD thesis, Hemvati Nandan Bahuguna Garhwal
             Wildlife Warden, Uttarakhand, for granting research permission;                            University, Uttarakhand, India.
             Sandeep Kumar, Divisional Forest Officer and former Deputy                              C O R L AT T I , L., F A T T O R I N I , L. & N E L L I , L. () The use of block
             Director, Gangotri National Park, and Shrawan Kumar for their sup-                         counts, mark–resight and distance sampling to estimate population
             port and cooperation; and L. Corlatti for reviewing the manuscript.                        size of a mountain-dwelling ungulate. Population Ecology, , –.

                                   Oryx, Page 8 of 9 © The Author(s), 2021. Published by Cambridge University Presson behalf of Fauna & Flora International doi:10.1017/S003060532000071X
Downloaded from https://www.cambridge.org/core. IP address: 94.175.106.18, on 31 May 2021 at 09:26:36, subject to the Cambridge Core terms of use, available at
https://www.cambridge.org/core/terms. https://doi.org/10.1017/S003060532000071X
Distance sampling with camera traps                          9

           D E N D U P , P. & L H A M , C. () Winter distribution and poaching of                       R O Y L E , J.A. () N-mixture models for estimating population size
               musk deer, Moschus chrysogaster and Moschus leucogaster in Jigme                                 from spatially replicated counts. Biometrics, , –.
               Dorji National Park, Bhutan. International Journal of Conservation                           R O Y L E , J.A. & N I C H O L S , J.D. () Estimating abundance from
               Science, , –.                                                                             repeated presence–absence data or point counts. Ecology, , –.
           F O R S Y T H , D.M. & H I C K L I N G , G.J. () An improved technique for                   S AT H YA K U M A R , S. () Habitat ecology of major ungulates in
               indexing abundance of Himalayan thar. New Zealand Journal of                                     Kedarnath Musk Deer Sanctuary. PhD thesis, Saurashtra University,
               Ecology, , –.                                                                             Rajkot, Gujrat, India.
           G R E E N , M.J.B. () Aspects of the ecology of the Himalayan musk                           S AT H YA K U M A R , S. & B H AT N A G A R , Y.V. () Mountain Ungulates
               deer. PhD thesis, University of Cambridge, Cambridge, UK.                                        ENVIS Bulletin: Wildlife and Protected Areas. Wildlife Institute of
           H O F M E E S T E R , T.R., R O W C L I F F E , J.M. & J A N S E N , P.A. () A simple            India, Dehradun, India.
               method for estimating the effective detection distance of camera                             SATHYAKUMAR, S., B H AT T AC H A R Y A , T., B A S H I R , T. & P O U DY A L , K.
               traps. Remote Sensing in Ecology and Conservation, , –.                                     (a) Developing Spatial Database on the Mammal
           H O W E , E.J., B U C K L A N D , S.T., D E S P R É S -E I N S P E N N E R , M.L. & K Ü H L ,        Distributions and Monitoring Programme for Large Carnivores,
               H.S. () Distance sampling with camera traps. Methods in                                      Prey Populations, and their Habitats in Khangchendzonga
               Ecology and Evolution, , –.                                                             Biosphere Reserve. Project report, Wildlife Institute of India,
           H O W E , E.J., B U C K L A N D , S.T., D E S P R É S -E I N S P E N N E R , M.L. & K Ü H L ,        Dehradun, India.
               H.S. () Model selection with overdispersed distance sampling                             S AT H YA K U M A R , S., G O P A L , S.R. & J O H N S I N G H , A.J.T. (b) Musk
               data. Methods in Ecology and Evolution, , –.                                               deer. In Mammals of South Asia. Volume  (eds A.J.T. Johnsingh &
           J A R V I S , A., G U E VA R A , E., R E U T E R , H.I. & N E L S O N , A.D. ()                  N. Manjrekar), pp. –. Universities Press, Hyderabad, India.
               Hole-Filled SRTM for the Globe, Version . CGIAR-CSI SRTM                                    S AT H YA K U M A R , S. & M A L I K , P.K. (). Release of Captive
               m Database, CGIAR Consortium for Spatial Information.                                          Himalayan Musk Deer in Kedarnath Wildlife Sanctuary, Western
               srtm.csi.cgiar.org [accessed August ].                                                       Himalaya. Final report, Wildlife Institute of India, Dehradun, India.
           K I T T U R , S., S AT H YA K U M A R , S. & R AWAT , G.S. () Assessment of                  S C H A L L E R , G.B., J U N R A N G , R. & M I N G J I A N G , Q. () Status of the
               spatial and habitat use overlap between Himalayan tahr and                                       snow leopard Panthera Uncia in Qinghai and Gansu Provinces,
               livestock in Kedarnath Wildlife Sanctuary, India. European                                       China. Biological Conservation, , –.
               Journal of Wildlife Research, , –.                                                   S I N G H , N.J. & M I L N E R -G U L L A N D , E.J. () Monitoring ungulates
           K U E H L , H.S., T O D D , A., B O E S C H , C. & W A L S H , P.D. ()                           in Central Asia: current constraints and future potential. Oryx,
               Manipulating decay time for efficient large-mammal density                                       , –.
               estimation: gorillas and dung height. Ecological Applications,                               S U L K AVA , R.T. & L I U K KO , U.M. () Use of snow-tracking methods
