research paper on indian grey wolf

We can help you reset your password using the email address linked to your BioOne Complete account.

• BioOne Complete
• BioOne eBook Titles
• By Publisher
• About BioOne Digital Library
• How to Subscribe & Access
• Library Resources
• Publisher Resources
• Instructor Resources
• FIGURES & TABLES
• SUPPLEMENTAL CONTENT
• DOWNLOAD PAPER SAVE TO MY LIBRARY

Wolves play a crucial role in shaping ecological communities as an apex predator in the dry-open forests of semi-arid landscapes in India. Large scale habitat loss pertaining to human expansion and retaliatory killing by human caused severe decline in the wolf population across its range. The estimated wolf population size is close to 2000–3000 individuals in India; however, these estimates were decades old and the present status of the wolf in the semi-arid landscape is largely unknown. We assessed the distribution of wolves in Kailadevi Wildlife Sanctuary, Rajasthan using occupancy models and identified important factors associated with habitat-use by wolves. Occupancy modelling shifts the focus from individual animal to a site, while accounting for detection probability. To assess the habitat-use we used sign-based surveys that rely on data collected from adjacent sampling sites (replicates). The habitat-use was assessed across 672.82 km 2 surveying 48 grid cells, each measuring 14.44 km 2 . Estimated habitat-use Ѱ ( SD ) was found to be 0.82 (0.14). Our findings suggested that availability of agriculture land had the significant positive influence on the habitat-use of wolves. Other factors such as availability of water, scrubland, and wild prey (nilgai and chinkara) also had a positive effect on the habitat use of wolves, but it was not significant. Forest cover has a negative influence on the habitat use of wolves. This study is the first rigorous assessment of the Indian grey wolf habitat-use at the level of wildlife reserve with potential conservation value that can be applied to other areas in India.

KEYWORDS/PHRASES

Publication title:, collection:, publication years.

Advertisement

Genetic diversity, structure, and demographic histories of unique and ancient wolf lineages in India

• Research Article
• Published: 07 August 2023
• Volume 25 , pages 33–48, ( 2024 )

Cite this article

• Yellapu Srinivas   ORCID: orcid.org/0000-0002-5412-4717 1 &
• Yadvendradev Jhala   ORCID: orcid.org/0000-0003-3276-1384 1  

600 Accesses

17 Altmetric

Explore all metrics

Assessing genetic diversity, population connectivity, demographic patterns, and phylogeographic relationships is vital for understanding the evolutionary history of species and thus aid in conservation management decisions. Indian wolves (currently, Canis lupus pallipes and Canis lupus chanco ) are considered ancient, unique and divergent lineages among grey wolves, yet their population genetics are poorly understood. To void this knowledge gap, we collected samples from Indian peninsular ( n  = 77) and Himalayan wolves ( n  = 24) and used a combination of maternal (mtDNA CR and Cyt b ) and bi-parental (nuclear microsatellites) markers to estimate levels of genetic diversity, examine the patterns of genetic structuring between them and within their distribution range, and assess their demographic histories. Both the wolf populations showed moderate levels of genetic variability, comparable to other grey wolves. Low levels of genetic differentiation were observed within both the Indian peninsular and Himalayan wolves indicating high levels of gene flow within their populations. On the other hand, high levels of genetic differentiation were observed between the two wolves indicating absence of gene flow. Molecular analysis highlighted the uniqueness of both the Indian wolves which was further supported by the presence of unique haplotypes indicating no admixture between them. Demographic analysis using both mtDNA and microsatellites revealed decline in population sizes of both the wolf lineages and both have undergone bottlenecks. Estimates of past effective population size revealed recent population declines of both lineages of Indian wolves at around 25–50 generations corresponding to about 100–200 years ago. Our results further support the designation of both lineages of Indian wolves as two distinct species Canis pallipes and Canis himalayensis and suggest increasing conservation efforts to save the unique and ancient wolf species from extinction.

This is a preview of subscription content, log in via an institution to check access.

Access this article

Subscribe and save.

• Get 10 units per month
• Download Article/Chapter or eBook
• 1 Unit = 1 Article or 1 Chapter
• Cancel anytime

Price includes VAT (Russian Federation)

Instant access to the full article PDF.

Rent this article via DeepDyve

Institutional subscriptions

Similar content being viewed by others

The role of the Caucasus, Carpathian, and Dinaric–Balkan regions in preserving wolf genetic diversity

Global Phylogeographic and Admixture Patterns in Grey Wolves and Genetic Legacy of An Ancient Siberian Lineage

Evidence of strong population bottleneck in genetics of endangered Brazilian Merganser ( Mergus octosetaceus)

Data availability.

The datasets presented in this study can be found in online repositories. Mitochondrial DNA sequence data can be found using accession numbers ON010580- ON010589 on GenBank ( https://www.ncbi.nlm.nih.gov/genbank/ ). Nuclear microsatellite genotypes data can be found on repository with https://doi.org/10.6084/m9.figshare.19385912

Aggarwal RK, Ramadevi J, Singh L (2003) Ancient origin and evolution of the Indian wolf: evidence from mitochondrial DNA typing of wolves from Trans-Himalayan region and Pennisular India. Genome Biol 4(6):1–30

Article   Google Scholar  

Aggarwal RK, Kivisild T, Ramadevi J, Singh L (2007) Mitochondrial DNA coding region sequences support the phylogenetic distinction of two Indian wolf species. J Zool Syst Evol Res 45(2):163–172

Akaike H (1974) A new look at the statistical model identification. IEEE Trans Autom Control 19(6):716–723

Allendorf FW, Luikart GH, Aitken SN (2013) Conservation and the genetics of populations, 2nd edn. Wiley-Blackwell Publishing, p 602

Google Scholar  

Aspi J, Roininen E, Ruokonen M, Kojola I, Vila C (2006) Genetic diversity, population structure, effective population size and demographic history of the Finnish wolf population. Mol Ecol 15(6):1561–1576

Article   PubMed   CAS   Google Scholar  

Banks SC, Cary GJ, Smith AL, Davies ID, Driscoll DA, Gill AM et al (2013) How does ecological disturbance influence genetic diversity? Trends Ecol Evol 28(11):670–679

Article   PubMed   Google Scholar  

Boitani, L., Phillips, M. & Jhala, Y. 2018. Canis lupus (errata version published in 2020). The IUCN Red List of Threatened Species 2018: e.T3746A163508960. https://doi.org/10.2305/IUCN.UK.2018-2.RLTS.T3746A163508960.en .

Boitani L, Phillips M, Jhala YV (2020) Canis lupus (errata version published in 2020). The IUCN red list of threatened species 2018: e. T3746A163508960.

Breen M, Jouquand S, Renier C, Mellersh CS, Hitte C, Holmes NG et al (2001) Chromosome-specific single-locus FISH probes allow anchorage of an 1800-marker integrated radiation-hybrid/linkage map of the domestic dog genome to all chromosomes. Genome Res 11(10):1784–1795

Article   PubMed   PubMed Central   CAS   Google Scholar  

Chevalet C, Nikolic N (2010) The distribution of coalescence times and distances between microsatellite alleles with changing effective population size. Theor Popul Biol 77(3):152–163

Coombs JA, Letcher BH, Nislow KH (2008) CREATE: a software to create input files from diploid genotypic data for 52 genetic software programs. Mol Ecol Resour 8(3):578–580

Cornuet JM, Luikart G (1996) Description and power analysis of two tests for detecting recent population bottlenecks from allele frequency data. Genetics 144(4):2001–2014

Crooks KR (2002) Relative sensitivities of mammalian carnivores to habitat fragmentation. Conserv Biol 16(2):488–502

Czarnomska SD, Jędrzejewska B, Borowik T, Niedziałkowska M, Stronen AV, Nowak S et al (2013) Concordant mitochondrial and microsatellite DNA structuring between Polish Lowland and Carpathian Mountain wolves. Conserv Genet 14(3):573–588

Earl DA, VonHoldt BM (2012) STRUCTURE HARVESTER: a website and program for visualizing STRUCTURE output and implementing the Evanno method. Conserv Genet Resour 4(2):359–361

Excoffier L, Lischer HE (2010) Arlequin suite ver 3.5: a new series of programs to perform population genetics analyses under Linux and Windows. Mol Ecol Resour 10(3):564–567

Fan Z, Silva P, Gronau I, Wang S, Armero AS, Schweizer RM et al (2016) Worldwide patterns of genomic variation and admixture in gray wolves. Genome Res 26(2):163–173

Fox JL, Chundawat RS (1995) Wolves in the trans-Himalayan region of India: the continued survival of a low-density population. Ecology and conservation of wolves in a changing world, proceedings of the 2nd North American symposium on wolves. Edmonton, Alberta, Canada, pp. 95–104

Francis RM (2017) pophelper: an R package and web app to analyse and visualize population structure. Mol Ecol Resour 17(1):27–32

Francisco LV, Langsten AA, Mellersh CS, Neal CL, Ostrander EA (1996) A class of highly polymorphic tetranucleotide repeats for canine genetic mapping. Mamm Genome 7(5):359–362

Frankham R (2006) Genetics and landscape connectivity. Connectivity conservation. Cambridge University Press, pp 72–96

Chapter   Google Scholar  

Fredholm M, Winterø AK (1995) Variation of short tandem repeats within and between species belonging to the Canidae family. Mamm Genome 6(1):11–18

Fu YX (1997) Statistical tests of neutrality of mutations against population growth, hitchhiking and background selection. Genetics 147(2):915–925

Garza JC, Williamson EG (2001) Detection of reduction in population size using data from microsatellite loci. Mol Ecol 10(2):305–318

Gaubert P, Bloch C, Benyacoub S, Abdelhamid A, Pagani P, Djagoun CAMS et al (2012) Reviving the African wolf Canis lupus lupaster in North and West Africa: a mitochondrial lineage ranging more than 6,000 km wide. PLoS ONE 7(8):e42740

Girod C, Vitalis R, Leblois R, Fréville H (2011) Inferring population decline and expansion from microsatellite data: a simulation-based evaluation of the Msvar method. Genetics 188(1):165–179

Article   PubMed   PubMed Central   Google Scholar  

Góis-Marques CA, Rubiales JM, de Nascimento L, de Sequeira MM, Fernández-Palacios JM, Madeira J (2020) Oceanic Island forests buried by Holocene (Meghalayan) explosive eruptions: palaeobiodiversity in pre-anthropic volcanic charcoal from Faial Island (Azores, Portugal) and its palaeoecological implications. Rev Palaeobot Palynol 273:104116

Gopalakrishnan S, Sinding MHS, Ramos-Madrigal J, Niemann J, Castruita JAS, Vieira FG et al (2018) Interspecific gene flow shaped the evolution of the genus Canis. Curr Biol 28(21):3441–3449

Goudet J (1995) FSTAT (version 12): a computer program to calculate F-statistics. J Hered 86(6):485–486

Hellmann AP, Rohleder U, Eichmann C, Pfeiffer I, Parson W, Schleenbecker U (2006) A proposal for standardization in forensic canine DNA typing: allele nomenclature of six canine-specific STR loci. J Forensic Sci 51(2):274–281

Hennelly LM, Habib B, Modi S, Rueness EK, Gaubert P, Sacks BN (2021) Ancient divergence of Indian and Tibetan wolves revealed by recombination-aware phylogenomics. Mol Ecol 30(24):6687–6700

Hindrikson M, Remm J, Maennil P, Ozolins J, Tammeleht E, Saarma U (2013) Spatial genetic analyses reveal cryptic population structure and migration patterns in a continuously harvested grey wolf ( Canis lupus ) population in North-Eastern Europe. PLoS ONE 8(9):e75765

Hodgson BH (1847) Description of the wild ass ( Asinus polydon ) and wolf of Tibet (Lupus laniger): Calcutta. J Nat Hist 7:474

Hughes AR, Inouye BD, Johnson MT, Underwood N, Vellend M (2008) Ecological consequences of genetic diversity. Ecol Lett 11(6):609–623

Ichikawa Y, Takahashi Y, Tsumagari S, Takeishi M, Ishihama K, Morita M et al (2002) Identification and characterization of 40 dinucleotide microsatellites in the dog genome. Anim Genet 33(5):400–401

Jhala YV (2003) Status, ecology and conservation of the Indian wolf Canis lupus pallipes Sykes. J Bombay Nat Hist Soc 100(2&3):293–307

Jhala YV, Giles RH Jr (1991) The status and conservation of the wolf in Gujarat and Rajasthan, India. Conserv Biol 5(4):476–483

Jhala Y, Saini S, Kumar S, Qureshi Q (2022) Distribution, status, and conservation of the Indian Peninsular wolf. Front Ecol Evo 10:814966

Johnson PC, Haydon DT (2007) Maximum-likelihood estimation of allelic dropout and false allele error rates from microsatellite genotypes in the absence of reference data. Genetics 175(2):827–842

Jombart T (2008) adegenet: a R package for the multivariate analysis of genetic markers. Bioinformatics 24(11):1403–1405

Joshi B, Lyngdoh S, Singh SK, Sharma R, Kumar V, Tiwari VP et al (2020) Revisiting the Woolly wolf ( Canis lupus chanco ) phylogeny in Himalaya: Addressing taxonomy, spatial extent and distribution of an ancient lineage in Asia. PLoS ONE 15(4):e0231621

Kalinowski ST, Taper ML, Marshall TC (2007) Revising how the computer program CERVUS accommodates genotyping error increases success in paternity assignment. Mol Ecol 16(5):1099–1106

Koblmüller S, Vilà C, Lorente-Galdos B, Dabad M, Ramirez O, Marques-Bonet T et al (2016) Whole mitochondrial genomes illuminate ancient intercontinental dispersals of grey wolves ( Canis lupus ). J Biogeogr 43(9):1728–1738

Kocher TD, Thomas WK, Meyer A, Edwards SV, Pääbo S, Villablanca FX, Wilson AC (1989) Dynamics of mitochondrial DNA evolution in animals: amplification and sequencing with conserved primers. Proc Natl Acad Sci 86(16):6196–6200

Korablev MP, Korablev NP, Korablev PN (2021) Genetic diversity and population structure of the grey wolf ( Canis lupus Linnaeus , 1758) and evidence of wolf × dog hybridisation in the centre of European Russia. Mamm Biol 101(1):91–104

Lande R (1993) Risks of population extinction from demographic and environmental stochasticity and random catastrophes. Am Nat 142(6):911–927

Leigh JW, Bryant D (2015) Popart: full-feature software for haplotype network construction. Methods Ecol Evol 6(9):1110–1116

Lucchini V, Fabbri E, Marucco F, Ricci S, Boitani L, Randi E (2002) Noninvasive molecular tracking of colonizing wolf ( Canis lupus ) packs in the western Italian Alps. Mol Ecol 11(5):857–868

Maurya KK, Shafi S, Mall A, Ojha G, Roy HK (2021) First photographic record of Indian wolf Canis lupus pallipes in Valmiki Tiger reserve, Bihar, India. Canid Biol Conserv 23(1):1–4

Mech LD (1974) Canis lupus. Mamm Species 37:1–6

Mech LD, Boitani L (2010) Wolf social ecology. Wolves: behavior, ecology, and conservation. University of Chicago Press, pp 1–34

Mech LD, Barber-Meyer SM, Erb J (2016) Wolf ( Canis lupus ) generation time and proportion of current breeding females by age. PLoS ONE 11(6):e0156682

Montana L, Caniglia R, Galaverni M, Fabbri E, Ahmed A, Bolfíková BČ et al (2017) Combining phylogenetic and demographic inferences to assess the origin of the genetic diversity in an isolated wolf population. PLoS ONE 12(5):e0176560

Moura AE, Tsingarska E, Dąbrowski MJ, Czarnomska SD, Jędrzejewska B, Pilot M (2014) Unregulated hunting and genetic recovery from a severe population decline: the cautionary case of Bulgarian wolves. Conserv Genet 15(2):405–417

Musiani M, Leonard JA, Cluff HD, Gates CC, Mariani S, Paquet PC, Villa C, Wayne RK (2007) Differentiation of tundra/taiga and boreal coniferous forest wolves: genetics, coat colour and association with migratory caribou. Mol Ecol 16(19):4149–4170

Musiani M, Boitani L, Paquet PC (eds) (2010) The world of wolves: new perspectives on ecology, behaviour, and management. University of Calgary Press, Calgary

Neff MW, Broman KW, Mellersh CS, Ray K, Acland GM, Aguirre GD et al (1999) A second-generation genetic linkage map of the domestic dog, Canis familiars . Genetics 151(2):803–820

Nikolic N, Chevalet C (2014) Detecting past changes of effective population size. Evol Appl 7(6):663–681

Ostrander EA, Sprague GF Jr, Rine J (1993) Identification and characterization of dinucleotide repeat (CA) n markers for genetic mapping in dog. Genomics 16(1):207–213

Parra D, García D, Mendez S, Cañon J, Dunner S (2010) High mutation rates in canine tetranucleotide microsatellites: too much risk for genetic compatibility purposes? Open Forensic Sci J 3(1).

