European Journal of Wildlife Research

, Volume 57, Issue 3, pp 603–613

Noninvasive genetic monitoring of tiger (Panthera tigris tigris) population of Orang National Park in the Brahmaputra floodplain, Assam, India

Authors

    • Aaranyak
  • Rumi Dev Barman
    • Aaranyak
  • Chatrapati Das
    • Aaranyak
  • Ajit Basumatary
    • Aaranyak
  • Anjan Talukdar
    • Aaranyak
  • M. Firoz Ahmed
    • Aaranyak
  • Bibhab Kumar Talukdar
    • Aaranyak
  • Rupjyoti Bharali
    • Department of BiotechnologyGauhati University
Original Paper

DOI: 10.1007/s10344-010-0471-0

Cite this article as:
Borthakur, U., Barman, R.D., Das, C. et al. Eur J Wildl Res (2011) 57: 603. doi:10.1007/s10344-010-0471-0
  • 293 Views

Abstract

The Brahmaputra Valley of Assam, India, is one of the prime habitats for the endangered Royal Bengal tiger Panthera tigris tigris. With dwindling global population, estimation of the minimum number of tigers has always been a curiosity to wildlife researchers as well as to protected area managers. In the present study, DNA-based techniques were used for identifying individual tigers present in Orang National Park of Assam, from 57 faecal samples collected during February 2009. Orang National Park stands as an island of a single forest patch along the north bank of river Brahmaputra. The present study confirms the presence of 17 individual tigers in Orang National Park, with five male and 12 female. DNA-based capture–recapture analysis yielded minimum range estimate of 18 and 19 individuals, with possible overestimates of population size following two models of capture probability in CAPWIRE. The results of our genetic counting of tigers are compared with the estimates of 19 tigers based on pugmark analysis by the state Forest Department in 2000 and an independent capture–recapture estimate of 14 (±3.6) individuals based on photographic identity study in 2009. Looking at high mortality of tigers in the area, with 19 reported deaths during 2000 to 2009, our results indicate high individual turnover in the area. This study shows that Orang National Park harbours a healthy breeding population of tigers. However, the possibility of a source-sink dynamics operating in the landscape could not be ruled out, with possible immigration from nearby Kaziranga National Park on the south bank of Brahmaputra, which has the highest reported density of the species in the world.

Keywords

Panthera tigris tigrisNoninvasive geneticsPopulation estimationMicrosatellitesOrang National ParkAssamNortheast India

Introduction

The tiger (Panthera tigris) is one of the most widely distributed big cats, yet highly endangered throughout the range countries in which it occurs (Seidensticker et al. 1999; Sanderson et al. 2006). Therefore, the Royal Bengal tiger (Panthera tigris tigris) is listed in the Schedule I species of the Indian Wildlife (Protection) Act, 1972, and categorised as Endangered by IUCN Red List (IUCN 2010). Population monitoring of large carnivores such as tigers are difficult to conduct because they are rare and roam over large distances and remote areas (Schipper et al. 2008). Over the years, remotely triggered camera-based photographic identity has been increasingly in use in wildlife studies for determining population size or other ecological aspects of elusive, secretive and nocturnal species such as tiger (Karanth 1995; Karanth and Nichols 1998; Karanth and Nichols 2002; Jhala et al. 2008). The opportunity to gain genetic data from natural populations has been enhanced by the development of noninvasive genetic techniques in the past two decades. The technique involves the extraction of DNA from materials like faeces, shed hair, feathers, urine and sloughed skin (Höss et al. 1992; Kohn and Wayne 1997; Goossens et al. 1998; Segelbacher 2002; Fernando et al. 2003). The methodology has been increasingly applied to a wide variety of species (Taberlet et al. 1997; Reed et al. 1997; Kohn et al. 1999; Ernest et al. 2000; Garnier et al. 2001; Eggert et al. 2003; Goossens et al. 2003; Hedmark et al. 2004; Vidya and Sukumar 2005; Smith et al. 2006; Bhagavatula and Singh 2006). Studies on the genetic diversity of tigers have focused on mitochondrial DNA (Zhang et al. 2006a), and microsatellites (Zhang et al. 2006b; Xu et al. 2005; Williamson et al. 2002). By employing both mitochondrial and nuclear genetic markers, DNA isolated from tissue sources have been used to update and delineate the taxonomic status among the five living tiger subspecies (Luo et al. 2004). Studies have demonstrated the feasibility of using the noninvasive genetic technique for counting tiger individuals through genetic identification of individuals from field collected faecal samples of wild tigers in India (Bhagavatula and Singh 2006) and for species identification (Mukherjee et al. 2007) and sex identification (Sugimoto et al. 2006).