               , –.                                                                                   to estimate the abundance of otter (Lutra lutra) in Finland with
           K U M A R , A., A D H I K A R I , B.S. & R AWAT , G.S. () Biogeographic                          evaluation of one-visit census for monitoring purposes. Annales
               delineation of the Indian trans-Himalaya: need for revision.                                     Zoologici Fennici, , –.
               Current Science, , –.                                                             S U R Y AWA N S H I , K.R., B H AT N A G A R , Y.V. & M I S H R A , C. () Why
           L A I N G , S.E., B U C K L A N D , S.T., B U R N , R.W., L A M B I E , D. & A M P H L E T T ,       should a grazer browse? Livestock impact on winter resource use
               A. () Dung and nest surveys: estimating decay rates. Journal                                 by bharal Pseudois nayaur. Oecologia, , –.
               of Applied Ecology, , –.                                                           S U R Y AWA N S H I , K.R., B H AT N A G A R , Y.V. & M I S H R A , C. ()
           M C C A R T H Y , K.P., F U L L E R , T.K., M I N G , M., M C C A R T H Y , T.M.,                    Standardizing the double-observer survey method for estimating
               W A I T S , L. & J U M A B A E V , K. () Assessing estimators of snow                        mountain ungulate prey of the Endangered snow leopard. Oecologia,
               leopard abundance. Journal of Wildlife Management, , –.                                , –.
           M C N A U G H T O N , S.J. () Grazing as an optimization process: grass–                     S U R Y AWA N S H I , K.R., K H A N YA R I , M., S H A R M A , K., L K H A G VA J AV , P. &
               ungulate relationships in the Serengeti. The American Naturalist,                                M I S H R A , C. () Sampling bias in snow leopard population
               , –.                                                                                    estimation studies. Population Ecology, , –.
           M I S H R A , C., VA N W I E R E N , S.E., K E T N E R , P., H E I T KO N I G , I.M.A. &         S U R Y AWA N S H I , K.R., M U D A P P A , D., K H A N Y A R I , M., R A M A N , T.R.S.,
               P R I N S , H.H.T. () Competition between domestic livestock and                             R A T H O R E , D., K U M A R , M.A. & P AT E L , J. () Population
               wild bharal Pseudois nayaur in the Indian trans-Himalaya. Journal                                assessment of the Endangered Nilgiri tahr Nilgiritragus hylocrius
               of Applied Ecology, , –.                                                                 in the Anamalai Tiger Reserve, using the double-observer survey
           N A M G A I L , T., F O X , J.L. & B H AT N A G A R , Y.V. () Habitat shift and                  method. Oryx, , –.
               time budget of the Tibetan argali: the influence of livestock grazing.                       T A K E S H I T A , K., I K E D A , T., T A K A H A S H I , H., Y O S H I D A , T., I G O TA , H.,
               Ecological Research, , –.                                                                  M AT S U U R A , Y. & K A J I , K. () Comparison of drive counts and
           O’N E I L L , H. () Designing robust ranger based monitoring                                     mark-resight as methods of population size estimation of highly
               strategies for the saiga antelope Saiga tatarica tatarica. PhD thesis,                           dense sika deer (Cervus Nippon) populations. PLOS ONE,
               Imperial College London, London, UK.                                                             , e.
           P A L , R., T H A K U R , S., A R Y A , S., B H AT TA C H A R YA , T. & S AT H Y A K U M A R ,   T H O M A S , L., B U C K L A N D , S.T., R E X S TA D , E.A., L A A K E , J.L.,
               S. () Mammals of the Bhagirathi basin, Western Himalaya:                                     S T R I N D B E R G , S., H E D L E Y , S.L. et al. () Distance software: design
               understanding distribution along spatial gradients of habitats and                               and analysis of distance sampling surveys for estimating population
               disturbances. Oryx, published online  July .                                               size. Journal of Applied Ecology, , –.
           P R A T E R , S.H. () The Book of Indian Mammals. Bombay Natural                             T I M M I N S , R.J. & D U C K W O R T H , J.W. () Moschus leucogaster. In
               History Society, Mumbai, India.                                                                  The IUCN Red List of Threatened Species : e.TA.
           R O D G E R S , W.A., P A N WA R , H.S. & M AT H U R , V.B. () Wildlife                          dx.doi.org/./IUCN.UK.-.RLTS.TA.en
               Protected Area Network in India: A Review. Executive summary,                                    [accessed  June ].
               st edition. Wildlife Institute of India, Dehradun, India.                                   Y O C C O Z , N.G., N I C H O L S , J.D. & B O U L I N I E R , T. () Monitoring of
           R O V E R O , F. & Z I M M E R M A N N , F. () Camera Trapping for Wildlife                      biological diversity in space and time. Trends in Ecology &
               Research. Pelagic Publishing Ltd, Exeter, UK.                                                    Evolution, , –.

           Oryx, Page 9 of 9 © The Author(s), 2021. Published by Cambridge University Presson behalf of Fauna & Flora International doi:10.1017/S003060532000071X
Downloaded from https://www.cambridge.org/core. IP address: 94.175.106.18, on 31 May 2021 at 09:26:36, subject to the Cambridge Core terms of use, available at
https://www.cambridge.org/core/terms. https://doi.org/10.1017/S003060532000071X
You can also read