Paxinos E, McIntosh C, Ralls K, Fleischer R (1997) A noninvasive method for distinguishing among canid species: amplification and enzyme restriction of DNA from dung. Mol Ecol 6(5):483–486

Peery MZ, Kirby R, Reid BN, Stoelting R, Doucet-Bëer ELENA, Robinson S et al (2012) Reliability of genetic bottleneck tests for detecting recent population declines. Mol Ecol 21(14):3403–3418

Pilot M, Branicki W, Jędrzejewski W, Goszczyński J, Jędrzejewska B, Dykyy I et al (2010) Phylogeographic history of grey wolves in Europe. BMC Evol Biol 10(1):1–11

Pilot M, Moura AE, Okhlopkov IM, Mamaev NV, Alagaili AN, Mohammed OB et al (2019) Global phylogeographic and admixture patterns in grey wolves and genetic legacy of an ancient Siberian lineage. Sci Rep 9(1):1–13

Article   CAS   Google Scholar  

Piry S, Luikart G, Cornuet JM (1999) Computer note: BOTTLENECK: a computer program for detecting recent reductions in the effective size using allele frequency data. J Hered 90(4):502–503

Plumer L, Keis M, Remm J, Hindrikson M, Jõgisalu I, Männil P et al (2016) Wolves recolonizing islands: genetic consequences and implications for conservation and management. PLoS ONE 11(7):e0158911

Posada D (2008) jModelTest: phylogenetic model averaging. Mol Biol Evol 25(7):1253–1256

Pritchard JK, Stephens M, Donnelly P (2000) Inference of population structure using multilocus genotype data. Genetics 155(2):945–959

R Core Team (2015) R version 3.2. R Project for Statistical Computing, Vienna

Raymond M, Rousset F (1995) GENEPOP (version 1.2): population genetics software for exact tests and ecumenicism. J Hered 86(3):248–249. https://doi.org/10.1093/oxfordjournals.jhered.a111573

Reed DH, Frankham R (2003) Correlation between fitness and genetic diversity. Conserv Biol 17(1):230–237

Rice WR (1989) Analyzing tables of statistical tests. Evolution 43(1):223–225

Ripple WJ, Estes JA, Beschta RL, Wilmers CC, Ritchie EG, Hebblewhite M et al (2014) Status and ecological effects of the world’s largest carnivores. Science. https://doi.org/10.1126/science.1241484

Roelke ME, Martenson JS, O’Brien SJ (1993) The consequences of demographic reduction and genetic depletion in the endangered Florida panther. Curr Biol 3(6):340–350

Ronquist F, Teslenko M, Van Der Mark P, Ayres DL, Darling A, Höhna S et al (2012) MrBayes 3.2: efficient Bayesian phylogenetic inference and model choice across a large model space. System Biol 61(3):539–542

Rozas J, Ferrer-Mata A, Sánchez-DelBarrio JC, Guirao-Rico S, Librado P, Ramos-Onsins SE, Sánchez-Gracia A (2017) DnaSP 6: DNA sequence polymorphism analysis of large data sets. Mol Biol Evol 34(12):3299–3302

Rueness EK, Asmyhr MG, Sillero-Zubiri C, Macdonald DW, Bekele A, Atickem A, Stenseth NC (2011) The cryptic African wolf: Canis aureus lupaster is not a golden jackal and is not endemic to Egypt. PLoS ONE 6(1):e16385

Sastre N, Vila C, Salinas M, Bologov VV, Urios V, Sánchez A et al (2011) Signatures of demographic bottlenecks in European wolf populations. Conserv Genet 12(3):701–712

Sharma DK, Maldonado JE, Jhala YV, Fleischer RC (2003) Ancient wolf lineages in India. Proc R Soc London Ser B 271(suppl 3):S1–S4

Srinivas Y, Jhala Y (2021) Morphometric variation in wolves and golden jackal in India (Mammalia, Carnivora). Biodivers Data J 9.

Szewczyk M, Nowak S, Niedźwiecka N, Hulva P, Špinkytė-Bačkaitienė R, Demjanovičová K et al (2019) Dynamic range expansion leads to establishment of a new, genetically distinct wolf population in Central Europe. Sci Rep 9(1):1–16

Tajima F (1989) Statistical method for testing the neutral mutation hypothesis by DNA polymorphism. Genetics 123(3):585–595

Talala MS, Bondarev AY, Zakharov ES, Politov DV (2020) Genetic differentiation of the wolf Canis lupus L populations from Siberia at microsatellite loci. Russian J Genet 56(1):59–68

Thalmann O, Shapiro B, Cui P, Schuenemann VJ, Sawyer SK, Greenfield DL et al (2013) Complete mitochondrial genomes of ancient canids suggest a European origin of domestic dogs. Science 342(6160):871–874

Van Oosterhout C, Hutchinson WF, Wills DP, Shipley P (2004) MICRO-CHECKER: software for identifying and correcting genotyping errors in microsatellite data. Mol Ecol Notes 4(3):535–538

Vanak AT, Gompper ME (2009) Dogs Canis familiaris as carnivores: their role and function in intraguild competition. Mammal Rev 39(4):265–283

Verardi A, Lucchini V, Randi E (2006) Detecting introgressive hybridization between free-ranging domestic dogs and wild wolves ( Canis lupus ) by admixture linkage disequilibrium analysis. Mol Ecol 15(10):2845–2855

Vilà C, Wayne RK (1999) Hybridization between wolves and dogs. Conserv Biol 13(1):195–198

Vilà C, Amorim IR, Leonard JA, Posada D, Castroviejo J, Petrucci-Fonseca F et al (1999) Mitochondrial DNA phylogeography and population history of the grey wolf Canis lupus. Mol Ecol 8(12):2089–2103

vonHoldt BM, Pollinger JP, Earl DA, Knowles JC, Boyko AR, Parker H, Geffen E, Pilot M, Jedrzejewski W, Jedrzejewska B, Sidorovich V, Greco C, Randi E, Musiani M, Kays R, Bustamante CD, Ostrander EA, Novembre J, Wayne RK. A genome-wide perspective on the evolutionary history of enigmatic wolf-like canids. Genome Res. 2011 Aug;21(8):1294-305. https://doi.org/10.1101/gr.116301.110 .

vonHoldt BM, Cahill JA, Fan Z, Gronau I, Robinson J, Pollinger JP et al (2016) Whole-genome sequence analysis shows that two endemic species of North American wolf are admixtures of the coyote and gray wolf. Sci Adv 2(7):e1501714

Wang MS, Wang S, Li Y, Jhala Y, Thakur M, Otecko NO et al (2020) Ancient hybridization with an unknown population facilitated high-altitude adaptation of canids. Mol Biol Evol 37(9):2616–2629

Wang MS, Thakur M, Jhala Y, Wang S, Srinivas Y, Dai SS et al (2022) Genome sequencing of a gray wolf from peninsular India provides new insights into the evolution and hybridization of gray wolves. Genome Biol Evol 14(2):evac012

Weckworth BV, Dawson NG, Talbot SL, Flamme MJ, Cook JA (2011) Going coastal: shared evolutionary history between coastal British Columbia and Southeast Alaska wolves (Canis lupus). PLoS ONE 6(5):e19582

Werhahn G, Senn H, Kaden J, Joshi J, Bhattarai S, Kusi N et al (2017) Phylogenetic evidence for the ancient Himalayan wolf: towards a clarification of its taxonomic status based on genetic sampling from western Nepal. R Soc Open Sci 4(6):170186

Werhahn G, Senn H, Ghazali M, Karmacharya D, Sherchan AM, Joshi J et al (2018) The unique genetic adaptation of the Himalayan wolf to high-altitudes and consequences for conservation. Glob Ecol Conserv 16:e00455

Wilson DE, Reeder DM (eds) (2005) Mammal species of the world: a taxonomic and geographic reference, vol 1. JHU Press

Yumnam B, Jhala YV, Qureshi Q, Maldonado JE, Gopal R, Saini S et al (2014) Prioritizing tiger conservation through landscape genetics and habitat linkages. PLoS ONE 9(11):e111207

Download references

Acknowledgements

The authors would like to thank the officials of state forest departments and zoological parks for the logistic support with sample collection. We specifically thank the sample providers without whom this study would not have been possible. Special thanks to Dhruv Jain and Soham Seal from Wildlife Institute of India for their help with GIS map and Images beautification. Our sincere thanks to the editor and an anonymous reviewer for constructive comments to improve the manuscript.

The authors declare that no funds or grants were received during the preparation of this manuscript.

Author information

Authors and affiliations.

Wildlife Institute of India, Chandrabani, Dehradun, 248001, India

Yellapu Srinivas & Yadvendradev Jhala

You can also search for this author in PubMed   Google Scholar

Contributions

YVJ and YS conceptualised the study and collected samples. YS performed the experiments, analysed the data and wrote the manuscript. YVJ supervised the study, reviewed the drafts and finalised the manuscript.

Corresponding authors

Correspondence to Yellapu Srinivas or Yadvendradev Jhala .

Ethics declarations

Conflict of interest.

The authors declare no conflict of interest.

Additional information

Publisher's note.

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Supplementary Information

Below is the link to the electronic supplementary material.

10592_2023_1553_MOESM1_ESM.jpg

Supplementary file1 (JPG 395 KB). Supplementary Figure 1: Bayesian phylogenetic relationships of the Indian peninsular and Himalayan wolves to Holarctic wolves and other Canis species based on Cyt b DNA sequences with GenBank accession numbers.

10592_2023_1553_MOESM2_ESM.tiff

Supplementary file2 (TIFF 78 KB) Supplementary Figure 2: Mismatch distributions of pairwise differences of CR haplotypes for the Indian peninsular wolves (A) and Himalayan wolves (B). Depicted are observed (dashed lines) and expected (solid lines) frequencies.

10592_2023_1553_MOESM3_ESM.tiff

Supplementary file3 (TIFF 2808 KB) Supplementary Figure 3: Population genetic structure of Indian peninsular wolves using 25 nuclear microsatellites implemented in STRUCTURE v 2.3 with no prior location information. Plot of STRUCTURE Harvester showing ΔK peaking at K=4 (A) and log-likelihood change in probability (B). Summary bar plot of STRUCTURE runs at K=2, 4, and 6 (C) showing population assignments for each individual. The populations 1, 2, 3, 4, 5 and 6 represents Gujarat, Rajasthan, Uttar Pradesh, Bihar, Maharashtra, and Karnataka respectively.

10592_2023_1553_MOESM4_ESM.tiff

Supplementary file4 (TIFF 742 KB) Supplementary Figure 4: Population genetic structure of Himalayan wolves obtained using 25 nuclear microsatellites implemented in STRUCTURE v 2.3 with no prior location model. Plot of STRUCTURE Harvester based on ΔK showing K=2 (A) and log-likelihood change in probability (B). Summary bar plot of STRUCTURE runs at K = 2, 3, and 4 (C) showing population assignments for each individual. The sampling localities 1, 2, and 3 represents Ladakh, Himachal Pradesh and North Sikkim respectively.

Supplementary file5 (DOCX 45 KB)

Rights and permissions.

Springer Nature or its licensor (e.g. a society or other partner) holds exclusive rights to this article under a publishing agreement with the author(s) or other rightsholder(s); author self-archiving of the accepted manuscript version of this article is solely governed by the terms of such publishing agreement and applicable law.

Reprints and permissions

About this article

Srinivas, Y., Jhala, Y. Genetic diversity, structure, and demographic histories of unique and ancient wolf lineages in India. Conserv Genet 25 , 33–48 (2024). https://doi.org/10.1007/s10592-023-01553-y

Download citation

Received : 16 June 2022

Accepted : 21 July 2023

Published : 07 August 2023

Issue Date : February 2024

DOI : https://doi.org/10.1007/s10592-023-01553-y

Share this article

Anyone you share the following link with will be able to read this content:

Sorry, a shareable link is not currently available for this article.

Provided by the Springer Nature SharedIt content-sharing initiative

• Phylogeography
• Phylogenetics
• Population structure
• Find a journal
• Publish with us
• Track your research

Thank you for visiting nature.com. You are using a browser version with limited support for CSS. To obtain the best experience, we recommend you use a more up to date browser (or turn off compatibility mode in Internet Explorer). In the meantime, to ensure continued support, we are displaying the site without styles and JavaScript.

• View all journals
• Explore content
• About the journal
• Publish with us
• Sign up for alerts
• Open access
• Published: 31 March 2021

Identifying unknown Indian wolves by their distinctive howls: its potential as a non-invasive survey method

• Sougata Sadhukhan 1 ,
• Holly Root-Gutteridge 2 , 3 &
• Bilal Habib 1  

Scientific Reports volume  11 , Article number:  7309 ( 2021 ) Cite this article

6560 Accesses

14 Citations

61 Altmetric

Metrics details

• Animal behaviour
• Behavioural ecology
• Conservation biology
• Ecological modelling
• Environmental sciences
• Population dynamics

Previous studies have posited the use of acoustics-based surveys to monitor population size and estimate their density. However, decreasing the bias in population estimations, such as by using Capture–Mark–Recapture, requires the identification of individuals using supervised classification methods, especially for sparsely populated species like the wolf which may otherwise be counted repeatedly. The cryptic behaviour of Indian wolf ( Canis lupus pallipes ) poses serious challenges to survey efforts, and thus, there is no reliable estimate of their population despite a prominent role in the ecosystem. Like other wolves, Indian wolves produce howls that can be detected over distances of more than 6 km, making them ideal candidates for acoustic surveys. Here, we explore the use of a supervised classifier to identify unknown individuals. We trained a supervised Agglomerative Nesting hierarchical clustering (AGNES) model using 49 howls from five Indian wolves and achieved 98% individual identification accuracy. We tested our model’s predictive power using 20 novel howls from a further four individuals (test dataset) and resulted in 75% accuracy in classifying howls to individuals. The model can reduce bias in population estimations using Capture-Mark-Recapture and track individual wolves non-invasively by their howls. This has potential for studies of wolves’ territory use, pack composition, and reproductive behaviour. Our method can potentially be adapted for other species with individually distinctive vocalisations, representing an advanced tool for individual-level monitoring.

Similar content being viewed by others

Vocal fingerprinting reveals a substantially smaller global population of the Critically Endangered cao vit gibbon ( Nomascus nasutus ) than previously thought

Supplemental structured surveys and pre-existing detection models improve fine-scale density and population estimation with opportunistic community science data

Mark-recapture validates the use of photo-identification for the widely distributed blue-spotted ribbontail ray, Taeniura lymma

Introduction.

Accurate population estimates are a critical part of wildlife biology, conservation and inform management strategies 1 . Informed management decisions rely on accurate estimates which can be hard to achieve but are critical as the conservation status of any species is dependent on its population size, which is inversely correlated with extinction risk 2 . Therefore, it is of the foremost importance to have a statistically robust population estimation technique. However, widely used population estimation methods such as camera trapping and sighting-based distance sampling fall short in analysing population trends of certain elusive species or species living in extensive home ranges 3 , 4 . Many of these species are vocally active, which inspired scientists to study the effectiveness of an acoustics-based population abundance model for these species 5 , 6 , 7 , 8 . Acoustic monitoring has long been used to monitor the presence of aquatic animals, amphibians, insects, and birds 9 , 10 , 11 , 12 , 13 . The critical advantages of acoustic monitoring are that it can be operative at day and night 14 and detect visually cryptic species or those spread over large home ranges 7 , 15 , 16 . Like camera traps, passive acoustics devices can operate throughout the day for weeks without intervention, and this perpetual data can be analysed easily with the advancement of methodologies for automating the process 17 . Recordings from these devices can be used in calculating a wide range of metrics including acoustic indices 18 , 19 , animal diversity 19 , 20 , animal localisation 21 , 22 , 23 , and density 24 , 25 estimation. This density estimation is mostly obtained through Spatially Explicit Capture-Recapture (SECR) that relies on multiple recording stations for Capture-Mark-Recapture (CMR), and instead of ‘recapture’ with time, it considers ‘redetection’ in different points in space 24 , 25 , 26 . This methodology is applied to individuals that are not identifiable from their calls 25 , 27 . The conventional CMR model requires identification at the individual level 27 , 28 , but it provides a robust population estimation 28 and much additional information such as home-range, survival rate, animal movement pattern, and population viability analysis 29 . However, the difficulty of successfully identifying unknown individuals from their calls has prevented its use for more species, though new techniques are being developed for some species, including the use of unsupervised classifiers to cluster calls 30 . Here, we explore the potential of identifying individuals through supervised classification from their vocal features to potentially improve identification to the point where CMR surveys are possible for an elusive and wide-ranging species.

Like other grey wolf subspecies, Indian grey wolves are known for their long-ranging communication via howls 31 . Howling is a social communication process, vital for the overall behaviour of many canid species 32 . It has evolved in wolves to communicate with other group members over a long distance as well as to demarcate their vast territories 33 . Due to its high amplitude and low frequency, a howl can travel for six kilometres or more 34 , 35 , 36 . Hence, an acoustics survey using howling may be more advantageous than a visual survey or other methods, such as snow-tracking 22 , 23 , 35 , 37 . As vocalisations of wolves were found to be highly variable within and among individuals 31 , 38 , the howl is a useful tool to identify individuals 39 , 40 , 41 ; thus, wolves are ideal candidates for acoustic monitoring techniques.

Previously the ‘ Howlbox ’, a self-contained active acoustics-monitoring device that broadcasts howls and records the responses automatically, was tested for its capability to detect wolves 42 , 43 . This device was unsuccessful in surveying wolves due to low detection rate as, instead of howling back, the wolves visited the device site without howling, and various technical failures 42 . A few studies using passive acoustic devices show the potentiality of successful localisation and monitoring of the grey wolf 23 , 44 . However, these only allowed for presence to be detected and stopped short of individual identification. In contrast, the identification of wolves from their distinctive howls will open an opportunity for more conventional CMR methods 45 , and this will improve population estimation without bias and help to measure other ecological variables, such as site occupancy and home-range. With the ability to identify individual wolves from howl recordings, information on population sizes, dispersal patterns, pack composition and the presence of pups could be obtained. These would be used to develop conservation management strategies and to examine population trends with howl surveys conducted over multiple years. Therefore, our study aimed to record howls from Indian wolves ( Canis lupus pallipes ) and test the feasibility of identifying unknown individuals from their howls alone using a supervised classification method.

Study species

Indian wolf, subspecies of the grey wolf is among the keystone species found in the Central Indian landscape 46 and reside in arid grasslands, floodplains, and the buffer of dense forests 46 , 47 , 48 , 49 . The Indian wolf plays a significant ecological role in controlling ungulate populations in the human-dominated landscapes 50 , 51 . The population status of Indian wolves is entirely unknown 52 . It is known that Indian wolves face a major threat from humans as their habitat is increasingly used by humans, and human-wildlife conflict is increasing 53 . Therefore, time is a critical factor to their conservation. The major challenges for population estimation of the wolf are its vast home range of ~ 230 km 2 48 and that they actively avoid camera traps because of camera sound, light, and odour emission 54 . Since implementing standard population monitoring tools in these landscapes is a tremendous challenge, monitoring their population through howls can be an essential technique. The average fundamental frequency and duration of Indian wolf howls are 422 Hz and 5.21 s, respectively 55 . Due to its low-frequency range and longer duration, it can be heard from an extended distance like howls of other subspecies 23 , 35 , 36 .

The study was conducted on captive individuals of Jaipur Zoo and free-ranging, wild wolves of Maharashtra, India.

Jaipur Zoo is situated at the heart of Jaipur City, Rajasthan, India. Since Jaipur is one of the major tourist destination and capital of Rajasthan, the anthropogenic noise is reasonably high in and around the zoo. All the wolves (n = 10) in Jaipur zoo were offspring of captive-bred individuals except one adult male recently captured from a wild population of Rajasthan.