Genetic capture–recapture-based population estimators have been evaluated through various studies, either in comparison with simulated data or direct count estimates (Miller et al. 2005; Puechmaille and Petit 2007). Studies have compared indirect and genetic methods showing both over and under estimates through these methods (Zhan et al. 2006; Hájková et al. 2009). A study by Arandjelovic et al. (2010) has compared the feasibility of genetic capture–recapture using opportunistically collected faecal samples. Capture–recapture-based genetic population estimation of wild tigers with multi-session sampling has successfully been carried out in India (Mondol et al. 2009), thus opening a new era of population monitoring of this species in the country.

The state of Assam with the Brahmaputra floodplains and parts of north-eastern hills in India possibly harbours high potential of long-term tiger conservation in India and also among the areas of range distribution of the species (Jhala et al. 2008). Protected areas in Assam such as Manas National Park, Nameri National Park and Orang National Park (ONP) on the northern bank of river Brahmaputra and Kaziranga National Park (KNP) on the southern bank provide highly suitable habitat and prey base for tigers (Ahmed et al. 2010a). According to the Forest Department’s pugmark-based census, 19 tigers were found in ONP in the year 2000. During 2008 and 2009, seven and 12 individuals were identified in ONP by Ahmed et al. (2009a and 2010b). On the basis of this camera trapping based study, a 2009 capture–recapture population estimate of 14 (±3.6) tigers in ONP was obtained (Ahmed et al. 2010b).

As part of this study, genetic analysis of tiger scat samples collected from Orang National Park in February 2009 were carried out, in order to standardise genetic individual identification of tigers and to determine the minimum number of tiger presents in the area. We have also attempted genetic capture–recapture based population estimation of tigers in ONP, based on single session sampling design.

Materials and methods

Study area

Orang National Park with an area of 78.8 km2 (92°16′ to 92°27′ E, 26°29′ to 26°40′ N) is located within Darrang and Sonitpur districts of Assam (Fig. 1). The river Brahmaputra marks the southern boundary of ONP, while tributaries such as Pachnoi, Belsiri and Dhansiri flow along the boundary of the park. The habitat of ONP is composed of dry grassland (32%), swampy grassland (31%), woodland (19%), scrub forest (5%), water body (7%) and river sand (6%; Talukdar and Sarma 2007). ONP has likely demographic and genetic exchange with KNP through Laokhowa-Burachapori forest complex along the Brahmaputra River (Ahmed et al. 2009b), thus forming a part of the greater Kaziranga-Karbi Anglong population described by Jhala et al. (2008).
https://static-content.springer.com/image/art%3A10.1007%2Fs10344-010-0471-0/MediaObjects/10344_2010_471_Fig1_HTML.gif
Fig. 1

Map showing locations of Orang National Park in India and distribution of genetically identified 48 genuine tiger scats in the area

Apart from tigers, the park harbours Greater One-horned Rhino (Rhinoceros unicornis), Hog Deer (Axis porcinus), Wild Pig (Sus scrofa), Fishing Cat (Felis viverrina), Jungle Cat (Felis chaus) and Leopard Cat (Prionailurus bengalensis) as major mammalian species.

Sample collection

Scat samples were collected from ONP as part of a carnivore sign survey undertaken in February 2009. Global positioning system coordinate readings were recorded while collecting scat samples. Fifteen to 20 g of fresh scat samples were collected in plastic vials containing silica gel (desiccant) and later transported to the laboratory. Samples were further dried in hot-air oven, with precaution for avoiding sample intermixing. After drying, samples were kept in air-tight plastic vials at –20°C Freezer until DNA extraction. We have undertaken opportunistic sampling of 2 to 3 g of muscle tissue from dead wild tigers in ONP and KNP during the 2 years period of 2009 to 2010.