The data of free-ranging wild wolves were collected from six districts of Maharashtra. Pune, Ahmednagar, Solapur and Osmanabad (Fig.  1 ) districts fall under the semi-arid drought-prone area of the Deccan peninsula Biogeographic Zone (Zone 6) 56 . The dominant habitat type in our sampling areas was Deccan thorn scrub forests 57 . The terrain is gently undulating with mild slopes and flat-topped hillocks with intermittent shallow valleys, which forms the primary drainage channels. Grassland area is distributed in fragmented patches, creating a mosaic of grazing land, agricultural land and human settlements. Striped hyenas ( Hyaena hyaena ), golden jackals ( Canis aureus indicus ), and Indian leopards ( Panthera pardus fusca ) are the co-predators in this landscape 48 , 58 . Wild prey include blackbucks ( Antilope cervicapra ), chinkaras ( Gazella bennettii ) and wild pigs ( Sus scrofa cristatus ); but a significant part of their diet is domestic livestock 48 , 50 , 59 .

Map showing howling recording locations of the free-ranging wolf in six districts of Maharashtra. Yellow round bullets indicate the survey locations and Red triangular bullets represent the howling recording sites.

In Maharashtra, Nagpur and Gondia districts come under the central Deccan Plateau with Tropical dry deciduous broadleaf forests 56 , 57 . Due to moderate to high rainfall, vegetation is dense in most of the areas. Our sampling areas were mostly packed with open forest and modest density forest. The terrain is generally flat. Nagpur division is surrounded by Many National parks and Sanctuaries. Wolves are primarily found in the buffer areas of these protected areas. Co-predators in those stretches are tigers ( Panthera tigris tigris ), Indian leopards, sloth bears ( Melursus ursinus ), striped hyenas, dholes ( Cuon alpinus ), and golden jackals. Prey species are sambar ( Rusa unicolor ), nilgai ( Boselaphus tragocamelus ), chital ( Axis axis ), chousingha ( Tetracerus quadricornis ), and wild pigs.

Data collection

The howls from the Indian wolves were recorded from November 2015 to July 2016. The howls were recorded during the systematic howling surveys accompanied by the opportunistic and spontaneous recordings of captive and free-ranging wolf howls. Howling surveys were done in the early morning (from 4:30 am onwards) and early evening hours (up to 7:45 pm) [time varies depending on sunrise and sunset]. The survey protocol was adapted from Harrington and Mech 60 . Each howling session consisted of five trials with three-minute intervals. A series of 50-s-long pre-recorded solo howls (from an individual in Jaipur Zoo) was played three times with increasing amplitude; the session was followed by a 50-s-long chorus howl (from three individuals in Jaipur Zoo) in the order of mid and highest amplitude level of the speaker respectively. A 40-W JBL Xtreme speaker (Harman International Industries, 2014) was used for the playbacks. If howling responses were recorded, the playback session was terminated and repeated after 15 to 20 min. All howls were recorded in a single microphone setup, using a Blue Yeti Pro USB Condenser Microphone (Blue Microphone, 2011) attached with Zoom H4N Handheld Audio Recorder (Zoom Corporation, 2009) with a sampling rate of 44.1 kHz and 16-bit depth.

Ethical approval

The study on captive wolves in zoos was done with the permission of the Director of Jaipur Zoo and the Forest Department of Rajasthan, India [Letter no- 3(04)-II/CCFWL/2013/4586–87; Dated 30th Oct 2015]. The survey of free-ranging wolves of Maharashtra was performed with the consent of the Principal Chief Conservator of Maharashtra Forest Department [Letter no- 22(8)/WL/CR-947(14–15)/1052/2015–16; Dated- 6th Aug 2015]. No animal was harmed during the study, and the standard non-invasive protocol of howling survey was maintained. All the data collection were approved by the Animal Ethics committee of Wildlife Institute of India, Dehradun, India.

Feature extraction

The howls were sorted, and spectrograms were generated using a Discrete Fourier Transform (DFT) algorithm in Raven Pro 1.5 software 61 . Discrete Fourier Transform (DFT) algorithm transforms the same length sequence of equally spaced sample points (N, where N is a prime number) with circular convolution being implemented on the points 62 . All the spectrograms were produced using Hann windows  at the rate of 1800 samples on 35.2 Hz 3 dB filter (Fig.  2 ). Only recordings with low levels of background noise and without any overlapping sounds, where the howls were clearly visible as contours, were selected for further analysis. Spectral images were digitised using Web Plot Digitizer Software 63 . Thirteen different features (Table 1 ) were measured from the digitised value by using Microsoft Excel. The details methodology is represented in Fig.  3 .

Spectrogram of Gangewadi Wolf howl (160203-001_Gangewadi2_A5) showing how different variables were measured.

The pictorial representation of methodology for identifying unknown Indian wolves by their howls.

One hundred and thirty-three howls that were longer than 5-s were used for further analysis, with more than ten individual wolves included. The 5 s cut off were chosen to avoid social squeak calls that are very similar to howl but shorter ( \({\overline{\text{x}}}\)  = 3.87 s) and high-frequency variable calls, described by Sadhukhan et al. 55 . Also, the longer howls might contain more identification features than the shorter howls do. Principal Component Analysis (PCA) was conducted on measured parameters of 133 howls to reduce the dimension and emphasise the variation between each howl. Out of 133 howls, only 69 howls were identified to an individual. The 69 howls were from nine wolves with known identities: three were captive wolves and six wild, free-ranging wolves, which were identified from their visual features when they were howling in front of the observer and thus howls could be attributed to them individually. The data was further subdivided into training and test datasets. Forty-nine howls from five individuals (2 captives; 3 wild) were used as the training data, and 20 howls from four different individuals (1 captive, 3 wild) as test data to ensure the validity of the method (Table 2 ). Since the known wolf howls were used test data never used in building model, it provides ‘unbiased sense of model effectiveness’ 64 .

Discriminant function analysis

Linear discriminant function analysis (DFA) was performed with 49 howls from five individuals (training data) using seven PCA values that contributed more than 5% variation (Table 3 ) [The cut off value was chosen from scree plot, See Supp. Material 1 : PCA.pdf]. The objective of DFA was to construct the linear combination of independent principal component variables (PC1–PC7) that will discriminate howls of different individuals. The howls were plotted with discriminant functions at two-dimensional space followed by the group prediction (Fig.  4 ).

Figure showing a two-dimensional plot of discriminant function analysis using LD1 (Linear Discriminant) and LD2. Each colour represents each wolf. 100% accuracy was achieved in identifying 49 howls from five Indian wolves.

Hierarchical clustering

To test the success rate of identifying different individuals from their howls with Linear Discriminant (LD) score, an Agglomerative Nesting hierarchical clustering (AGNES) was executed on 49 howls (training data) that were used in DFA. AGNES initially considers each howl as a different cluster and use a ‘ bottom-up ’ algorithm to join different clusters based on the similarities 65 . The analysis was performed in R using ‘ agnes ’ function in the package ‘dendextend’ and ‘ manhattan ’ metric was used to build the cluster 66 . The same analysis was performed on the test data to determine the accuracy of identifying unknown individuals and estimating the number of wolves from their howls. While the test data contained howls from known individuals, the wolves’ identities were not included in the model. The variables of these 20 howls were calculated from the equation of DFA of 49 howls for cluster analysis.

Dimensions reduction to emphasis on variation among howls

Seven Principal Components (PC) that explained more than five percent of the variance (Table 3 ) each were generated from 13 scalar variables (Table 1 ). These seven PCs together explained 94.8% variance among different howls (Fig.  5 ). SD of the fundamental frequency (f 0 ), Frequency (f 0 ) range, Maximum f 0 and the number of abrupt change (> 25 Hz) were the most important contributing factors for building PC1 which contributed 41.2% explaining the variable (Fig.  5 ).

The spider web bubble plot is describing how the Simple Scalar Variables (SSV) are ultimately contributing to two LD functions through PC values. The bubble size of each SSV represents the contribution for building each PC function. The blue line represents LD 1, and Orange represents LD2. Since PC1 and PC2 contribute 85% for LD1, the most important SSVs are Stdv f 0 , Min f 0 , Max f 0 and Mean f 0 . Similarly Duration, Abrupt changes, Co-fv contribute the most in building the LD2 function via PC4 and PC5. LD1 was best defined by the different fundamental frequency factors, while LD2 was best defined through the shape of the frequency contour. Therefore, the critical factors for individuality were encoded in X and Y variables.

Building discriminant function to emphasis on howl variation among different individuals

The objective of DFA was to build an equation that discriminates the howls of different individuals. The LD score also highlights the variation among howls from different individuals. DFA achieved 100% accuracy in identifying five individuals from 49 howls (Fig.  4 ). As the first two Linear Discriminants (LD1 and LD2) were responsible for 96.2% of the variance to differentiate between howls of different individuals (LD1 = 87.57% and LD2 = 8.63%), we calculated LD1 and LD2 for rest of the howls using the same function (equation) from last DFA. PC1 and PC2 contributed 85% in building LD1; PC4 and PC5 are the most crucial factor (65%) for LD2 function (Fig.  5 ).

Identifying individuals from their howls in testing dataset

First, we tested AGNES on the training dataset (49 howls from 5 individuals) and found 48 howls (~ 97.9% accuracy) were identified correctly at 2.2 clustering scale (Fig.  6 ). When the same analysis was performed on 20 howls of four different individuals to test the accuracy for the non-training dataset, 15 out of 20 howls from (75% accuracy) four individuals were identified correctly at 2.2 clustering scale (Fig.  7 ; Table 4 ). Two howls from wolf BMT.A were misclassified to wolves BMT.SA2 and CG2.A2; Three howls from wolf NU.A were misclassified to wolves BMT.SA2 (1 howl) and CG2.A2 (2 howls) (Fig.  7 ; Table 4 ).

Hierarchical Clustering of 49 howls from five individuals. These 49 howls were used in training the data. 48 howls were identified correctly with the accuracy of 97.9%. The wrongly identified howl is marked in red.

Hierarchical Clustering of 20 howls from four Indian wolves. None of the 20 howls was used in training the data. 15 howls were identified correctly with the accuracy of 75%, and all the four individuals were identified correctly as different clusters. The correctly identified howls are marked in black, and the five wrongly identified howls are marked in red.

Here, we presented a new approach to train the classification model, which can identify individuals from their howls and determine the number of wolves present in a certain number of howls, allowing for fine-scale population surveys. In this study, we built an identification model with known training data which was verified with novel test data. The testing data included howls from the known individuals of both captive and wild Indian wolves but independent from the training dataset so that we can cross-check the identification accuracy without bias. The key finding of our study was 97.9% wolf howls were identified correctly from training data, whereas the accuracy of the model on the testing data was 75%. Moreover, we were able to identify four individuals accurately from the testing dataset. The primary significance of this study is that it can be replicated for any other wolf sub-species with a set of a known wolf howls. This study increases the feasibility of wolf pack census using a howling survey 35 , 60 . Since wolves may actively avoid camera traps 54 and photo-identification of wolf requires arduous effort 3 , 67 , identifying wolves from their howls is a big step towards population estimation using CMR.

Although CMR associated with camera trapping provides population estimation without bias for an identifiable animal like a tiger 68 , camera trapping has several limitations for non-identifiable and long-ranging species like the wolf 3 . Other non-invasive methods like DNA-based CMR resulted in biased population estimation due to the animals’ non-uniform scent-marking patterns 59 , 69 . However, acoustics based surveys allow vast area sampling with limited resources as compared to camera trapping and other non-invasive methods 3 . Furthermore, our field observations of wolves have shown that the whole pack typically howls during choruses and that all individuals are acoustically active.

For population size estimation through an acoustics-based survey, a combination of CMR and Distance Sampling is required to reduce bias and heterogeneity in detection probability 27 , 70 . Identifying individual wolves from their howls close this gap of implementing the CMR technique for the population assessment of this elusive and challenging to track species 7 , 25 , 27 . While a few studies have established that howls carry individuality information 38 and known howls can be distinguished from each other 39 , 45 , 71 , no study has been successful before in identifying unknown individuals from a set of howls. Furthermore, attempts to count the number of individuals present in a recording have been limited by difficulties in minimising confidence intervals 18 , 72 . There are two ways to identify individual wolves or packs—supervised clustering and unsupervised clustering. While supervised clustering requires a set of known training data and cluster validation is straightforward, unsupervised clustering requires ground-truthing before it can be used to monitor populations at a survey level and does not allow individual level CMR or tracking 30 .

Although DNA-based identification from faecal sampling is more accurate in identifying individuals than our result, it has drawbacks, such as biased population estimation and the increased cost and effort required to collect and analyse the faeces 59 , 69 . Nevertheless, the acoustics-based identification model requires further work to increase its accuracy, though we believe that the successful implementation of this method as a CMR-based supervised population estimation model is already possible.

Wolves mostly live in packs that habitually howl together, and it is challenging to identify the specific wolf that is howling, particularly in choruses. If included and incorrectly attributed to a particular wolf, these howls could lead to erroneous predictions by the model. Therefore, this limited our potential data set to those howls which were conclusively attributed to a known individual, and we dropped many howls, especially the chorus howls, from the analysis to avoid misleading the model. However, larger training datasets from different wolf populations might increase the efficacy of the identification model and verification with more wolf howls conceding better reliability as found for Southwestern Willow Flycatcher 73 . Thus, our result of 75% may represent a baseline, not a limit, on the accuracy we could achieve. The inclusion of multiple series of howls from every individual would give a more precise result. However, since none of the free-ranging wolves was radio-collared or marked, this was not possible for the wild wolves. Studying howls of collared wolves would help in adding multiple howl sequences from many free-ranging wolves in the training data and may fill this research gap.

This study revealed that the number of wolves present in the recordings could be determined from their howls and the individuality information is sufficient for supervised population estimation through CMR techniques 7 , 25 , 27 , 30 . Therefore, wolves recorded in one location can be acoustically recaptured at another location, and we can identify them individually. Since our model is exclusively built on fundamental frequency, changes in terrain or vegetation should not affect the accuracy of the model. The information gained from recapturing wolves across different locations would help in deriving territoriality (home-range) information, and this information is crucial for spatially explicit individual-based point process models. This is a clear advancement for developing howling playback surveys as a wolf pack census method. Regular population monitoring will help towards conserving and saving this cryptic species before its population falls beyond a recovery level. Furthermore, since wolf howls can be detected across distances of more than 6 km, identifying wolves from their howls also opens up a new opportunity for non-invasive tracking of this species across large landscapes.

Guidelines to implement the methodology on the field

We used this methodology to identify individual Indian wolf howl. However, one can use this methodology to identify species, sub-species or individual from their calls. This requires a set of calls to make up the training dataset and a set of calls to make up the testing dataset. We recommend some precautions and step by step guidelines for adapting this method.

Before the data collection, one should be cautious about choosing the recorder and data collection methodology. Although we are not definite about the impact of multi-recorder setup in identification accuracy, we recommend using a single microphone set up to keep consistency, especially for individual identification as differences in sensitivity and recording parameters can influence acoustic integrity [See 45 ].

The multiple groups in the training dataset should be carefully selected to represent distinct group member calls with high confidence (e.g. species/sub-species/individuals), as a single incorrectly identified call in the training dataset can lead the model to erroneous results.

The selection of appropriate spectral features is important. While many species encode their identity in the same features, some encoding is species-specific. We tested a wide range of software which fell short in feature extraction for overlapping calls or where background noise was present. The feature description is only as reliable as the extraction. Here, we used web-plot digitiser software for spectrogram digitisation. We recommend the use of any semi-automated graph digitiser tool for noisy or overlapping spectral data.

The training data should contain only known groups (multi-species/multiple sub-species/multiple individuals). Each training group should have at least three to five calls and recordings from multiple sessions will increase the accuracy of the model as the animals may have higher intra-individual variation across days than within them. Thus, the higher the intra-individual or intra-group variation, the greater the number of vocalisations and individuals that should be included in the training dataset to make a robust model for the testing dataset.

Even though one can choose an unknown dataset as test data, we recommend using a known dataset when originally validating the model. Using multiple test datasets will increase the model’s confidence.

We recommend using multiple small batches as test data (50–100 sample of calls) instead of large data to avoid confusion in cluster groups that may represent other variation in the calls.

To allow study replication, we have made our data and codes available in the Supplementary Materials. While the data needs to be replaced for each study, the system of analysis and classification should be robust and replicable.

Our study reached substantial accuracy in identifying wolves from their howls. Since the methodology was validated using known wolf data and was found to be reasonably reliable, unknown howls can also be classified. This opens up a new opportunity for population estimation and tracking of wolves through howling surveys. Although we analysed our data with Indian wolf howls, the procedure is replicable for other subspecies that have a set of known howls from different individuals and could potentially be applied to other species with individually distinctive vocalisations. This would refine and improve both population estimates and the ability to monitor individuals in situ, with global implications for conservation and ecology.

Buckland, S. T., Anderson, D. R., Burnham, K. P. & Laake, J. L. Introductory concepts. In Distance Sampling. Estimating Abundance of Biological Populations 446 (1993). https://doi.org/10.1002/9780470752784.part1 .

Mace, G. M. et al. Quantification of extinction risk: IUCN’s system for classifying threatened species. Conserv. Biol. 22 , 1424–1442 (2008).

Article   PubMed   Google Scholar  

Garland, L., Crosby, A., Hedley, R., Boutin, S. & Bayne, E. Acoustic vs. Photographic monitoring of gray wolves ( Canis lupus ): a methodological comparison of two passive monitoring techniques. Can. J. Zool. 98 , 219–228 (2020).

Article   Google Scholar  

Crunchant, A. S., Borchers, D., Kühl, H. & Piel, A. Listening and watching: do camera traps or acoustic sensors more efficiently detect wild chimpanzees in an open habitat?. Methods Ecol. Evol. 11 , 542–552 (2020).

Wood, C. M. et al. Using the ecological significance of animal vocalizations to improve inference in acoustic monitoring programs. Conserv. Biol. https://doi.org/10.1111/cobi.13516 (2020).

Article   PubMed   PubMed Central   Google Scholar  

Rhinehart, T. A., Chronister, L. M., Devlin, T. & Kitzes, J. Acoustic localization of terrestrial wildlife: current practices and future opportunities. Ecol. Evol. 10 , 6794–6818 (2020).

Kidney, D. et al. An efficient acoustic density estimation method with human detectors applied to gibbons in Cambodia. PLoS ONE 11 , 1–16 (2016).

Article   CAS   Google Scholar  

Thompson, M. E., Schwager, S. J., Payne, K. B. & Turkalo, A. K. Acoustic estimation of wildlife abundance: methodology for vocal mammals in forested habitats. Afr. J. Ecol. 48 , 654–661 (2010).

Google Scholar  

Parra, J. M. Passive acoustic aquatic animal finder apparatus and method. US patent 5,099,455 (1992).

Riede, K. Acoustic monitoring of Orthoptera and its potential for conservation. J. Insect Conserv. 2 , 217–223 (1998).

Petrusková, T., Pišvejcová, I., Kinštová, A., Brinke, T. & Petrusek, A. Repertoire-based individual acoustic monitoring of a migratory passerine bird with complex song as an efficient tool for tracking territorial dynamics and annual return rates. Methods Ecol. Evol. 7 , 274–284 (2016).

Sanders, C. E. & Mennill, D. J. Acoustic monitoring of nocturnally migrating birds accurately assesses the timing and magnitude of migration through the Great Lakes. Condor 116 , 371–383 (2014).

Acevedo, M. A. & Villanueva-Rivera, L. J. From the field: Using automated digital recording systems as effective tools for the monitoring of birds and amphibians. Wildl. Soc. Bull. 34 , 211–214 (2006).

Wrege, P. H., Rowland, E. D., Keen, S. & Shiu, Y. Acoustic monitoring for conservation in tropical forests: examples from forest elephants. Methods Ecol. Evol. 8 , 1292–1301 (2017).

Pérez-Granados, C. et al. Vocal activity rate index: a useful method to infer terrestrial bird abundance with acoustic monitoring. Ibis (Lond. 1859) 161 , 901–907 (2019).

Kimura, S. et al. Comparison of stationary acoustic monitoring and visual observation of finless porpoises. J. Acoust. Soc. Am. 125 , 547–553 (2009).

Article   ADS   PubMed   Google Scholar  

Gibb, R., Browning, E., Glover-Kapfer, P. & Jones, K. E. Emerging opportunities and challenges for passive acoustics in ecological assessment and monitoring. Methods Ecol. Evol. https://doi.org/10.1111/2041-210X.13101 (2018).

Papin, M., Aznar, M., Germain, E., Guérold, F. & Pichenot, J. Using acoustic indices to estimate wolf pack size. Ecol. Indic. 103 , 202–211 (2019).

Depraetere, M. et al. Monitoring animal diversity using acoustic indices: implementation in a temperate woodland. Ecol. Indic. 13 , 46–54 (2012).

Wheeldon, A., Mossman, H. L., Mathenge, J. & De Kort, S. R. Comparison of acoustic and traditional point count methods to assess bird diversity and composition in the Aberdare National. Afr. J. Ecol. https://doi.org/10.1111/aje.12596 (2019).

Wilson, S. J. & Bayne, E. M. Use of an acoustic location system to understand how presence of conspecifics and canopy cover influence Ovenbird ( Seiurus aurocapilla ) space use near reclaimed wellsites in the boreal forest of Alberta. Avian Conserv. Ecol. https://doi.org/10.5751/ACE-01248-130204 (2018).

Gable, T. D., Windels, S. K. & Bump, J. K. Finding wolf homesites: improving the efficacy of howl surveys to study wolves. PeerJ 6 , e5629 (2018).

O’Gara, J. R. et al. Efficacy of acoustic triangulation for gray wolves. Wildl. Soc. Bull. https://doi.org/10.1002/wsb.1089 (2020).

Dawson, D. K. & Efford, M. G. Bird population density estimated from acoustic signals. J. Appl. Ecol. 46 , 1201–1209 (2009).

Stevenson, B. C. et al. A general framework for animal density estimation from acoustic detections across a fixed microphone array. Methods Ecol. Evol. 6 , 38–48 (2015).

Royle, J. A., Chandler, R. B., Sollmann, R. & Gardner, B. Spatial Capture–Recapture (Academic Press, 2013).

Marques, T. A. et al. Estimating animal population density using passive acoustics. Biol. Rev. 88 , 287–309 (2013).

Adi, K., Johnson, M. T. & Osiejuk, T. S. Acoustic censusing using automatic vocalization classification and identity recognition. J. Acoust. Soc. Am. 127 , 874–883 (2010).

Lettink, M. & Armstrong, D. P. An introduction to using mark-recapture analysis for monitoring threatened species. Dep. Conserv. Tech. Ser. 28A , 5–32 (2003).

Clink, D. J. & Klinck, H. Unsupervised acoustic classification of individual gibbon females and the implications for passive acoustic monitoring. Methods Ecol. Evol. 1 , 1–2 (2020).

Theberge, J. B. & Falls, J. B. Howling as a means of communication in timber wolves. Am. Zool. 7 , 331–338 (1967).

Kershenbaum, A. et al. Disentangling canid howls across multiple species and subspecies: structure in a complex communication channel. Behav. Process. 124 , 149–157 (2016).

Harrington, F. H. & Mech, D. L. Wolf howling and its role in territory maintenance. Behaviour 68 , 207–249 (1978).

Joslin, P. Summer Activities of Two Timber Wolf (Canis lupus) Packs in Algonquin Park (University of Toronto, 1966).

Suter, S. M., Giordano, M., Nietlispach, S., Apollonio, M. & Passilongo, D. Non-invasive acoustic detection of wolves. Bioacoustics 4622 , 1–12 (2016).

Harrington, F. H. & Mech, D. L. Wolf vocalization. In Wolf and man , 109–132 (Elsevier, 1978). https://doi.org/10.1016/B978-0-12-319250-9.50014-1 .

Blanco, J. C. & Cortés, Y. Surveying wolves without snow: a critical review of the methods used in Spain. Hystrix 23 , 35–48 (2012).

Tooze, Z. J., Harrington, F. H. & Fentress, J. C. Individually distinct vocalizations in timber wolves, Canis lupus . Anim. Behav. 40 , 723–730 (1990).

Root-Gutteridge, H. et al. Improving individual identification in captive Eastern grey wolves ( Canis lupus lycaon ) using the time course of howl amplitudes. Bioacoust. Int. J. Anim. Sound Rec. 23 , 39–53 (2014).

Hull, C., McCombe, C. & Dassow, A. Acoustic identification of wild gray wolves, Canis lupus , using low quality recordings. Am. J. Undergrad. Res. 16 , 41–49 (2020).

Wasser, S. K., Smith, H., Madden, L., Marks, N. & Vynne, C. Scent-matching dogs determine number of unique individuals from scat. J. Wildl. Manag. 73 , 1233–1240 (2009).

Brennan, A., Cross, P. C., Ausband, D. E., Barbknecht, A. & Creel, S. Testing automated howling devices in a wintertime wolf survey. Wildl. Soc. Bull. 37 , 389–393 (2013).

Ausband, D. E., Skrivseth, J. & Mitchell, M. S. An automated device for provoking and capturing wildlife calls. Wildl. Soc. Bull. 35 , 498–503 (2011).

Papin, M., Pichenot, J., Guérold, F. & Germain, E. Acoustic localization at large scales: a promising method for grey wolf monitoring. Front. Zool. 15 , 1–10 (2018).

Root-Gutteridge, H. et al. Identifying individual wild Eastern grey wolves ( Canis lupus lycaon ) using fundamental frequency and amplitude of howls. Bioacoust. Int. J. Anim. Sound Rec. 23 , 55–66 (2014).

Singh, M. & Kumara, H. N. Distribution, status and conservation of Indian gray wolf ( Canis lupus pallipes ) in Karnataka, India. J. Zool. 270 , 164–169 (2006).

Jhala, Y. V. & Giles, R. H. The status and conservation of the wolf in Gujarat and Rajasthan, India. Conserv. Biol. 5 , 476–483 (1991).

Habib, B. Ecology of Indian wolf [ Canis lupus pallipes sykes. 1831), and modeling its potential habitat in the great Indian bustard sanctuary, Maharashtra, India (Aligarh Muslim University, Aligarh, India, 2007).

Dey, S., Sagar, V., Dey, S. & Choudhary, S. K. 2 Sight record of the Indian wolf Canis lupus pallipes in the river Gandak floodplains. J. Bombay Nat. Hist. Soc. 107 , 51 (2010).

Jethva, B. D. & Jhala, Y. V. Foraging ecology, economics and conservation of Indian wolves in the Bhal region of Gujarat, Western India. Biol. Conserv. 116 , 351–357 (2004).

Jethva, B. D. & Jhala, Y. V. Computing biomass consumption from prey occurrences in Indian wolf scats. Zoo Biol. 23 , 513–520 (2004).

Jhala YV. Human conflict in India. In “Beyond: Realties of Global Wolf Restoration” Symposium February, 23–26 (2020).

Habib, B. & Kumar, S. D. shifting by wolves in semi-wild landscapes in the Deccan Plateau, Maharashtra, India. J. Zool. 272 , 259–265 (2007).

Meek, P. D. et al. Camera traps can be heard and seen by animals. PLoS ONE 9 , e110832 (2014).

Article   ADS   PubMed   PubMed Central   CAS   Google Scholar  

Sadhukhan, S., Hennelly, L. & Habib, B. Characterising the harmonic vocal repertoire of the Indian wolf ( Canis lupus pallipes ). PLoS ONE 14 , e0216186 (2019).

Article   CAS   PubMed   PubMed Central   Google Scholar  

Rodgers, W. A. & Panwar, S. H. Biogeographical classification of India. New For. Dehra Dun, India (1988).

Reddy, C. S., Jha, C. S., Diwakar, P. G. & Dadhwal, V. K. Nationwide classification of forest types of India using remote sensing and GIS. Environ. Monit. Assess. 187 , 777 (2015).

Article   PubMed   CAS   Google Scholar  

Majgaonkar, I. et al. Land-sharing potential of large carnivores in human-modified landscapes of western India. Conserv. Sci. Pract. 1 , e34 (2019).

Morin, D. J., Kelly, M. J. & Waits, L. P. Monitoring coyote population dynamics with fecal DNA and spatial capture-recapture. J. Wildl. Manag. 80 , 824–836 (2016).

Harrington, F. H. & Mech, D. L. An analysis of howling response parameters useful for wolf pack censusing. J. Wildl. Manag. 46 , 686–693 (1982).

Bioacoustics Research Program. Raven Pro: interactive sound analysis software. The Cornell Lab of Ornithology (2014).

Rader, C. M. Discrete Fourier transforms when the number of data samples is prime. Proc. IEEE 56 , 1107–1108 (1968).

Rohatgi, A. WebPlotDigitizer. (2017).

Kuhn, M. et al. Applied Predictive Modeling Vol. 26 (Springer, 2013).

Book   MATH   Google Scholar  

Kaufman, L. & Rousseeuw, P. J. Agglomerative nesting (Program AGNES). In Finding Groups in Data 199–252 (Wiley, 2009).

Galili, T. dendextend: an R package for visualizing, adjusting, and comparing trees of hierarchical clustering. Bioinformatics https://doi.org/10.1093/bioinformatics/btv428 (2015).

Galaverni, M. et al. Monitoring wolves ( Canis lupus ) by non-invasive genetics and camera trapping: a small-scale pilot study. Eur. J. Wildl. Res. 58 , 47–58 (2012).

Jhala, Y. V, Qureshi, Q. & Nayak, A. K. Status of tigers, co-predators and prey in India 2018: summary report. (2019).

López-Bao, J. V. et al. Toward reliable population estimates of wolves by combining spatial capture-recapture models and non-invasive DNA monitoring. Sci. Rep. 8 , 1–8 (2018).

Laake, J. L. & Borchers, D. L. Methods for incomplete detection at distance zero. Advance in Distance Sampling (eds Buckland, S. T., Andersen, D. R., Burn, K. P., Laake, J. L. & Thomas, L.) 108–189 (2004).

Palacios, V., Font, E. & Márquez, R. Iberian wolf howls: acoustic structure, individual variation, and a comparison with North American populations. J. Mammal. 88 , 606–613 (2007).

Passilongo, D., Mattioli, L., Bassi, E., Szabó, L. & Apollonio, M. Visualizing sound: counting wolves by using a spectral view of the chorus howling. Front. Zool. 12 , 12–22 (2015).

Fernández-Juricic, E., del Nevo, A. J. & Poston, R. Identification of individual and population-level variation in vocalizations of the endangered Southwestern Willow Flycatcher ( Empidonax traillii extimus ). Auk 126 , 89–99 (2009).

Download references

Acknowledgements

We want to thank the authority and staffs of the Jaipur Zoo and Maharashtra Forest Department for permitting to conduct this research. We sincerely acknowledge the funding agencies, the Department of Science and Technology, the Govt of India (No. EMR/2015/000036) and the Forest Department of Maharashtra (No. 1852). We appreciate all our field personals (Daut Shaikh, Shivkumar More, Sarang Mhamane) and wildlife enthusiast groups (Pune Wolfgang, Mihir Godbole, Vineet Arora, R. V. Kasar, Rajesh Pardeshi, Sawan Behkar and others) from Maharashtra who helped in local information gathering and various logistic arrangement during data collection. The first author is thankful for the effort of Shivam Shrotriya assisting during the initial analysis. The authors are grateful to Dr Arik Kershenbaum, foe his consistent support. We are delighted for having continuous support from the Wildlife Institute of India, Dehradun and our lab members.

This research was funded by Maharashtra Forest Department ( http://mahaforest.gov.in ) (Grant No. 1852) and Department of Science and Technology, Govt of India ( http://www.dst.gov.in/ ) (Grant No. EMR/2015/000036). BH was the principal investigator of both the project, and SS was the researcher in those projects. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.

Author information

Authors and affiliations.

Animal Ecology and Conservation Biology, Wildlife Institute of India, Dehradun, 248001, India

Sougata Sadhukhan & Bilal Habib

Animal Behaviour, Cognition and Welfare Group, University of Lincoln, Lincoln, UK

Holly Root-Gutteridge

Reby Lab, School of Psychology, University of Sussex, Brighton, UK

You can also search for this author in PubMed   Google Scholar

Contributions

B.H. conceptualised the study. S.S. collected all the data and did data extraction, analysis and writing the manuscript. H.R.G. and B.H. both supervised in data interpretation along with the manuscript writing. B.H. played a sole role in funding acquisition. All authors contributed critically to the drafts and gave final approval for publication.

Corresponding author

Correspondence to Bilal Habib .

Ethics declarations

Competing interests.

The authors declare no competing interests.

Additional information

Publisher's note.

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Supplementary Information

Supplementary information 1., supplementary information 2., supplementary information 3., supplementary information 4., appendices: supplementary materials.

All the data and R code require to recreate the analysis are hosted in https://github.com/bhlabwii/wolf_howlID platform. Raw sound files are available on request to the corresponding author. Compiled reports from R Scripts can be found in following supporting material:

Rights and permissions

Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/ .

Reprints and permissions

About this article

Cite this article.

Sadhukhan, S., Root-Gutteridge, H. & Habib, B. Identifying unknown Indian wolves by their distinctive howls: its potential as a non-invasive survey method. Sci Rep 11 , 7309 (2021). https://doi.org/10.1038/s41598-021-86718-w

Download citation

Received : 27 December 2020

Accepted : 19 March 2021

Published : 31 March 2021

DOI : https://doi.org/10.1038/s41598-021-86718-w

Share this article

Anyone you share the following link with will be able to read this content:

Sorry, a shareable link is not currently available for this article.

Provided by the Springer Nature SharedIt content-sharing initiative

This article is cited by

Using the birdnet algorithm to identify wolves, coyotes, and potentially their interactions in a large audio dataset.

• Daniel Sossover
• Kelsey Burrows
• Connor M. Wood

Mammal Research (2024)

The potential for acoustic individual identification in mammals

• Pavel Linhart
• Mathieu Mahamoud-Issa
• Daniel T. Blumstein

Mammalian Biology (2022)

By submitting a comment you agree to abide by our Terms and Community Guidelines . If you find something abusive or that does not comply with our terms or guidelines please flag it as inappropriate.

Quick links

• Explore articles by subject
• Guide to authors
• Editorial policies

Sign up for the Nature Briefing: Anthropocene newsletter — what matters in anthropocene research, free to your inbox weekly.

An official website of the United States government

The .gov means it’s official. Federal government websites often end in .gov or .mil. Before sharing sensitive information, make sure you’re on a federal government site.

The site is secure. The https:// ensures that you are connecting to the official website and that any information you provide is encrypted and transmitted securely.

• Publications
• Account settings

Preview improvements coming to the PMC website in October 2024. Learn More or Try it out now .

• Advanced Search
• Journal List
• Genome Biol Evol
• v.14(2); 2022 Feb

Genome Sequencing of a Gray Wolf from Peninsular India Provides New Insights into the Evolution and Hybridization of Gray Wolves

Ming-shan wang.

1 Howard Hughes Medical Institute, University of California Santa Cruz, USA

2 Department of Ecology and Evolutionary Biology, University of California Santa Cruz, USA

Mukesh Thakur

3 Zoological Survey of India, New Alipore, Kolkata, West Bengal, India

4 State Key Laboratory of Genetic Resources and Evolution, Kunming Institute of Zoology, Chinese Academy of Sciences, Kunming, China

Yadvendradev Jhala

5 Wildlife Institute of India, Chandrabani, Dehradun, Uttarakhand, India

Yellapu Srinivas

Shan-shan dai, zheng-xi liu.

6 College of Animal Science, Jilin University, Changchun, China

Hong-Man Chen

7 College of Animal Science and Technology, Yunnan Agricultural University, Kunming, China

Richard E Green

8 Department of Biomolecular Engineering, University of California Santa Cruz, USA

Klaus-Peter Koepfli

9 Smithsonian-Mason School of Conservation, George Mason University, USA

10 Center for Species Survival, Smithsonian Conservation Biology Institute, National Zoological Park, Washington, District of Columbia, USA

11 Computer Technologies Laboratory, ITMO University, St. Petersburg, Russia

Beth Shapiro

Associated data.

The genome sequencing raw reads were deposited in the NCBI-SRA database, under Bio-Project accession: PRJNA714797.