DNA extraction

DNA extractions for all the scats collected were performed by using commercial Stool DNA isolation kit (QIAamp DNA Stool Kit, QIAGEN Ag., Germany) following minor modifications. Modifications over the standard kit protocol are: (1) samples in ASL buffer were vortexed for 10 min; (2) incubation with Inhibitex for 15 min; and (3) elution was performed in 80 μl AE buffer. For every extraction, negative controls composed of reagent blanks without the scat sample were included to monitor contamination. For reference tissue samples from dead tigers, DNeasy Blood and Tissue Kit (QIAGEN Ag., Germany) were used following standard kit protocols. Extracted DNA were run on 1% agarose gel and visualised by ethidium bromide staining on an UV transilluminator.

Genetic species identification of tiger scats

Genetic markers developed by Mukherjee et al. (2007) were used for identification of genuine tiger scats collected from ONP, thus avoiding analysis of scats of other sympatric carnivores. Here, the species identity is based on the presence or absence of tiger-specific mitochondrial polymerase chain reaction (PCR) products of specific size, determined through agarose gel electrophoresis. PCR reactions were carried out using QIAGEN Multiplex PCR Kit (QIAGEN, Germany) following conditions described by Mukherjee et al. (2007) in reaction volume of 10 μl including 2.5 μl of scat DNA.

Sex identification of tiger scats

Genuine tiger scat samples were sexed using primers to amplify the Y chromosome linked SRY (sex determining region) loci as demonstrated in the domestic cat individualization panel, MEOWPLEX (Butler 2002; Butler et al. 2002). This SRY primer pair gives a single PCR product of 99 base pairs in the male and no amplification in the female. This primer pair was first tested on five reference tiger tissue samples with known sex to evaluate their efficacy. Forward primer of this marker was fluorescently tagged, which was screened along with microsatellite loci by capillary electrophoresis. Here microsatellite loci also act as positive controls in distinguishing between PCR failure and false female identification. Sex identification for all the identified tiger scats were performed, irrespective of their individual identity. Thus the sex identity acts as an additional confirmation of multiple scats belonging to the same tiger individual.

Genetic identification of individuals from tiger scats

Selection of polymorphic microsatellite markers

We have used 19 microsatellite loci (Table 1) originally developed from Sumatran tiger (Williamson et al. 2002), Amur tiger (Wu et al. 2008), Asiatic lion (Singh et al. 2002) and Domestic cat (Menotti-Raymond et al. 1999) to screen for a set of polymorphic markers for individual identification of tigers in ONP. The 19 loci include some of the microsatellite loci for individual identification of tigers, as suggested by Bhagavatula and Singh (2006) and Mondol et al. (2009). The forward primers of each microsatellite marker used in the study were labelled at the 5′ end with one of the fluorescent dyes 6–FAM (blue), PET (red), VIC (green) or NED (yellow), while the reverse primers were not labelled. The PCR products generated after amplification were loaded in a capillary electrophoresis based ABI 3130 Genetic Analyzer (Applied Biosystems, USA) and the raw data scored for allele sizes using the software GENEMAPPER v3.7 (Applied Biosystems, USA).
Table 1

Details of 19 microsatellite loci screened on 12 test tiger scat samples

Locus name

No. of alleles

Allele range

% PCR success

Allelic dropout

False alleles

Hexp

Hobs

Reference

F41

4

113–121

60

0.02

0

0.66

0.25

a,b

F53

4

149–157

73

0

0

0.68

0.33

a

Fca126

4

136–146

67

0

0

0.61

0.17

a,b

Fca304

4

124–144

67

0

0

0.71

0.42

a,c

Fca391

4

201–211

67

0.04

0

0.7

0.25

a,b

Fca441

3

150–157

73

0.02

0

0.55

0.42

a,b

Fca453

5

156–171

80

0.05

0

0.7

0.42

a,b

Fca506

4

210–216

60

0

0

0.72

0.17

a

Fca628

5

89–115

73

0

0

0.76

0.33

a,b

HDZ007

5

199–235

47

0.02

0

0.75

0.17

d

HDZ056

6

170–191

67

0

0

0.81

0.42

d

HDZ089

3

206–223

47

0.1

0

0.62

0.08

d

HDZ170

5

200–218

93

0.02

0

0.65

0.33

d

HDZ700

5

139–141

100

0.01

0

0.79

0.42

d

HDZ859

4

260–298

67

0.03

0

0.66

0.25

d

Pati09

4

119–125

80

0

0

0.7

0.31

e

Pati18

6

203–221

67

0

0

0.67

0.42

e

Ple51

3

171–175

80

0

0

0.81

0.17

f

Ple55

7

137–155

93

0.04

0

0.7

0.5

f

aMenotti-Raymond et al. (1999)

bUsed by Mondol et al. (2009)

cUsed by Bhagavatula and Singh (2006)

dWilliamson et al. (2002)

eWu et al. (2008)

fSingh et al. (2002)