The gray wolf ( Canis lupus ) is among the few large carnivores that survived the Late Pleistocene megafaunal extinctions. Thanks to their complex history of admixture and extensive geographic range, the number of gray wolf subspecies and their phylogenetic relationships remain poorly understood. Here, we perform whole-genome sequencing of a gray wolf collected from peninsular India that was phenotypically distinct from gray wolves outside India. Genomic analyses reveal that the Indian gray wolf is an evolutionarily distinct lineage that diverged from other extant gray wolf lineages ∼110 thousand years ago. Demographic analyses suggest that the Indian wolf population declined continuously decline since separating from other gray wolves and, today, has exceptionally low genetic diversity. We also find evidence for pervasive and mosaic gene flow between the Indian wolf and African canids including African wolf, Ethiopian wolf, and African wild dog despite their current geographical separation. Our results support the hypothesis that the Indian subcontinent was a Pleistocene refugium and center of diversification and further highlight the complex history of gene flow that characterized the evolution of gray wolves.

Significance

The gray wolf ( Canis lupus ) is one of the few megafaunal carnivores that survived the Late Pleistocene megafaunal extinctions. Despite extensive research on living and extinct gray wolves, the evolutionary history of this lineage remains unclear. Here, we sequence and analyze a draft genome of a gray wolf collected from peninsular India. We find that the Indian wolf lineage, which is both highly threatened and phenotypically distinct from other gray wolves, is the most deeply diverging lineage of extant gray wolves and, despite their physical isolation from other wolf lineages, has a long history of gene flow with other canid lineages.

Introduction

The gray wolf ( Canis lupus ) first appears in the fossil records of Eurasia and North America some 500,000 years ago ( Nowak 1979 ) and later diversified into more than 37 named subspecies ( Wilson and Reeder 2005 ). Numerous morphological and genomic analyses of gray wolves have presented a complex and sometimes contradictory view of their evolutionary history ( Leonard et al. 2007 ; Sinding et al. 2018 ; Smeds et al. 2019 ; Loog et al. 2020 ). For example, analyses of mitochondrial DNA have revealed a lack of strong haplotype structure among populations across the Northern hemisphere ( Thalmann et al. 2013 ; Loog et al. 2020 ), whereas nuclear genomic analyses have identified distinct lineages in Eurasia and North America ( Fan et al. 2016 ; Gopalakrishnan et al. 2018 ). These studies have also revealed widespread admixture among domestic dogs, gray wolves, and other species in the genera Canis and Cuon ( Freedman et al. 2014 ; Fan et al. 2016 ; Gopalakrishnan et al. 2018 ; Pilot et al. 2019 ). This evolutionary history of dynamic long-distance dispersal, population replacements, and cross-species gene flow has complicated efforts to understand both how gray wolf populations are related to each other and the location, origin, and timing of dog domestication ( Koepfli et al. 2015 ; Perri et al. 2021 ).

Among the least studied populations of gray wolves are those that inhabit the Indian subcontinent. Early taxonomists described two species endemic to this region: the Himalayan wolf, Canis laniger ( Hodgson 1847 ), found in the highland regions of the Tibetan Plateau and eastern Kashmir, and the Indian wolf Canis pallipes ( Sykes 1831 ), distributed within the arid/semi-arid lowland plains of peninsular India. Since these first descriptions, Himalayan and Indian wolves have been reclassified as subspecies within the gray wolf complex, Canis lupus chanco and C. l. pallipes , respectively ( Allen 1938 ). The current range of C. l. pallipes extends from the eastern Mediterranean region of western Asia eastward to peninsular India, where several isolated populations are reported ( Nowak 1995 ; Jhala 2003 ).

Genetic studies using mitochondrial and nuclear markers have shown that the Himalayan wolf is distinct from other gray wolf populations ( Aggarwal et al. 2007 ; Ersmark et al. 2016 ; Werhahn et al. 2017 , 2018 ). Similarly, Indian wolves are morphologically, behaviorally, and genetically distinct from other wolf subspecies ( Aggarwal et al. 2003 ; Sharma et al. 2004 ; Aggarwal et al. 2007 ). Compared with other wolves, Indian wolves are smaller in size (18–22 kg) with less and relatively shorter hair that is light brown in color with black hair tips ( fig. 1 A ). Indian wolves are also among the most threatened canid subspecies in the world, with an estimated population size of ∼2,000–3,000 individuals ( Jhala 2003 ).

Sampling location and mitochondrial phylogeny. ( A ) A photograph of an Indian wolf from peninsular India (provided by Y. Shah). ( B ) Map showing the distribution of samples used in this study. The red dot depicts the location where the Indian wolf IW01 ( supplementary fig. S1 , Supplementary Material online ) was sampled. ( C ) Maximum-likelihood tree estimated from mitochondrial genomes (15,462 bp). The Indian wolf (IW01) is a sister clade to domestic dogs and other gray wolves but inside the lineage of Tibetan wolf+Himalayan wolf. IDs in brackets are the GenBank accession numbers.

More recently, relationships among gray wolves have been analyzed using whole-genome sequences. In one study examining admixture among gray wolves and domestic dogs ( Fan et al. 2016 ), a wolf presumed to originate from India but lacking precise locality information (NCBI accession: SRS661487) clustered with wolves from Iran and Israel, which together were grouped within a larger cluster of gray wolves from Eurasia. This result was, however, at odds with earlier phylogenetic analyses based on mitochondrial sequences that suggested that Himalayan wolves and wolves from peninsular India are the earliest branchings and most divergent lineages among all gray wolf populations, with Indian wolves diverging from other lineages ∼270–400 thousand years ago (ka) ( Aggarwal et al. 2003 ; Sharma et al. 2004 ; Aggarwal et al. 2007 ). To resolve this inconsistency, additional analyses using wolf samples of unambiguous provenance are necessary, in particular as the complex history of admixture among canids can lead to discordance among individual gene trees (mitochondrial and nuclear) and the population/species tree ( Degnan and Rosenberg 2009 ; Toews and Brelsford 2012 ).

Here, we address this by generating and analyzing a high-coverage nuclear and mitochondrial genome from a male Indian wolf captured for a radio-telemetry study in Velavadar Blackbuck National Park, Gujarat State, in western India ( Jhala 2003 ), which we call IW01. IW01 had the morphological traits ( supplementary fig. S1 , Supplementary Material online ) and mitochondrial sequence of a typical Indian peninsular wolf ( Sharma et al. 2004 ). We analyze IW01 in conjunction with previously published genomic data from gray wolves sampled from Eurasia and North America ( supplementary table S1 , Supplementary Material online ), including SRS661487, the wolf mentioned above whose precise origins remain ambiguous. We find strong evidence that IW01, along with Himalayan/Tibetan wolves, comprise lineages that are basal to all other gray wolves in both mitochondrial and nuclear phylogenies. Reconstruction of demographic histories also reveals that IW01 has a distinct effective population size trajectory compared with other wolves. Finally, we uncover evidence of historical admixture between IW01 and several canid lineages from Africa despite their current geographical separation, as well as gene flow between the domestic dog + gray wolf clade and these African canids. Our analyses indicate, however, that despite this history of admixture, the Indian wolf lineage has been evolving in isolation from other gray wolf lineages for around 110 thousands years.

Results and Discussion

Genome sequencing and mitochondrial phylogeny.

We extracted genomic DNA from IW01 using a whole blood sample collected in 1995. We prepared four pair-end sequencing libraries from which we sequenced 93.5 G nucleotide bases. We mapped sequencing reads to the domestic dog CanFam3.1 reference genome assembly, which yielded a 30.7-fold coverage genome for IW01. In addition, we de novo assembled the mitochondrial genome from IW01 to 2,557-fold coverage. From this whole mitochondrial genome, we extracted the cytochrome b and 16S rRNA gene sequences, which we used to estimate a phylogeny including IW01 and previously published mitochondrial data from Indian and other gray wolves for which full mitochondrial genomes were unavailable. Maximum-likelihood trees based on these two genes place IW01 in a previously reported clade containing other wolves from peninsular India that, along with Himalayan/Tibetan wolves, is basal to Holarctic gray wolves and domestic dogs ( supplementary fig. S2 , Supplementary Material online ).

Using published raw read data, we also de novo assembled mitochondrial genomes of wolves putatively originating from India (SRS661487) and Iran (SRS661488), both of which lack precise locality information ( Fan et al. 2016 ). We aligned these to a data set of 36 previously published mitochondrial genomes representing different Eurasian and North American gray wolf populations, including one Tibetan wolf and one Himalayan wolf, domestic dogs, and other species belonging to the genera Canis , Cuon , and Lycaon . As with the single gene analyses, IW01 was basal to all Holarctic gray wolves but inside the clade containing the Himalayan and Tibetan wolves, and distant from the SRS661487 (India) and SRS661488 (Iran), which cluster within the clade comprising Holarctic wolves and domestic dogs ( fig. 1 C ).

Phylogenetic Relationship between the Indian Wolf IW01 and Other Gray Wolves

We combined gene trees estimated from 5,000 randomly selected 20-kb regions across the nuclear genomes of IW01 and 18 other canids and reconstructed a species tree using ASTRAL-III ( Zhang et al. 2018 ). As observed previously ( Koepfli et al. 2015 ; Gopalakrishnan et al. 2018 ), the African wolf and golden jackal are basal to the coyote and gray wolf clades ( fig. 2 A and B ), and the Ethiopian wolf is an outgroup to the golden jackal. Domestic dogs and East Asian gray wolves formed a clade sister to European gray wolves, but with low support ( fig. 2 A and supplementary fig. S3 A , Supplementary Material online ). Quartet frequencies of gene trees comprising domestic dog, East Asian wolf, and European wolf were similar ( supplementary fig. S3 A , Supplementary Material online ). When IW01, SRS661487 (India), and SRS661488 (Iran) are included in the ASTRAL tree, these three lineages form a well-supported clade basal to North American and Eurasian wolves following the split of Himalayan and Tibetan wolves, the latter of which comprises the earliest diverging lineage in the gray wolf/domestic dog clade ( fig. 2 A ). This result is inconsistent with the phylogenetic tree presented in Fan et al. (2016) , based on a supermatrix analysis of genome-wide SNP data that do not account for gene tree discordance. In Fan et al. (2016) , SRS661487 and SRS661488 fall in the clade with European wolves, as they do in our mitochondrial phylogeny ( fig. 1 C ). When we estimated the ASTRAL tree excluding IW01, SRS661487, and SRS661488 cluster with European wolves ( supplementary fig. S4 , Supplementary Material online ) as in Fan et al. (2016) . When the ASTRAL tree includes IW01 but excludes SRS661487 and SRS661488, IW01 falls basal to all gray wolves and the domestic dog, including the Himalayan and Tibetan wolf clade, with strong support ( fig. 2 C and D , panel 10). However, the placement of the Himalayan/Tibetan wolf clade has low support ( supplementary fig. S3 B , Supplementary Material online ), suggesting that the phylogenetic relationship among IW01, Himalayan/Tibetan wolf, and the domestic dog + gray wolf clade is not well resolved, possibly due to incomplete sorting and/or gene flow among these lineages.

Phylogenetic analysis of Indian wolf IW01 and other canids based on nuclear genomes. ( A ) Consensus phylogenetic tree obtained using ASTRAL-III, estimated from 5,000 20-kb regions sampled across the nuclear genomes of the domestic dog, representative gray wolves, and other canid species. The blue-colored values at each node show the mean posterior probability of that node. White numbers with black squares denote branch numbers. As domestic dogs are a monophyletic group within the clade containing gray wolves from Eurasia ( Fan et al. 2016 ; Wang et al. 2016 ), we only chose one high-coverage domestic dog genome for this analysis. ( B ) Quartet frequencies of three possible topologies for branch 9 in ( A ). The format “15,16|6,8” indicates the quartet topology with branches 15 and 16 together on one side and branches 6 and 8 on the other side. Quartet topology frequencies for 16 branches in the underlying unrooted phylogeny are shown in supplementary figure S3 A , Supplementary Material online . The red bar indicates the frequency of the topology shown in ( A ) and the other blue-colored bars represent frequencies of the two alternative topologies. The dotted line represents the one-third frequency cut-off of the true topology for each quartet ( Allman et al. 2011 ). ( C ) Phylogenetic tree estimated from the nuclear genome using ASTRAL-III but excluding the previously reported gray wolf genomes of SRS661487 (India) and SRS661488 (Iran). ( D ) The quartet frequencies of three possible topologies for branch 10 in ( C ). Quartet topology frequencies for 14 branches in the underlying unrooted phylogeny are shown in supplementary figure S3 B , Supplementary Material online .

To further explore the placement of IW01, we aligned the high-coverage nuclear genomes from IW01, a Tibetan wolf, a Chinese wolf, and a dhole and divided the alignment into 250-kb, 500-kb, and 1-Mb nonoverlapping segments, and then estimated maximum-likelihood phylogenetic trees for each segment. The most commonly observed topology, which accounted for 48–57% of windows, placed IW01 as basal to the Tibetan wolf and the Chinese wolf ( supplementary fig. S5 , Supplementary Material online ).

Given that the most commonly observed topology placed IW01 as basal to Tibetan wolves, which previously estimated contained as much as 39% ancestry from a deeply divergent “ghost” lineage ( Wang et al. 2020 ), it is possible that all or some component of the ancestry of IW01 is also from this “ghost” lineage. To test this, we constructed a neighbor-joining tree using only genomic segments characterized as of “ghost” origin in Himalayan and Tibetan wolves ( Wang et al. 2020 ). Similar to the mitochondrial tree ( fig. 1 C ), IW01 and Himalayan/Tibetan wolves formed two distinct clades in this analysis, with the latter clade basal to other gray wolves, including IW01, with high bootstrap support ( supplementary fig. S6 , Supplementary Material online ). These results suggest that IW01 is not the possible source of the “ghost” lineage ancestry. Instead, the “ghost” lineage is likely basal to IW01.

Finally, we modeled the genetic makeup and phylogenetic assignments of IW01 using admixture graphs. Because this analysis is based on genotype calls, we prioritized genomes with sequence coverage over 10-fold. In agreement with the above analyses, our data fit the graph models (no f4 outliers) in which IW01 is assigned to a lineage basal to Eurasian gray wolves and shows no signals of admixture with other gray wolf populations ( fig. 3 ). Our results also indicate Tibetan wolves have admixed ancestry that is perhaps derived from ancient hybridization between a lineage basal to IW01 and Eurasian gray wolves. Interestingly, this analysis suggests that the Mongolian wolf is also admixed ( fig. 3 ), with the majority of its ancestry coming from European wolves, and the remainder from a lineage connecting them to Himalayan/Tibetan wolves. Previous studies have suggested that the range of Himalayan/Tibetan wolves was probably expanded across much of Mongolia and Northwest China ( Sharma et al. 2004 ), although these wolves maintain different distributions and represent distinct genetic lineages today ( Werhahn et al. 2017 ).

Fitted admixture graphs (no f4 outliers) showing the genetic makeup for IW01, European wolves (represented here by the Spanish wolf), and two highland wolves (represented by the Mongolian wolf and Tibetan wolf). Dashed lines indicate inferred admixture events and the admixture proportions are reported next to the dashed lines. The likelihood is shown at the top of each graph. The first graph has the highest likelihood of support.

Gene Flow between Indian Wolf IW01 and Other Canids

The uncertainty of the phylogenetic placement of gray wolves SRS661487 (India) and SRS661488 (Iran), as well as previous reports of admixture among canid lineages ( Koepfli et al. 2015 ; Skoglund et al. 2015 ; Gopalakrishnan et al. 2018 ), suggest that one or more of the sampled Middle Eastern and Indian wolf lineages may have admixed ancestry. We explored genetic affinity and admixture between IW01 and other gray wolves using TreeMix ( Pickrell and Pritchard 2012 ) and D -statistics ( Green et al. 2010 ) by analyzing 32 nuclear genomes ( supplementary table S1 , Supplementary Material online ). Our results support IW01 as a diverged wolf lineage basal to other Eurasian gray wolves and that SRS661487 is closely related to Iranian and European gray wolves ( fig. 4 and supplementary figs. S4–S9 , Supplementary Material online ).

TreeMix tree graph allowing two migration edges. This configuration reveals IW01 is basal to other gray wolves and domestic dogs. Two admixture events are shown, one between the African wolf and Ethiopian wolf, and the other between IW01 and SRS661487 (India)+SRS661488 (Iran). Tree graphs and Treemix residuals inferred by allowing zero to five migration edges are shown in supplementary figures S7 and S8 , Supplementary Material online . The graph with two migrations has the lowest residual distance.

We also found evidence of gene flow between IW01 and the two wolves of suspect origin (SRS661487 [India] and SRS661488 [Iran]) ( figs. 4 and 5 A and B ; supplementary figs. S7–S9 , Supplementary Material online ), as well as between IW01 and three more recently reported Iranian wolves ( supplementary fig. S10 , Supplementary Material online ) ( Amiri Ghanatsaman et al. 2020 ). Admixture among these lineages is expected, given the lack of reproductive barriers and any major geographic barriers separating these populations. We did not find evidence, however, of gene flow between IW01 and the Himalayan wolf. This is surprising, given the proximity of their ranges but consistent with previous findings based on mitochondrial sequences ( Sharma et al. 2004 ) and our phylogenetic results. It is possible that differences in local adaptation between highland wolves of the trans-Himalayan and Tibetan plateau ( Werhahn et al. 2018 ; Wang et al. 2020 ) versus lowland wolves of the semi-arid habitats in peninsular India, along with the small population sizes and fragmented habitat of Indian wolves may lessen chances for admixture between these lineages ( Owen et al. 2002 ; Blinkhorn and Petraglia 2017 ). However, given that our analyses are currently limited to a single Indian wolf sample of known origin, additional genomes from wolves sampled across peninsular India and the Himalayan region will be required to reveal the extent of gene flow among these lineages.

D-statistics testing the amount of allele sharing between IW01 and other canid species. ( A ) Schematic plot showing the topology used for calculating D -statistics. The calculation with | Z |≥3 was considered statistically significant. ( B ) D -statistics find no clear evidence of admixture between IW01 and Himalayan wolf. Most D values are positive, suggesting that IW01 shares more derived alleles with other gray wolves (H2) than with the Himalayan wolf. We note that D(IW01, H2; Himalayan wolf, Andean fox) calculated when H2 includes North American wolves showed significant negative values. However, this should not be taken to signify admixture between IW01 and the Himalayan wolf. Rather, this is likely due to North American wolves sharing ancestry from more divergent species like the coyote ( vonHoldt et al. 2016 ; Sinding et al. 2018 ). ( C ) D -statistics plot showing the amount of allele sharing between dhole and IW01 and other gray wolves. ( D ) D -statistics plot showing allele sharing between the African wolf and IW01. ( E ) D -statistics plot depicting allele sharing between the Ethiopian wolf and IW01. This test also finds evidence of admixture between the African wolf and the Ethiopian wolf. ( F ) D -statistics plot showing allele sharing between the African wild dog and IW01. ( G ) D -statistics plot showing the amount of allele sharing between the golden jackal and the ancestral clade of gray wolves. D -statistics were significantly positive when the domestic dog, East Asian wolves, and European wolves were in position H2, but became insignificant when North American wolves and Tibetan wolves were in the H2 position. These results suggest past gene flow between the Eurasian golden jackal and the ancestor of domestic dog and Eurasian gray wolves, supporting previous studies ( Gopalakrishnan et al. 2018 ; Chavez et al. 2019 ). This also suggests that the Eurasian golden jackal has no or less gene flow with IW01 compared with domestic dogs and Eurasian gray wolves, despite the overlapping distributions of the former two species.

Using D -statistics, we did not find any evidence of admixture between IW01 and the Asiatic dhole when the domestic dog, East Asian wolf, Croatian wolf, Spanish wolf, or North American wolf are in position H2 ( fig. 5 C ). However, we detected significant gene flow between IW01 and Kenyan African wolf ( fig. 5 D ), Ethiopian wolf ( fig. 5 E ), and African wild dog ( fig. 5 F ). This is consistent with the recent radiation of including Lycaon , Cuon , and Canis , which has been estimated at ∼1.72 Ma in models that include the possibility of gene flow among lineages ( Chavez et al. 2019 ). Such gene flow may have been mediated through an unknown, earlier diverging donor species ( Gopalakrishnan et al. 2018 ). We also found evidence of gene flow between IW01 and each of three recently reported northwestern African wolves (from Senegal, Morocco, and Algeria) ( Liu et al. 2018 ), although the proportion of shared ancestry varied among individuals sampled ( supplementary tables S2 and S3 , Supplementary Material online ). Moreover, past gene flow has been reported in other geographically distant canid species ( fig. 5 E and F ; supplementary table S2 , Supplementary Material online ), such as between Ethiopian wolf and Eurasian gray wolves and golden jackals, and between Ethiopian wolf and lineage ancestral to northwestern and eastern African wolves ( Gopalakrishnan et al. 2018 ).

We constructed admixture graph models to further investigate admixture among IW01 and African canids. Because this analysis requires a specified graph topology for testing, it is challenging to implement this test with a large number of populations or species with histories involving complex admixture events. Following previous canid genomic studies ( Sinding et al. 2018 ), we simplified admixture graphs by beginning with a model that includes European wolf, Tibetan wolf, IW01, and Andean fox (as an out-group), and then adding African canid species to fit all possible f4-statistics ( Lipson 2020 ). In agreement with our D -statistics results, the fitted admixture graphs (no f4 outliers) indicated that IW01 had gene flow with the African wolf, Ethiopian wolf, and African wild dog ( fig. 6 and supplementary fig. S11 , Supplementary Material online ). We found more gene flow between IW01 and the African wolf and Ethiopian wolf than between IW01 and with the African wild dog. Because the admixture history among gray wolf and canid species is complex, these fitted graphs reflect a parsimonious summary of our data and may not reflect the complete admixture history for these lineages.

Admixture graph modeling the gene flow between IW01 and African canids. ( A ) Fitted model (no f4 outliers) showing the genetic makeup of the Ethiopian wolf. ( B ) Fitted model (no f4 outliers) showing that the African wild dog carries 1% ancestry from IW01, which supports the D -statistics analysis ( fig. 5 F ). Admixture between IW01 and African wild dogs is also detected when African wolf or golden jackal are included ( supplementary fig. S11 , Supplementary Material online ). ( C ) Fitted model (no f4 outliers) showing admixture between IW01 and the African wolf and the Ethiopian wolf. This result is also supported by an alternative fitted model ( supplementary fig. S11 , Supplementary Material online ). Both support a previous conclusion that African wolves carry admixed ancestries ( Gopalakrishnan et al. 2018 ). Dashed lines indicate inferred admixture events and admixture proportions are reported beside the dashed lines. Because this analysis required genotype calls, we included only genomes with sequencing coverage >10-fold. As genome sequence coverage for the Himalayan wolf is 7-fold, we used the Tibetan wolf to represent the highland gray wolf for admixture graph construction.

Lastly, we applied PCAdmix ( Brisbin et al. 2012 ) to perform local ancestry inference for IW01, with African canids, Eurasian gray wolf, and domestic dog as source populations. Although this analysis has low power and resolution to infer small tracts reflecting anciently admixed ancestry, IW01 shared some potentially admixed tracts (posterior probabilities > 0.9) with each of the three African canid species, the African wolf, Ethiopian wolf, and African wild dog ( supplementary table S4 , Supplementary Material online ). The identified admixed tracts were short and few in number, indicative of ancient gene flow. IW01 shared the largest number and length of admixed blocks with the African wolf, followed by the Ethiopian wolf.

The above analyses support pervasive ancient gene flow between IW01 and African canids. Compared with the two wolves SRS661487 (India) and SRS661488 (Iran), IW01 shares less ancestry with African wolves and a comparable amount of ancestry with the Ethiopian wolf ( fig. 5 D and E ). A possible explanation for this pattern is that gene flow between IW01 and African canids was mediated through Middle Eastern wolves. However, this model does not explain the shared ancestry between IW01 and African wild dogs ( figs. 5 F and ​ and6; 6 ; supplementary fig. S11 , Supplementary Material online ).

Further, our results show that Iranian wolf genomes shared a large excess of genetic ancestry with IW01 ( fig. 4 and supplementary fig. S7 , Supplementary Material online ). This suggests that the lineage leading to IW01 may have been more widely distributed in the past, from the Indian subcontinent to the Arabian Peninsula ( Sharma et al. 2004 ), and overlapping in range and potentially hybridizing with Middle-Eastern gray wolves and African canid lineages in the past.

Our results support the hypothesis that the Sinai Peninsula and Southwest Levant are important hubs of canid evolution, where pervasive interspecific hybridization has been detected among gray wolves, African wolves, and Eurasian golden jackals ( Koepfli et al. 2015 ; Gopalakrishnan et al. 2018 ). Assemblages of Early Pleistocene mammalian fossils from the Pinjor Formation in India, including remains of at least two species of Canis , suggest paleobiogeographic linkages with African and Middle Eastern faunas ( Patnaik and Nanda 2010 ). The connections between the faunas of India and Africa are also supported by the vertebrate fossil records from Late Pleistocene deposits in Gujurat, which includes a Canis sp. that is larger and more robust than the present-day Indian wolf ( Costa 2017 ), and from other taxa, as Asiatic lions in India have experienced extensive gene flow with African lions ( de Manuel et al. 2020 ), and African leopards are known to have admixed with leopards from the Middle East (Palestine region) and Central Asia (Afghanistan) ( Paijmans et al. 2021 ). Our model is, of course, speculative, and additional data from both fossils and living animals will be helpful to understand the history of admixture among these canid lineages.

Intriguingly, D -statistics tests of allele sharing between IW01 and African canids revealed the Himalayan wolf as distinct from other wolf lineages ( fig. 5 C – G ), leading us to hypothesize that the Himalayan wolf was less admixed. To test this, we computed D -statistics with the Himalayan wolf as H1, domestic dog and gray wolves as H2, and African wolf, Ethiopian wolf, African wild dog, or golden jackal as H3. All analyses resulted in significant positive D values ( Z > 3), suggesting that the domestic dog and gray wolves also shared excess derived alleles with African canids and golden jackals ( supplementary fig. S12 , Supplementary Material online ). This analysis provides support for the idea that present-day wolves and domestic dogs have admixed ancestries ( Fan et al. 2016 ; Frantz et al. 2016 ) and that the Himalayan wolf is relatively isolated (less or unadmixed with other canids) compared with other wolves ( fig. 5 B ).

Demographic History and Divergence Time for the IW01 and Other Gray Wolves

To place the evolution of IW01 in a chronological context along with other gray wolves, we calculated relative cross-coalescence rates (CCR, the ratio between the cross- and the within-coalescence rates) for each pair of populations using the multiple sequentially Markovian coalescent (MSMC) model ( Schiffels and Durbin 2014 ), including genomes with a sequence coverage >20-fold. Using 50% CCR as a cutoff to estimate divergence time, these analyses suggest that IW01 diverged from domestic dogs and Chinese, Tibetan, European, and American wolves ∼110 ka ( fig. 7 A ). This divergence date is much older than the previous estimates of ∼68–81 ka for divergence between the Tibetan wolf and domestic dog/East Asian gray wolves ( Wang et al. 2020 ) and supports our phylogenetic result that IW01 is basal to the Tibetan/Himalayan wolf and domestic dog + gray wolf clade ( figs. 2 C and ​ and4; 4 ; supplementary figs. S5–S7 , Supplementary Material online ). This analysis also showed that IW01 split from SRS661487 (India) and SRS661488 (Iran) more recently, around 86 and 81 ka, respectively ( fig. 7 A ), although these estimates will be impacted by the admixed ancestry of these three individuals ( fig. 4 ). We estimated that SRS661487 diverged from the domestic dog and Chinese wolf ∼68–85 ka and from European wolf ∼17 ka ( supplementary fig. S13 , Supplementary Material online ), and that SRS661487 separated from the Iranian wolf (SRS661488) ∼5.5 ka, consistent with these two samples clustering together in the phylogeny ( fig. 4 and supplementary figs. S6 and S7 , Supplementary Material online ). Therefore, SRS661487 likely represents a gray wolf that recently descended from Middle Eastern and European wolf lineages that then admixed with the IW01 lineage, whereas IW01 is a distinct and deeply diverged lineage.

Inferences of the time of divergence and demographic history of Indian wolf IW01 and other gray wolves. ( A ) MSMC estimation of splitting time for Indian wolves (IW01) from the domestic dog and other representative gray wolves. Gray-dashed vertical line indicates the estimated split time at ∼110 ka. ( B ) Results of PSMC analysis showing the demographic trajectories of seven representative gray wolves. For each sample, we performed 100 bootstrap replicates.

We used the pairwise sequentially Markovian coalescent (PSMC) model ( Li and Durbin 2011 ) to reconstruct historical patterns of effective population size over time for IW01 and other gray wolves with sequencing coverage ≥20-fold ( fig. 7 B ). Generally, all gray wolves shared similar demographic trajectories up to ∼150 ka. Thereafter, IW01 and the Tibetan wolf diverged first around 110 ka and then experienced continuous contractions in population size. Generally consistent with MSMC and PSMC, we used Coal-HMM ( Mailund et al. 2011 ) and estimated that IW01 diverged from dogs and other gray wolves ∼130–140 ka ( supplementary fig. S14 , Supplementary Material online ). In contrast, SRS661487 shared a similar demographic trajectory with European, Iranian, and North American wolves whose population size expanded slightly between 100 and 50 ka, which was then followed by contraction ( fig. 7 B ). These results corroborate that IW01 and SRS661487 represent two different gray wolf lineages.

To explore the recent history of the Indian wolf population, we examined nucleotide diversity and runs of homozygosity (ROH) for IW01 and compared this with the estimates from other gray wolves. Because such analyses are sensitive to genotyping errors, we focused on genomes with sequencing coverage ≥20-fold. IW01 had a nucleotide diversity of approximately 0.00104 ± 0.00098 (mean±SD), slightly higher than that of the Tibetan wolf, but lower than that estimated for European wolves, SRS661487, the Iranian wolf (SRS661488), the Mongolian wolf, and the North American wolf ( supplementary fig. S15 , Supplementary Material online ). IW01 had 11 blocks of ROH with a length >1 Mb, the longest of which was 1.57 Mb, whereas the Tibetan wolf had 48 blocks of ROH >1 Mb and 5 ROH >2 Mb ( supplementary fig. S16 , Supplementary Material online ). We found that 33% of the IW01 genome and 43% of the Tibetan wolf genome were homozygous, which was higher than that observed in other gray wolves except for the Chinese wolf ( supplementary fig. S16 , Supplementary Material online ). These results are consistent with the long-term small effective population sizes inferred in our PSMC analysis and with earlier ecological studies ( Aggarwal et al. 2003 , 2007 ; Sharma et al. 2004 ), and also suggest recent inbreeding.

Our results suggest that IW01 represents an evolutionarily distinct gray wolf lineage living in the semi-arid lowland region of the Indian subcontinent that diverged from other gray wolf populations ∼110 ka. IW01 shares ancestry with other gray wolves (SRS661487 and SRS661488) that fall within the geographic range described for C. l. pallipes . Consistent with our previous study, gray wolves from the Trans-Himalayan mountain range and Tibetan Plateau also carry deeply diverged ancestries ( Wang et al. 2020 ). The persistence of these ancient and diverged lineages in the Indian subcontinent may be due in part to the region’s unique topography and paleoenvironmental history. Similar patterns of locally divergent lineages have been observed in Trans-Himalayan red pandas ( Hu et al. 2020 ) and Chinese mountain cats ( Yu et al. 2021 ). Together, these findings point to the importance of the Indian subcontinent and Trans-Himalayan region as refugia during the Pleistocene ( Sharma et al. 2004 ; Costa 2017 ) that enabled the persistence of divergent lineages.

During the Pleistocene ice ages, the Indian subcontinent was dry and cold, and much of the Himalayan and Trans-Himalayan regions and southern Tibet ( Owen et al. 2002 ) were covered by ice. Regional unglaciated refugia persisted, however, within which small populations of gray wolves may have become isolated, leading to the evolution of distinct lineages ( Blinkhorn and Petraglia 2017 ). Our estimate of the timing of divergence between IW01 and other gray wolves coincides roughly with the end of the Last Interglacial period (Eemian), when warmer, wetter conditions occurred in the northern latitudes of Eurasia, whereas the Indian subcontinent and neighboring lower latitude regions experienced a cooler, drier climate ( Pedersen et al. 2017 ). These paleoclimatic differences, combined with geographic isolation, may have facilitated ecological and genetic divergence of the Indian wolf lineage.

Despite the relative isolation and small population size of Indian wolves today, we find that the IW01 lineage harbors evidence of a mosaic of past gene flow with the African wolf, Ethiopian wolf, African wild dog, and western Asian gray wolves. We also find that the Himalayan wolf shares significantly less admixed ancestry with modern-day African canids ( supplementary fig. S12 , Supplementary Material online ), which is consistent with its isolation and adaptation to the high-altitude arid environments of the Himalayan and Tibetan plateaus. It is possible that the distribution of gray wolves and African canids overlapped in the past, possibly in the Sinai Peninsula or Southwest Levant where several canid species are hypothesized to have hybridized ( Gopalakrishnan et al. 2018 ).

Our results present a scenario of pervasive gene flow between gray wolves and other canid species, adding to the growing evidence of the important role of interspecific hybridization in the evolution of canid species and populations specifically and the role of network-linked and reticulated evolution of species more generally. Although our study is based on a single sample of precisely known provenance, our analyses of IW01 bridge a data gap for gray wolves and provide an important resource for future studies. Additional sampling of Indian wolves from other regions of peninsular India, of other wolves from across the range of C. l. pallipes , and perhaps from ancient samples will be necessary to inform the conservation of this threatened and elusive gray wolf subspecies.

Materials and Methods

Iw01: origins and sampling.

The Indian wolf (IW01; fig. 1 and supplementary fig. S1 , Supplementary Material online ) sequenced for this study was captured in 1995 inside Velavadar Blackbuck National Park in Gujarat state, India (latitude = 22.0438°N, longitude = 72.0202°E), for a radio-telemetry-based ecological study of the species. The wolf was captured using a rubberized-jaw McBride foot-hold trap (Minnesota) and anesthetized using Telozol ( Kreeger et al. 1989 ). Whole blood was drawn from the brachial vein for DNA profiling and disease study. Permissions for capture and collaring were obtained from the Ministry of Environment and Forest, government of India, and from the Chief Wildlife Warden, Gujarat state. The whole blood sample was stored in alcohol at −20 °C until genomic DNA was extracted.

Genome Sequencing and Variant Calling

Four paired-end DNA sequencing libraries were prepared for IW01, resulting in a total of 311,789,040 paired-end 150-bp reads (corresponding to 93.5 Gb) generated by the M/s Xcelris Labs Ltd. Ahmedabad, Gujarat, India, using the Illumina HiSeq 2500 platform. We downloaded published genomic sequences from 30 other canid samples from the NCBI SRA (accession IDs are available in supplementary table S1 , Supplementary Material online ) including domestic dogs, African wild dog ( Lycaon pictus ), dhole (Cuon alpinus), coyote ( Canis latrans ), Eurasian golden jackal ( Canis aureus ), African wolf ( Canis lupaster ), Ethiopian wolf ( Canis simensis ), and Andean fox ( Lycalopex culpaeus ). We used Btrim ( Kong 2011 ) to remove low-quality bases. Because a highly contiguous chromosome-level reference genome assembly is not yet available for the gray wolf, we aligned the remaining reads to the domestic dog CanFam3.1 reference genome ( Lindblad-Toh et al. 2005 ) using the BWA-MEM algorithm ( Li 2014 ) with the settings “-t 4 –M.” We processed the bam alignment by coordinate sorting, marking duplicated reads, performed local realignment, and recalibrated base quality scores using the Picard (version 1.56; http://broadinstitute.github.io/picard/ , last accessed January 27, 2022) and GATK (version 3.7.0) packages ( McKenna et al. 2010 ). We called SNPs for all samples together using the UnifiedGenotyper function in GATK. To increase the reliability of the data, SNPs were further filtered as previously described ( Wang et al. 2020 ) using the VariantFiltration command in GATK with parameters: “QUAL < 40.0 MQ < 25.0 MQ0 ≥ 4 && ((MQ0/(1.0×DP)) > 0.1) cluster 3 -window 10.” Index, depth, and mapping statistics were computed using available tools in SAMtools v1.3.1 ( Li et al. 2009 ).