Five tissue samples of tigers collected from dead individuals from ONP and KNP during the period of 2009 to 2010 were used as reference samples for all the microsatellite standardisation work. Twelve tiger scat samples, ten from ONP and two from KNP were used for screening all the 19 microsatellite loci, in order to determine their level of polymorphism in the study population. These samples were selected from the sample set of genuine tiger scats on the basis the following criteria—(1) distant geographic location in the study area, thus with higher probability of including samples from different individuals, (2) freshness during the time of collection and (3) PCR amplification success with all the four species identification markers (mitochondrial) and one sexing marker (nuclear). The set of the genotype data from these samples were used in selecting a panel of microsatellite markers, based on following three criteria: (1) polymorphism information content and expected heterozygosity, with no linkage disequilibrium (2) mean PCR success from scat DNA samples and (3) multiplexing compatibilities. The selected panel of markers were screened on five wild collected reference tiger tissue samples to demonstrate the power of individual identification as well as standardisation of multiplexing PCR reactions.

Individual identification of tiger scats

A panel of selected polymorphic microsatellite loci were used further to genotype all the genuine tiger scats. All the PCR were carried out in multiplex of four loci in a single 10 μl reaction, each locus labelled with a separate fluorescent label. Multiplexing was carried out using QIAGEN Multiplex PCR Kit (QIAGEN, Germany) following standard kit protocols for reagent concentration with 0.2 μM of each primer and 2.5 μl template DNA in a 10 μl PCR reaction. The thermal cycling was performed with 95°C initial denaturation/activation of 15 min, followed by 40 cycles of 94°C for 45 s, 55° for 45 s and 72°C for 45 s followed by a 72°C for 15 min final extension step.

Assessing and minimising microsatellite genotyping error

Microsatellite genotyping errors principally arise due to allelic dropout (ADO) or false alleles (FA). ADO in the detection of only one allele in a heterozygous individual, and FA creates a new allele, which generates spurious heterozygotes. Each sample and loci were typed at three or more replicates depending on the genotype discrepancy (Navidi et al. 1992) and consensus genotypes were created from these repeat results. Per locus and per multilocus genotype error rates were estimated as a measure of the reliability of the genotypes obtained, which is important in genetic individual identification (Bellemain 2004; Pompanon et al. 2005). To minimise electropherogram stutter patterns and prevent allele dropout, dedicated microsatellite PCR kits (Multiplex PCR Kit, QIAGEN, Germany) were employed throughout for genotyping. Moreover, all the work were carried out in a room dedicated for low copy number DNA analysis and aerosol-barrier tips were used to prevent sample to sample contamination.

Genetic data analysis

Allele sizing was carried out by combining automated allele calling and visual inspection of electropherogram data for each locus in each sample. This process provides a balance between the efficiency and consistency of automated allele-calling software (GENEMAPPER v3.7, Applied Biosystems, USA) and the accuracy provided by human inspection in detecting novel alleles outside of the expected range of a locus, stochastic amplifications within the size range and potential mistypes due to stutter or large-allele dropout (Pompanon et al. 2005; Dewoody et al. 2006). Once samples were scored at all loci, the data was tested for scoring errors. Consensus genotypes were checked using the programme GIMLET v 1.3.3 (Valiere 2002), and also to estimate error rates. In order to select a final set of samples for individual identification, quality index value was assigned to each genotype as per Miquel et al. (2006). Samples that revealed the same genotype in all three repetitions had a quality index of one, whereas, samples that yielded two same genotypes out of the three amplifications had a quality index of 0.667. For the purpose of selection of samples for final data analysis, quality index of 0.667 was kept as the cut-off value.

For purposes of individual identification, five to seven microsatellite loci are usually sufficient, and act as a trade-off between achieving a reasonably low probability of identity (PID) value and limiting the genotyping error (the more the loci added, the higher will be the error rates introduced into the study). From the microsatellite genotype data, allele frequency, observed and expected heterozygosity and probability of identity (PID) and probability of identity among siblings (PID-sibs) were calculated using the software GIMLET v 1.3.3 (Valiere 2002). Tests for linkage disequilibrium and Hardy–Weinberg equilibrium were performed using the software Arlequin V3.0 (Excoffier et al. 2005). The unique multilocus microsatellite genotypes, i.e. individual tigers were identified using the identity analysis module of the programme CERVUS (Marshall et al. 1998). Here, incomplete multilocus genotypes were also considered, i.e. samples with identical genotypes for at least five or more common loci with missing data for the rest are considered as same individual.

Programme CAPWIRE (Miller et al. 2005) was used to estimate population size in single-session sampling scheme. Both the models of capture probability incorporated in CAPWIRE were used to derive population estimates with 95% confidence intervals with parametric bootstrap of 10,000 replicates. The even capture model (ECM) assumes there is no capture heterogeneity in the data set while the two innate rates model (TIRM) assigns individuals as having either a high or a low capture probability. Both the models have been used, as the method to detect capture heterogeneity has been shown to be inaccurate in several cases (Puechmaille and Petit 2007). The samples collected at the same time and location (coordinates) were considered as a single observation, as adopted by Zhan et al. (2006). Samples of Individuals from ONP with missing coordinates in this study were not included in the final capture–recapture analysis.

Results

A total of 58 carnivore scats were collected from ONP in February 2009. DNA extractions were performed on 57 scats, whereas one sample was discarded due to fungal growth. Forty-eight of the total scats could be confirmed as genuine tiger scats with PCR-based genetic species identification, whereas two were of non-tiger origin. Seven of the scats failed to produce any results. The map distribution of the genuine tiger scats with missing location data for six scats obtained from ONP is shown in Fig. 1.

The number of alleles, allele size range, percentage PCR success, ADO, FA, expected (Hexp) and observed (Hobs) heterozygosity for the 19 microsatellite loci in 12 test scat DNA samples are presented in Table 1. All the 19 loci were found to be polymorphic, with three to seven number of allele per locus. The observed heterozygosity varied from 0.08 to 0.50 across the 19 loci. PCR success rate of 47% to 100% were obtained across the 19 loci in 12 test scat DNA samples. The per locus ADO value varied from 0 to 0.05, whereas no FA was obtained. Eight of the 19 loci, Fca126, Fca304, Fca628, HDZ170, HDZ700, Pati18, Ple51 and Ple55 showed significant deviation from Hardy–Weinberg equilibrium, with no linkage disequilibrium. On the basis of the three criteria mentioned in the methodology, eight polymorphic microsatellite loci were selected for genotyping genuine tiger scat samples from ONP. Cumulative PID value of 1.24 × 10−7 and PID-sibs value of 1.33 × 10−3 were obtained for eight selected loci in 12 scat samples, which denote high resolving power of this set of markers in individual identification of tigers from the study population. The eight polymorphic loci were screened on five reference tiger tissue samples for assessing the reliability in identifying known different individuals from the wild population (Table 2). These eight loci with three to seven alleles per locus could successfully identify all the five different individual tigers. PCR amplification success of 100% was obtained for the eight loci on five reference tissue samples, with no ADO and FA. Figure 2 shows a graphical representation of product PID and PID-sibs values of eight selected polymorphic microsatellite loci in 12 test scat DNA samples and five reference tissue samples.
Table 2

Details of eight polymorphic microsatellite loci screened on five wild tiger reference tissue samples

Locus

No. of alleles

Allele range

% PCR success

Allelic dropout

False alleles

Hexp

Hobs

Multiplex PCR

Fca441

3

150–157

100

0

0

0.64

0.8

I

HDZ170

4

202–217

100

0

0

0.7

0.8

I

Pati09

6

115–129

100

0

0

0.78

0.8

I

Ple51

4

169–179

100

0

0

0.58

0.6

I

F53

3

153–167

100

0

0

0.46

0.6

II

Fca304

2

122–126

100

0

0

0.5

0.2

II

HDZ700

3

139–149

100

0

0

0.58

0.6

II

Ple55

3

147–151

100

0

0

0.46

0.2

II

https://static-content.springer.com/image/art%3A10.1007%2Fs10344-010-0471-0/MediaObjects/10344_2010_471_Fig2_HTML.gif
Fig. 2

Graphical representation Product PID and PID-sibs values of eight selected polymorphism microsatellite loci in a 12 test scat DNA samples and b five reference tissue samples