Mitochondrial Assembly and Phylogenetic Analysis

Because no complete mitochondrial genome is available in GenBank for the Indian wolf, we performed de novo assembly of the mitochondrial genome for IW01, SRS661487 (India), and SRS661488 (Iran) using NOVOPlasty v2.7.2 ( Dierckxsens et al. 2017 ) with a k-mer size of 31 based on whole-genome sequencing data. The domestic dog mitochondrial genome (GenBank accession: NC_002008.4 ) was used as a seed/reference sequence. We downloaded mitochondrial genomes for coyote, African dog, dhole, African wolf, and other gray wolves and domestic dogs from NCBI (GenBank accessions are shown in fig. 1 C ) and included the Tibetan and Himalayan wolf sequences from a previous study ( Wang et al. 2020 ). A total of 39 mitogenomes were analyzed in this study. These sequences were aligned using MUSCLE v3.8.31 ( Edgar 2004 ) and the alignments were checked manually. After removing poorly aligned and control regions, an alignment file with a length of 15,462 bp was used for phylogenetic analysis. A maximum-likelihood tree was reconstructed using RAxML v8.2.12 ( Stamatakis 2014 ) with the GTR+G model of DNA substitution, and 1,000 bootstraps were run to assess node support.

We also downloaded previously reported mitochondrial cytochrome b and 16S rRNA sequences for Indian wolf, domestic dog, and other gray wolves from GenBank (accessions are shown in supplementary fig. S2 , Supplementary Material online ) and aligned and analyzed these data (554 bp for 16S rRNA and 332 bp for cytochrome b ) for phylogenetic analysis using the same methods described above.

Nuclear Phylogeny Construction

We constructed phylogenetic trees using nuclear genome sequences to explore the relationship of IW01 with other gray wolves and canid species. For each canid taxon, only one sample was used. Given that domestic dogs constitute a monophyletic clade ( Fan et al. 2016 ; Wang et al. 2016 ), we chose the high-coverage Dingo genome (31.3-fold; SRR7120191) to represent the domestic dog lineage. As a result, a total of 19 samples were used to construct phylogenetic trees ( fig. 2 A and supplementary fig. S3 , Supplementary Material online ). We generated a consensus genome for each sample using ANGSD v0.931 ( Korneliussen et al. 2014 ) (-doFasta 1). Reads with a minimum mapping quality lower than 25 were discarded (-minMapQ 25). For genomes with an average sequencing depth of over or less than 10-fold, the minimum depth for each base was set to 4-fold (-setMinDepth 4) or to 3-fold (-setMinDepth 3), respectively. Additional filter parameters implemented were: -doCounts 1 -uniqueOnly 1 -nThreads 2. We selected 5,000 random regions with a length of 20 kb from across the genome of the domestic dog reference assembly and the other 18 canid taxa using the “random” function in BEDTools v2.28.0 ( Quinlan and Hall 2010 ) (-l 20,000 -n 5,000). Sequences for each region were retrieved using the “faidx” function of SAMtools v1.3.1 ( Li et al. 2009 ). For each region, a maximum-likelihood tree was constructed by RAxML v8.2.12 ( Stamatakis 2014 ) with 100 bootstrap replicates using the command: raxmlHPC-PTHREADS-SSE3 -x 12,345 -k -# 100 -p 321 -m GTRGAMMAI -T 4 -s myseq.fas -f a -n myseq.ml.tre. The 5,000 gene trees were then concatenated and used as input for ASTRAL-III v5.7.5 ( Zhang et al. 2018 ) to generate a species tree, using default parameters. We used DiscoVista ( Sayyari et al. 2018 ) to analyze the discordance frequencies between the ASTRAL species tree and the 5,000 gene trees.

We retrieved and concatenated genotypes for 31 samples (in VCF format) within regions containing the signal of diverged origin in high-altitude wolves (Himalayan and Tibetan wolves) ( Wang et al. 2020 ), and converted into .fas format files. A neighbor-joining tree was constructed using the mega-cc tool ( Kumar et al. 2012 ) in MEGA7 ( Kumar et al. 2016 ) and nodal support was evaluated with 1,000 bootstrap replicates. Lastly, following ( Wang et al. 2020 ), we split four high-coverage genomes from the Chinese wolf, IW01, Tibetan wolf, and dhole into 250-, 500-, and 1,000-kb windows across autosomes and constructed phylogenies for each window using TreeMix v1.13 ( Pickrell and Pritchard 2012 ) with dhole as the outgroup. The frequency of each topology was calculated using APE v5.5 ( Popescu et al. 2012 ).

PSMC Analysis

We used the PSMC model to infer historical demographic trajectories for the sampled gray wolves ( Li and Durbin 2011 ). We only analyzed genomes with coverage >20-fold to ensure the accurate calling of heterozygotes ( Nadachowska-Brzyska et al. 2016 ), although some studies used low coverage genomes with false negative rate corrections ( Kim et al. 2014 ; Hawkins et al. 2018 ). A diploid consensus sequence for each individual was generated using the “mpileup” command of the SAMtools package (v1.3.1) ( Li et al. 2009 ) with the option “-C50.” Variants with less than about 1/3 (“-d” option) or over two times (“-D” option) of average read depth were marked as missing and excluded from consensus sequence assignment. Sequences with consensus quality lower than 20 were also filtered out. The program “fq2psmcfa” from the PSMC package was used to convert the consensus sequences into 100-bp bin-input files for PSMC. We ran PSMC with parameters “-N25 -t15 -r5 -p 4 + 25×2 + 4 + 6.” A total of 100 bootstraps were analyzed for each sample. These PSMC estimates are scaled using a generation time ( g ) of 3 years and a mutation rate (µ) of 4e − 9 substitutions per site per generation as used previously ( Skoglund et al. 2015 ). This mutation rate was comparable to a recent estimation based on pedigree analysis ( Koch et al. 2019 ).

MSMC and Coal-HMM Inference of Splitting Time

We used the multiple sequential Markovian coalescent (MSMC2) model to infer the divergence time for the domestic dog and gray wolf population pairs ( Schiffels and Durbin 2014 ). Genotypes for all dogs and wolves were phased together using Beagle V.4.1 ( Browning and Browning 2016 ). The MSMC input files comprising four haplotypes (two individuals) were generated as suggested by the authors using available tools from the MSMC-tool package ( https://github.com/stschiff/msmc-tools , last accessed January 27, 2022). We ran MSMC for each pair of genomes using default settings and the time when the relative cross-coalescent rate was dropped to 50% as an approximate estimate of the splitting time ( Malaspinas et al. 2016 ). For each calculation, four haplotypes were analyzed, and estimations were scaled using a generation time ( g ) of 3 years and a mutation rate (µ) of 4e −9 substitutions per site per generation ( Skoglund et al. 2015 ). Similar to the PSMC analysis, we restricted this analysis to genomes with coverage >20-fold.

We also used Coal-HMM ( Mailund et al. 2011 ), a coalescent hidden Markov model-based approach, to measure the divergence time for the Indian wolf (IW01) and dog and other wolves. We performed estimation for each population pair using 1-Mb nonoverlapping sliding window segments across each chromosome. We filtered out windows with over 10% missing rate for such analysis. We also removed results for each segment where: 1) the recombination rate was lower 0.1 or over 10 cM/Mb, 2) the ancestral effective population size below 1,000 or above 1,000,000, and 3) the split time was below 1,000 years or above 1,000,000 years.

Nuclear Diversity and ROH Analysis

Nucleotide diversity (π) ( Nei and Li 1979 ) was calculated for each sample across the autosomes using VCFtools v0.1.13 ( Danecek et al. 2011 ) in 50-kb sliding windows with a step of 25 kb. ROH was calculated for each sample across the autosomes using the “roh” function in the BCFtools v1.4-7-g41827a3 ( Narasimhan et al. 2016 ) with default parameters.

TreeMix, ABBA-BABA, and AdmixtureGraph Analyses

To explore the phylogenetic relationships and admixture among gray wolves and other canid species, we also used TreeMix v1.13 ( Pickrell and Pritchard 2012 ) to construct maximum-likelihood tree graphs by allowing gene flow. TreeMix analysis was run for all variants located on autosomes using 1,000 variants per block (-k 1,000) and allowing zero to five migrations, with Andean fox used as the outgroup.

We used the ABBA-BABA test, also known as D -statistics ( Green et al. 2010 ) to detect the amount of allele sharing between gray wolf populations. This analysis is based on the topology (((H1, H2), H3), Outgroup) as shown in figure 5 A . D = 0 suggests no gene flow between ingroup (H1 or H2) and H3; D > 0 suggests gene flow between H3 and H2; and D < 0 suggests gene flow between H3 and H1. We used the function “-doAbbababa 1” in ANGSD v0.931 ( Korneliussen et al. 2014 ) to perform this analysis with the additional settings “-doCounts 1 -minMapQ 25 -minQ 25 -uniqueOnly 1 -nThreads 6.”

To assess the genetic makeup and relationships among IW01, gray wolves, and three African canid species (African wolf, Ethiopian wolf, and African wild dog), we constructed admixture graph models using the qpGraph tool from AdmixTools package ( Patterson et al. 2012 ), the admixturegraph R package ( Leppala et al. 2017 ), and qpBrute ( Ni Leathlobhair et al. 2018 ; Liu et al. 2019 ). Because this analysis requires high-confidence genotype calls, we chose one sample with genome sequencing coverage over 10-fold from each population or species for constructing admixture graphs. To resolve the relationship between IW01, Himalayan/Tibetan wolves, and Eurasian gray wolves, we tested all possible graph models to fit all possible f4-statistics. The phylogenetic tree based on “ghost” admixed sequences and mitochondrial genomes from Himalayan or Tibetan wolves showed that the “ghost” lineage was basal to IW01. Therefore, we considered graphs in which Himalayan or Tibetan wolves were modeled as a product of admixture with one source from the lineage basal to IW01. To investigate the admixture between IW01 and African canids, we constructed admixture models starting with three populations (IW01, European wolf, and Himalayan or Tibetan wolf) and the fitted graph was then used as the base model in which we successively added each of the three African canid species.

Local Ancestry Inference

To identify potential admixed tracts along each chromosome in IW01, we performed local ancestry inference using PCAdmix ( Brisbin et al. 2012 ). We used phased genotypes as mentioned above as input, with IW01 designated as an admixed population and each of the African canid species, domestic dog, and Eurasian gray wolves as source populations. We performed two independent runs using 20 (default by the software) and 40 SNPs per window (“-w” parameter), respectively. The identified regions with posterior probabilities >0.9 were considered as potentially admixed.

Supplementary Material

Supplementary data are available at Genome Biology and Evolution online.

evac012_Supplementary_Data

Acknowledgments.

We thank the research communities for making their genomic data public, which makes this study possible. We also thank Y. Shah for the Indian wolf photos, and Robert Wayne for helpful discussions. This project was supported in part by DST-INSPIRE Faculty funding awarded to M.T. (04/2016/002246).

Author Contributions

M.T., Y.J., and B.S. conceived the idea and designed the research project. B.S., Y.J., K.-P.K., and R.E. G. supervised the analysis. M.-S.W., M.T., and S.W. performed the analysis with inputs from H.-M. C. and S.-S. D. Y.J., M.T., and Y.S. provided and coordinated genome sequencing of wolf sample IW01 within India. M.-S.W., M.T., and B.S. drafted the manuscript. B.S., Y.J., K.-P. K., M.-S.W., and M.T. revised the manuscript with input from all authors. Analysis of the Indian wolf sampled from western Gujarat, including sequencing and data analysis, was undertaken in India. Z.-X.L. submitted sequenced genome to NCBI. All authors read and improved the manuscript.

Data Availability

Literature cited.