As per the quality index criteria mentioned in the methodology, genotype data for 24 scat samples were selected for identification of unique multilocus genotypes and determining the number of individual tigers. These 24 scats include the initial ten test scats from ONP used for screening 19 loci. The number of alleles, allele size range, percentage PCR success rate, ADO and FA, Hexp and Hobs in 24 tiger scat samples from ONP are presented in Table 3. The mean observed heterozygosity of the eight loci in 24 scat samples is found to be 0.47. Calculated per locus ADO was 0.034 and per sample ADO of 0.032, whereas there was no FA amplification observed for the 24 final scat samples. Cumulative PID value of 1.09 × 10−6 and PID-sibs value of 2.85 × 10−3 were obtained for eight loci in 24 scat samples. Seventeen unique multilocus genotypes were obtained based on available genotype data with at least five or more loci and allowing zero locus mismatches, thus 17 individual tigers (Table 4). Locations of scats of all the 17 individual tigers are shown in Fig. 3. However, one sample location for each individual F2, M1 and M5 were missing in the dataset and subsequently in our map in Fig. 3. Sex identification data was incorporated for all the 24 genotyped tiger scats and all the matching multilocus genotypes, i.e. different samples from the same individual tiger, shared the same sex identity. This further confirms the individual identity of the matching samples. In total, identity of five male and 12 female tigers could be confirmed in the dataset. However, when one locus mismatch was allowed in comparing multilocus genotype matches among samples, two pairs of samples showed match at all but locus HDZ170, two pairs of samples matches at all but each locus Fca304 and F53. However, all these matching pairs except for one could be identified as different individuals on the basis of their sex identity. Only one pair of samples with a mismatch at locus HDZ170 shares the same sex identity. If this mismatch was allowed, the number of tiger individuals would come down to 16, with 11 female and five male tigers. As stringent data generation and selection criteria was followed, only unique multilocus genotypes based on zero locus mismatch was considered for individual identification in the present study. During camera trapping-based monitoring of tigers in ONP (Ahmed et al. 2010b), two male, four female and one individual of unknown sex were found. Thus, the sex ratio of tigers obtained through genetic (1:2.4) and camera trapping based monitoring (1:2.5) are in close agreement with each other.
Table 3

Details of eight polymorphic microsatellite loci used in individual identification of 24 tiger scat samples

Locus name

No. of alleles

Allele range

% PCR success

Allelic dropout

False alleles

Hexp

Hobs

Fca441

7

145–175

90

0.037

0

0.63

0.38

HDZ170

7

198–218

99

0.037

0

0.76

0.38

Pati09

5

107–125

89

0.03

0

0.65

0.46

Ple51

3

171–175

93

0

0

0.57

0.17

F53

7

149–167

86

0

0

0.76

0.67

Fca304

5

122–144

92

0.104

0

0.61

0.67

HDZ700

3

137–141

93

0.02

0

0.55

0.71

Ple55

4

147–155

90

0.048

0

0.5

0.29

Table 4

Multilocus microsatellite genotypes and sex identity of 17 individual tigers in Orang National Park

Individual ID

Sex (SRY)

Fca441

HDZ170

Pati09

Ple51

F53

Fca304

HDZ700

Ple55

  

A

B

A

B

A

B

A

B

A

B

A

B

A

B

A

B

F1

Female

145

175

204

204

?

?

173

173

?

?

122

126

139

141

147

147

F2

Female

153

165

214

214

123

123

173

173

131

139

132

144

137

139

151

151

F3

Female

157

157

212

216

123

125

173

173

153

153

122

126

?

?

?

?

F4

Female

149

149

216

218

117

125

171

171

149

157

122

126

139

141

147

147

F5

Female

151

157

218

218

123

123

173

173

?

?

?

?

139

141

155

155

F6

Female

151

151

216

216

123

125

171

171

149

171

122

122

139

141

147

147

F7

Female

149

149

212

216

125

125

171

171

167

167

122

126

141

141

147

147

F8

Female

149

149

212

216

125

125

171

171

153

157

122

126

139

141

147

147

F9

Female

149

165

212

212

125

125

171

171

149

167

122

122

139

141

147

147

F10

Female

149

149

216

216

123

125

?

?

149

149

122

126

139

139

149

149

F11

Female

149

149

212

216

125

125

171

171

157

167

122

126

139

141

147

147

F12

Female

?

?

216

216

117

125

171

171

149

167

122

122

139

141

147

147

M1

Male

149

149

198

198

117

125

171

175

153

167

122

126

139

141

147

151

M2

Male

153

153

216

216

107

107

173

173

167

167

130

144

137

139

147

147

M3

Male

149

165

216

216

119

119

171

171

149

167

130

130

139

141

147

147

M4

Male

?

?

212

212

?

?

171

171

?

?

122

126

139

141

147

147

M5

Male

149

165

202

216

123

125

173

173

157

167

122

122

139

139

147

149

? missing data

https://static-content.springer.com/image/art%3A10.1007%2Fs10344-010-0471-0/MediaObjects/10344_2010_471_Fig3_HTML.gif
Fig. 3

Map showing locations of genetically identified five male (M1, M2, etc.) and 12 female (F1, F2, etc.) tigers in Orang National Park

The population size estimates obtained by genetic capture–recapture analysis in CAPWIRE were 27 (18 to 41) and 36 (19 to 50) for ECM and TIRM models, respectively. The lowest range estimate of 18, obtained through even capture probability assumption, should serve as the genetic population estimate of tigers in ONP, considering the similarity under both the models (18 and 19, for ECM and TIRM, respectively).

Discussion

The present study, for the first time uses noninvasive genetic techniques for indentifying individual tigers from a protected area in north-eastern India. The eight microsatellite loci selected shows a considerably low PID and PID-sibs value for both reference tissue samples as well as field collected tiger scat samples. A PID-sibs value of 2.85 × 10−3 corresponds to one in every 350 individuals sharing the same multilocus genotype signature, being selected as the cut-off value, which is in agreement with the value suggested by other studies (Waits et al. 2001; Mondol et al. 2009). The cut-off PID-sibs value of 5 × 10−3 obtained by Mondol et al. (2009) using five loci and by Bhagavatula and Singh (2006) using six loci was obtained in the present study by using seven loci. However, the mean PCR amplification success from scat samples in the present study was 91.5%, which is higher compared to 85% obtained by Mondol et al. (2009) and 60% obtained by Bhagavatula and Singh (2006).

Result of genetic counting of tigers is supported by camera trapping-based population estimate of 14 (±3.6) obtained by Ahmed et al. (2010b). Prior to these studies, pugmark-based census undertaken by the Forest Department in year 2000 found 19 tigers in ONP. However, there are 19 tiger mortality cases reported from ONP during 2000 to 2009 (Table 5; data source: Orang National Park Authority), i.e. the time interval between Forest department census and photographic capture–recapture and genetic estimate. More than 70% of these mortality cases are due to tiger kill poisoning and suspected human-induced injuries. Thus, the presence of 17 individual tigers in the year 2009 shows that, in spite of high mortality, ONP seems to be maintaining a steady number of tigers in past one decade. So far, there is no available data on emigration rate, which is likely to add up to the total tiger loss from ONP, if considered. Karanth et al. (2006) estimated the annual rate of loss to a tiger population to be as high as 23% (which includes mortality and emigration) in a camera trapping-based study conducted in Nagarahole Reserve of Karnataka, India. Thus, in case of ONP, the annual mortality rate of nearly 20% is considerably high and the total annual tiger loss including emigration (as suggested by Karanth et al. 2006 for Nagarahole) may even be higher.
Table 5

Mortality records of tigers from Orang National Park in the duration 2000 to 2009 (data source: Orang National Park Authority)

Sl. no.

Date

Locality

Cause

Sex

Age class

1

23-02-2000

Bejimari

Poisoning

Female

Adult

2

25-03-2000

Bontapu

Infighting

Female

Adult

3

12-12-2000

Gandarmari

Natural

Female

Cub

4

23-01-2003

Rongagora Village

Poisoning

Female

Adult

5

28-11-2004

Solmari

Natural

Female

Adult

6

26-01-2005

Nislamari

Natural

Unknown

Cub

7

15-11-2005

Bhutiali

Poisoning

Male

Adult

8

14-11-2005

Pachnoi 2

Poisoning

Unknown

Adult

9

14-02-2006

Gaimari

Infighting

Male

Adult

10

16-02-2006

Kachomari

Head injury

Female

Adult

11

06-11-2006

Jahoni (Kathgora)

Poisoning

Unknown

Cub

12

07-11-2006

Jahoni (Kathgora)

Poisoning

Unknown

Cub

13

13-11-2006

Jahoni (Kathgora)

Poisoning

Unknown

Cub

14

13-11-2006

Jahoni (Kathgora)

Poisoning

Female

Adult

15

02-10-2007

Old Orang

Poisoning

Male

Adult

16

04-10-2007

Old Orang

Poisoning

Unknown

Adult

17

27-03-2008

Hazarbigha

Head injury

Female

Adult

18

15-5-2009

Hazarbigha

Natural

Male

Sub-adult

19

19-08-2009

Nislamari

Poisoning

Male

Adult

The genetic capture–recapture estimates (N^) of 27 (18 to 41) and 36 (19 to 50) through two deposition models of CAPWIRE were probable overestimates, in comparison with camera trapping-based estimate by Ahmed et al. (2010b) and Forest Department census figures. Therefore, the minimum range estimate of 18 individuals under even capture probability assumption is proposed to be a realistic figure, keeping in consideration other studies in the area.

ONP lies in a geographic location on the northern bank of the river Brahmaputra, close to Kaziranga National Park on the south bank. KNP is one of the highest tiger density areas in the world (Karanth and Nichols 1998; Ahmed et al. 2010a) and considered as the source population in the Brahmaputra Valley (Jhala et al. 2008). Field study undertaken by Ahmed et al. (2009b) have already shown that the islands on the river Brahmaputra may provide opportunity for tigers dispersing from source population (such as KNP) to other tiger bearing protected areas (such as ONP) that provide opportunity for breeding. During our study, one male tiger M5 can be found in the mainland of ONP as well as an island named Jahoni in river Brahmaputra (Fig. 3). This is a first time report that the tigers from the mainland are using the Brahmaputra river islands.

ONP may be maintaining a steady population size, as indicated by the present study, because of two main factors: first, the tiger population in ONP is a healthy breeding one and second, there is a possible source-sink dynamics operating in the landscape, with continuous immigration of tigers from nearby source populations such as KNP. However, as hypothesised by Karanth and Stith (1999), although tiger populations have high mortality rates from natural and anthropogenic causes, they can be demographically viable if supported on an abundant prey base. In order to understand the current pattern of high tiger population turnover in ONP, a long-term population monitoring of tigers and its prey base in the landscape, covering ONP, KNP and riverine islands of the Brahmaputra is required. For long-term conservation of tigers in the Brahmaputra Valley of Assam, the above mentioned factors must be studied in detail prior to formulation of any long-term management strategy, along with strong measures for minimising the anthropogenic causes of tiger mortality in ONP.

Acknowledgements

We are grateful to the Forest Department of Assam and local Forest Department of Orang National Park for providing access to the park. We are also grateful to the Ministry of Environment and Forests, Government of India, and the Chief Wildlife Warden of Assam for providing us permission to undertake sample collection in the protected areas of Assam. Our special thanks go to Mr. Pranjal Kumar Das, Researcher at Wildlife Genetics Laboratory, Aaranyak, for his support to this work. Thanks also go to Mr. Kamal Azad and Mr. Dhritiman Das for their support in the field. We sincerely thank Mr. Pranjit Kumar Sarma and the team of GIS experts at Aaranyak for their help and support during this work. Thanks to Dr. Hilloljyoti Sinha, Assistant Professor at Assam University, Silchar, India, for his valuable suggestions to the manuscript. We thank our field assistants Mr. Biraj Saikia and Mr. Anil Das for their support in the field. Thanks also to all the officials and staff of Aaranyak for their support to the smooth running of the field and laboratory work. We also thank Mr. Bibek Yumnam, Senior Research Fellow at Wildlife Institute of India, for his valuable input to this work. The laboratory work was financially supported by Aaranyak Rufford Small Grants and field study was supported by SeaWorld Busch Gardens Conservation Fund.

Authors’ contributions

UB conceived the study design, data analysis, wrote the manuscript and provided guidance to RDB in carrying out the laboratory work. RDB carried out the laboratory work. CD, AB and MFA undertook the field sampling. MFA also provided guidance to the field team in sample collection using the same grids where simultaneously camera traps are also being installed. AT contributed towards reference tissue sample collection. BKT provided overall administrative support, research planning and guidance for the smooth running of the project. RB provided inputs to the manuscript writing. All the authors have read and accepted the final manuscript.

Copyright information

© Springer-Verlag 2010