• Aggarwal RK, Kivisild T, Ramadevi J, Singh L.. 2007. Mitochondrial DNA coding region sequences support the phylogenetic distinction of two Indian wolf species . J Zool Syst . 45 ( 2 ):163–172. [ Google Scholar ]
• Aggarwal RK, Ramadevi J, Singh L.. 2003. Ancient origin and evolution of the Indian wolf: evidence from mitochondrial DNA typing of wolves from trans-Himalayan region and Pennisular India . Genome Biol . 4 ( 6 ):P6. [ Google Scholar ]
• Allen GM. 1938. The mammals of China and Mongolia. Natural history of Central Asia. XI, Part I . 1st ed. New York: The American Museum of Natural History. p.342–345. [ Google Scholar ]
• Allman ES, Degnan JH, Rhodes JA.. 2011. Identifying the rooted species tree from the distribution of unrooted gene trees under the coalescent . J Math Biol . 62 ( 6 ):833–862. [ PubMed ] [ Google Scholar ]
• Amiri Ghanatsaman Z, et al. 2020. Whole genome resequencing of the Iranian native dogs and wolves to unravel variome during dog domestication . BMC Genomics 21 ( 1 ):207. [ PMC free article ] [ PubMed ] [ Google Scholar ]
• Blinkhorn J, Petraglia MD.. 2017. Environments and cultural change in the Indian Subcontinent: implications for the dispersal of Homo sapiens in the Late Pleistocene . Curr Anthropol . 58 ( S17 ):S463–S479. [ Google Scholar ]
• Brisbin A, et al. 2012. PCAdmix: principal components-based assignment of ancestry along each chromosome in individuals with admixed ancestry from two or more populations . Hum Biol . 84 ( 4 ):343–364. [ PMC free article ] [ PubMed ] [ Google Scholar ]
• Browning BL, Browning SR.. 2016. Genotype imputation with millions of reference samples . Am J Hum Genet . 98 ( 1 ):116–126. [ PMC free article ] [ PubMed ] [ Google Scholar ]
• Chavez DE, et al. 2019. Comparative genomics provides new insights into the remarkable adaptations of the African wild dog ( Lycaon pictus ) . Sci Rep . 9 ( 1 ):8329. [ PMC free article ] [ PubMed ] [ Google Scholar ]
• Costa AG. 2017. A new Late Pleistocene fauna from arid coastal India: implications for inundated coastal refugia and human dispersals . Quat Int . 436 :253–269. [ Google Scholar ]
• Danecek P, et al. 2011. The variant call format and VCFtools . Bioinformatics 27 ( 15 ):2156–2158. [ PMC free article ] [ PubMed ] [ Google Scholar ]
• Degnan JH, Rosenberg NA.. 2009. Gene tree discordance, phylogenetic inference and the multispecies coalescent . Trends Ecol Evol . 24 ( 6 ):332–340. [ PubMed ] [ Google Scholar ]
• de Manuel M, et al. 2020. The evolutionary history of extinct and living lions . Proc Natl Acad Sci U S A . 117 :10927–10934. [ PMC free article ] [ PubMed ] [ Google Scholar ]
• Dierckxsens N, Mardulyn P, Smits G.. 2017. NOVOPlasty: de novo assembly of organelle genomes from whole genome data . Nucleic Acids Res . 45 ( 4 ):e18. [ PMC free article ] [ PubMed ] [ Google Scholar ]
• Edgar RC. 2004. MUSCLE: multiple sequence alignment with high accuracy and high throughput . Nucleic Acids Res . 32 ( 5 ):1792–1797. [ PMC free article ] [ PubMed ] [ Google Scholar ]
• Ersmark E, et al. 2016. From the past to the present: wolf phylogeography and demographic history based on the mitochondrial control region . Front Ecol Evol . 4 . doi: 10.3389/fevo.2016.00134. [ CrossRef ] [ Google Scholar ]
• Fan Z, et al. 2016. Worldwide patterns of genomic variation and admixture in gray wolves . Genome Res . 26 ( 2 ):163–173. [ PMC free article ] [ PubMed ] [ Google Scholar ]
• Frantz LA, et al. 2016. Genomic and archaeological evidence suggest a dual origin of domestic dogs . Science 352 ( 6290 ):1228–1231. [ PubMed ] [ Google Scholar ]
• Freedman AH, et al. 2014. Genome sequencing highlights the dynamic early history of dogs . PLoS Genet . 10 ( 1 ):e1004016. [ PMC free article ] [ PubMed ] [ Google Scholar ]
• Gopalakrishnan S, et al. 2018. Interspecific gene flow shaped the evolution of the genus Canis . Curr Biol . 28 ( 21 ):3441–3449.e3445. [ PMC free article ] [ PubMed ] [ Google Scholar ]
• Green RE, et al. 2010. A draft sequence of the Neandertal genome . Science 328 ( 5979 ):710–722. [ PMC free article ] [ PubMed ] [ Google Scholar ]
• Hawkins MTR, et al. 2018. Genome sequence and population declines in the critically endangered greater bamboo lemur ( Prolemur simus ) and implications for conservation . BMC Genomics 19 ( 1 ):445. [ PMC free article ] [ PubMed ] [ Google Scholar ]
• Hodgson B. 1847. Description of the wild ass and wolf of Tibet . Calcutta J Nat Hist . 7 :469–477. [ Google Scholar ]
• Hu Y, et al. 2020. Genomic evidence for two phylogenetic species and long-term population bottlenecks in red pandas . Sci Adv . 6 ( 9 ):eaax5751. [ PMC free article ] [ PubMed ] [ Google Scholar ]
• Jhala Y. 2003. Status, ecology and conservation of the Indian wolf Canis lupus pallipes sykes . J Bombay Nat Hist Soc . 100 :293–307. [ Google Scholar ]
• Kim HL, et al. 2014. Khoisan hunter-gatherers have been the largest population throughout most of modern-human demographic history . Nat Commun . 5 :5692. [ PMC free article ] [ PubMed ] [ Google Scholar ]
• Koch E, et al. 2019. De novo mutation rate estimation in wolves of known pedigree . Mol Biol Evol . 36 ( 11 ):2536–2547. [ PMC free article ] [ PubMed ] [ Google Scholar ]
• Koepfli KP, et al. 2015. Genome-wide evidence reveals that African and Eurasian golden jackals are distinct species . Curr Biol . 25 ( 16 ):2158–2165. [ PubMed ] [ Google Scholar ]
• Kong Y. 2011. Btrim: a fast, lightweight adapter and quality trimming program for next-generation sequencing technologies . Genomics 98 ( 2 ):152–153. [ PubMed ] [ Google Scholar ]
• Korneliussen TS, Albrechtsen A, Nielsen R.. 2014. ANGSD: analysis of next generation sequencing data . BMC Bioinformatics 15 :356. [ PMC free article ] [ PubMed ] [ Google Scholar ]
• Kreeger TJ, Mandsager RE, Seal US, Callahan M, Beckel M.. 1989. Physiological response of gray wolves to butorphanol-xylazine immobilization and antagonism by naloxone and yohimbine . J Wildl Dis . 25 ( 1 ):89–94. [ PubMed ] [ Google Scholar ]
• Kumar S, Stecher G, Peterson D, Tamura K.. 2012. MEGA-CC: computing core of molecular evolutionary genetics analysis program for automated and iterative data analysis . Bioinformatics 28 ( 20 ):2685–2686. [ PMC free article ] [ PubMed ] [ Google Scholar ]
• Kumar S, Stecher G, Tamura K.. 2016. MEGA7: molecular evolutionary genetics analysis version 7.0 for bigger datasets . Mol Biol Evol . 33 ( 7 ):1870–1874. [ PMC free article ] [ PubMed ] [ Google Scholar ]
• Leonard JA, et al. 2007. Megafaunal extinctions and the disappearance of a specialized wolf ecomorph . Curr Biol . 17 ( 13 ):1146–1150. [ PubMed ] [ Google Scholar ]
• Leppala K, Nielsen SV, Mailund T.. 2017. admixturegraph: an R package for admixture graph manipulation and fitting . Bioinformatics 33 ( 11 ):1738–1740. [ PMC free article ] [ PubMed ] [ Google Scholar ]
• Li H. 2014. Toward better understanding of artifacts in variant calling from high-coverage samples . Bioinformatics 30 ( 20 ):2843–2851. [ PMC free article ] [ PubMed ] [ Google Scholar ]
• Li H, Durbin R.. 2011. Inference of human population history from individual whole-genome sequences . Nature 475 ( 7357 ):493–496. [ PMC free article ] [ PubMed ] [ Google Scholar ]
• Li H, et al. 2009. The Sequence Alignment/Map format and SAMtools . Bioinformatics 25 ( 16 ):2078–2079. [ PMC free article ] [ PubMed ] [ Google Scholar ]
• Lindblad-Toh K, et al. 2005. Genome sequence, comparative analysis and haplotype structure of the domestic dog . Nature 438 ( 7069 ):803–819. [ PubMed ] [ Google Scholar ]
• Lipson M. 2020. Applying f4-statistics and admixture graphs: theory and examples . Mol Ecol Resour . 20 ( 6 ):1658–1667. [ PubMed ] [ Google Scholar ]
• Liu L, et al. 2019. Genomic analysis on pygmy hog reveals extensive interbreeding during wild boar expansion . Nat Commun . 10 :1992. [ PMC free article ] [ PubMed ] [ Google Scholar ]
• Liu Y-H, et al. 2018. Whole-genome sequencing of African dogs provides insights into adaptations against tropical parasites . Mol Biol Evol . 35 ( 2 ):287–298. [ PubMed ] [ Google Scholar ]
• Loog L, et al. 2020. Ancient DNA suggests modern wolves trace their origin to a late Pleistocene expansion from Beringia . Mol Ecol . 29 ( 9 ):1596–1610. [ PMC free article ] [ PubMed ] [ Google Scholar ]
• Mailund T, Dutheil JY, Hobolth A, Lunter G, Schierup MH.. 2011. Estimating divergence time and ancestral effective population size of Bornean and Sumatran orangutan subspecies using a coalescent hidden Markov model . PLoS Genet . 7 ( 3 ):e1001319. [ PMC free article ] [ PubMed ] [ Google Scholar ]
• Malaspinas AS, et al. 2016. A genomic history of Aboriginal Australia . Nature 538 ( 7624 ):207–214. [ PubMed ] [ Google Scholar ]
• McKenna A, et al. 2010. The Genome Analysis Toolkit: a MapReduce framework for analyzing next-generation DNA sequencing data . Genome Res . 20 ( 9 ):1297–1303. [ PMC free article ] [ PubMed ] [ Google Scholar ]
• Nadachowska-Brzyska K, Burri R, Smeds L, Ellegren H.. 2016. PSMC analysis of effective population sizes in molecular ecology and its application to black-and-white Ficedula flycatchers . Mol Ecol . 25 ( 5 ):1058–1072. [ PMC free article ] [ PubMed ] [ Google Scholar ]
• Narasimhan V, et al. 2016. BCFtools/RoH: a hidden Markov model approach for detecting autozygosity from next-generation sequencing data . Bioinformatics 32 ( 11 ):1749–1751. [ PMC free article ] [ PubMed ] [ Google Scholar ]
• Nei M, Li WH.. 1979. Mathematical model for studying genetic variation in terms of restriction endonucleases . Proc Natl Acad Sci U S A . 76 ( 10 ):5269–5273. [ PMC free article ] [ PubMed ] [ Google Scholar ]
• Ni Leathlobhair M, et al. 2018. The evolutionary history of dogs in the Americas . Science 361 ( 6397 ):81–85. [ PMC free article ] [ PubMed ] [ Google Scholar ]
• Nowak RM. 1979. North American quaternary Canis . Lawrence (KS): Museum of Natural History, University of Kansas. [ Google Scholar ]
• Nowak RM. 1995. Another look at wolf taxonomy. In: Carbyn LN, Fritts SH, Seip DR, editors. Ecology and conservation of wolves in a changing world . Edmonton (AB): Canadian Circumpolar Institute. p. 375–398. [ Google Scholar ]
• Owen LA, Finkel RC, Caffee MW.. 2002. A note on the extent of glaciation throughout the Himalaya during the global Last Glacial Maximum . Quat Sci Rev . 21 ( 1–3 ):147–157. [ Google Scholar ]
• Paijmans JLA, et al. 2021. African and Asian leopards are highly differentiated at the genomic level . Curr Biol . 31 ( 9 ):1872–1882. [ PubMed ] [ Google Scholar ]
• Patnaik R, Nanda AC.. 2010. Early Pleistocene mammalian faunas of India and evidence of connections with other parts of the world. In: Fleagle JG, Shea JJ, Grine FE, Baden AL, Leakey RE, editors. Out of Africa I . Dordrecht (The Netherlands): Springer Netherlands. p. 129–143. [ Google Scholar ]
• Patterson N, et al. 2012. Ancient admixture in human history . Genetics 192 ( 3 ):1065–1093. [ PMC free article ] [ PubMed ] [ Google Scholar ]
• Pedersen RA, Langen PL, Vinther BM.. 2017. The last interglacial climate: comparing direct and indirect impacts of insolation changes . Clim Dyn . 48 ( 9–10 ):3391–3407. [ Google Scholar ]
• Perri AR, et al. 2021. Dire wolves were the last of an ancient New World canid lineage . Nature 591 ( 7848 ):87–91. [ PubMed ] [ Google Scholar ]
• Pickrell JK, Pritchard JK.. 2012. Inference of population splits and mixtures from genome-wide allele frequency data . PLoS Genet . 8 ( 11 ):e1002967. [ PMC free article ] [ PubMed ] [ Google Scholar ]
• Pilot M, et al. 2019. Global phylogeographic and admixture patterns in grey wolves and genetic legacy of an ancient Siberian lineage . Sci Rep . 9 ( 1 ):17328. [ PMC free article ] [ PubMed ] [ Google Scholar ]
• Popescu AA, Huber KT, Paradis E.. 2012. ape 3.0: new tools for distance-based phylogenetics and evolutionary analysis in R . Bioinformatics 28 ( 11 ):1536–1537. [ PubMed ] [ Google Scholar ]
• Quinlan AR, Hall IM.. 2010. BEDTools: a flexible suite of utilities for comparing genomic features . Bioinformatics 26 ( 6 ):841–842. [ PMC free article ] [ PubMed ] [ Google Scholar ]
• Sayyari E, Whitfield JB, Mirarab S.. 2018. DiscoVista: interpretable visualizations of gene tree discordance . Mol Phylogenet Evol . 122 :110–115. [ PubMed ] [ Google Scholar ]
• Schiffels S, Durbin R.. 2014. Inferring human population size and separation history from multiple genome sequences . Nat Genet . 46 ( 8 ):919–925. [ PMC free article ] [ PubMed ] [ Google Scholar ]
• Sharma DK, Maldonado JE, Jhala YV, Fleischer RC.. 2004. Ancient wolf lineages in India . Proc Biol Sci . 271 ( Suppl 3 ):S1–S4. [ PMC free article ] [ PubMed ] [ Google Scholar ]
• Sinding MS, et al. 2018. Population genomics of grey wolves and wolf-like canids in North America . PLoS Genet . 14 ( 11 ):e1007745. [ PMC free article ] [ PubMed ] [ Google Scholar ]
• Skoglund P, Ersmark E, Palkopoulou E, Dalen L.. 2015. Ancient wolf genome reveals an early divergence of domestic dog ancestors and admixture into high-latitude breeds . Curr Biol . 25 ( 11 ):1515–1519. [ PubMed ] [ Google Scholar ]
• Smeds L, Kojola I, Ellegren H.. 2019. The evolutionary history of grey wolf Y chromosomes . Mol Ecol . 28 ( 9 ):2173–2191. [ PMC free article ] [ PubMed ] [ Google Scholar ]
• Stamatakis A. 2014. RAxML version 8: a tool for phylogenetic analysis and post-analysis of large phylogenies . Bioinformatics 30 ( 9 ):1312–1313. [ PMC free article ] [ PubMed ] [ Google Scholar ]
• Sykes WH. 1831. Catalogue of the Mammalia of Dukun (Deccan); with observations on the habits, etc., and characters of new species . Proc Zool Soc Lond . 1830–1831 :99–106. [ Google Scholar ]
• Thalmann O, et al. 2013. Complete mitochondrial genomes of ancient canids suggest a European origin of domestic dogs . Science 342 ( 6160 ):871–874. [ PubMed ] [ Google Scholar ]
• Toews DP, Brelsford A.. 2012. The biogeography of mitochondrial and nuclear discordance in animals . Mol Ecol . 21 ( 16 ):3907–3930. [ PubMed ] [ Google Scholar ]
• vonHoldt BM, et al. 2016. Whole-genome sequence analysis shows that two endemic species of North American wolf are admixtures of the coyote and gray wolf . Sci Adv . 2 ( 7 ):e1501714. [ PMC free article ] [ PubMed ] [ Google Scholar ]
• Wang GD, et al. 2016. Out of southern East Asia: the natural history of domestic dogs across the world . Cell Res . 26 ( 1 ):21–33. [ PMC free article ] [ PubMed ] [ Google Scholar ]
• Wang MS, et al. 2020. Ancient hybridization with an unknown population facilitated high-altitude adaptation of canids . Mol Biol Evol . 37 ( 9 ):2616–2629. [ PubMed ] [ Google Scholar ]
• Werhahn G, et al. 2018. The unique genetic adaptation of the Himalayan wolf to high-altitudes and consequences for conservation . Glob Ecol Conserv . 16 :e00455. [ Google Scholar ]
• Werhahn G, et al. 2017. Phylogenetic evidence for the ancient Himalayan wolf: towards a clarification of its taxonomic status based on genetic sampling from western . R Soc Open Sci . 4 ( 6 ):170186. [ PMC free article ] [ PubMed ] [ Google Scholar ]
• Wilson DE, Reeder DM.. 2005. Mammal species of the world: a taxonomic and geographic reference . Baltimore (MD): Johns Hopkins University Press. [ Google Scholar ]
• Yu H, et al. 2021. Genomic evidence for the Chinese mountain cat as a wildcat conspecific ( Felis silvestris bieti ) and its introgression to domestic cats . Sci Adv . 7 :eabg0221. [ PMC free article ] [ PubMed ] [ Google Scholar ]
• Zhang C, Rabiee M, Sayyari E, Mirarab S.. 2018. ASTRAL-III: polynomial time species tree reconstruction from partially resolved gene trees . BMC Bioinformatics 19 ( S6 ):153. [ PMC free article ] [ PubMed ] [ Google Scholar ]

Genome Sequencing a Gray Wolf from Peninsular India Provides New Insights into the Evolution and Hybridization of Gray Wolves

42 Pages Posted: 13 Jul 2021 Publication Status: Review Complete

Ming-Shan Wang

Chinese Academy of Sciences (CAS) - State Key Laboratory of Genetic Resources and Evolution; University of California, Santa Cruz - Howard Hughes Medical Institute

Mukesh Thakur

Wildlife Institute of India

Yadvendradev Jhala

Chinese Academy of Sciences (CAS) - Kunming Institute of Zoology

Yellapu Srinivas

Zheng-xi liu.

Jilin University (JLU) - College of Animal Sciences

Richard E. Green

University of California, Santa Cruz - Department of Biomolecular Engineering (BME)

Klaus Koepfli

George Mason University - Smithsonian-Mason School of Conservation

Beth Shapiro

University of California, Santa Cruz - Howard Hughes Medical Institute

The gray wolf ( Canis lupus ) is among the few large carnivores that survived the Late Pleistocene megafaunal extinctions. Thanks to a complex history of admixture and their extensive geographic ranges, the number of gray wolf subspecies and their phylogenic relationships remain poorly understood. 1-5 Here, we performed whole-genome sequencing of a gray wolf collected from peninsular India that was phenotypically distinct from other gray wolves outside India. Genomic analyses revealed that the gray wolf lineage from the Indian subcontinent is ancestral to other gray wolf lineages and diverged from other gray wolves ~116 thousand years ago. Despite their small long-term population size and large geographic range, we also found evidence for pervasive and mosaic gene flow between the Indian wolf and African canids including African golden wolf, Ethiopian wolf, and African wild dog, as well as with West Asian gray wolves. Our study highlights the complex history of gene flow that characterized the evolution of gray wolves and contributes to a better understanding of their evolutionary history.

Suggested Citation: Suggested Citation

Chinese Academy of Sciences (CAS) - State Key Laboratory of Genetic Resources and Evolution ( email )

No.32 Jiaochang Donglu Kunming, Yunnan 650223 China

University of California, Santa Cruz - Howard Hughes Medical Institute ( email )

Santa Cruz, CA 95604 United States

Wildlife Institute of India ( email )

Chinese academy of sciences (cas) - kunming institute of zoology ( email ).

52 Sanlihe Rd. Datun Road, Anwai Beijing, 100864 China

Jilin University (JLU) - College of Animal Sciences ( email )

Jilin China

University of California, Santa Cruz - Department of Biomolecular Engineering (BME) ( email )

1156 High Street 335 Baskin Engineering Building Santa Cruz, CA 95064 United States

George Mason University - Smithsonian-Mason School of Conservation ( email )

1500 Remount Road Front Royal, VA 22630 United States

Beth Shapiro (Contact Author)

Click here to go to cell.com, paper statistics, related ejournals, ecology ejournal.

Subscribe to this fee journal for more curated articles on this topic

Genetics eJournal

Medical genetics & genomics ejournal.

Subscribe to this free journal for more curated articles on this topic

• Search Menu
• Sign in through your institution
• Volume 16, Issue 8, August 2024 (In Progress)
• Volume 16, Issue 7, July 2024
• Advance articles
• High-Impact Research Collection
• Celebrate 40 Years of Publishing
• Special sections
• Virtual Issues
• Research articles
• Perspectives
• Genome resources
• Biographies
• Author Guidelines
• Submission Site
• Open Access
• Reasons to submit
• About Genome Biology and Evolution
• About the Society for Molecular Biology and Evolution
• Editorial Board
• Advertising and Corporate Services
• Journals Career Network
• Self-Archiving Policy
• Journals on Oxford Academic
• Books on Oxford Academic

Article Contents

Introduction, results and discussion, materials and methods, supplementary material, acknowledgments, literature cited.

• < Previous

Genome Sequencing of a Gray Wolf from Peninsular India Provides New Insights into the Evolution and Hybridization of Gray Wolves

Ming-Shan Wang, Mukesh Thakur, Yadvendradev Jhala contributed equally to this work.

• Article contents
• Figures & tables
• Supplementary Data

Ming-Shan Wang, Mukesh Thakur, Yadvendradev Jhala, Sheng Wang, Yellapu Srinivas, Shan-Shan Dai, Zheng-Xi Liu, Hong-Man Chen, Richard E Green, Klaus-Peter Koepfli, Beth Shapiro, Genome Sequencing of a Gray Wolf from Peninsular India Provides New Insights into the Evolution and Hybridization of Gray Wolves, Genome Biology and Evolution , Volume 14, Issue 2, February 2022, evac012, https://doi.org/10.1093/gbe/evac012

• Permissions Icon Permissions

The gray wolf ( Canis lupus ) is among the few large carnivores that survived the Late Pleistocene megafaunal extinctions. Thanks to their complex history of admixture and extensive geographic range, the number of gray wolf subspecies and their phylogenetic relationships remain poorly understood. Here, we perform whole-genome sequencing of a gray wolf collected from peninsular India that was phenotypically distinct from gray wolves outside India. Genomic analyses reveal that the Indian gray wolf is an evolutionarily distinct lineage that diverged from other extant gray wolf lineages ∼110 thousand years ago. Demographic analyses suggest that the Indian wolf population declined continuously decline since separating from other gray wolves and, today, has exceptionally low genetic diversity. We also find evidence for pervasive and mosaic gene flow between the Indian wolf and African canids including African wolf, Ethiopian wolf, and African wild dog despite their current geographical separation. Our results support the hypothesis that the Indian subcontinent was a Pleistocene refugium and center of diversification and further highlight the complex history of gene flow that characterized the evolution of gray wolves.

The gray wolf ( Canis lupus ) is one of the few megafaunal carnivores that survived the Late Pleistocene megafaunal extinctions. Despite extensive research on living and extinct gray wolves, the evolutionary history of this lineage remains unclear. Here, we sequence and analyze a draft genome of a gray wolf collected from peninsular India. We find that the Indian wolf lineage, which is both highly threatened and phenotypically distinct from other gray wolves, is the most deeply diverging lineage of extant gray wolves and, despite their physical isolation from other wolf lineages, has a long history of gene flow with other canid lineages.

The gray wolf ( Canis lupus ) first appears in the fossil records of Eurasia and North America some 500,000 years ago ( Nowak 1979 ) and later diversified into more than 37 named subspecies ( Wilson and Reeder 2005 ). Numerous morphological and genomic analyses of gray wolves have presented a complex and sometimes contradictory view of their evolutionary history ( Leonard et al. 2007 ; Sinding et al. 2018 ; Smeds et al. 2019 ; Loog et al. 2020 ). For example, analyses of mitochondrial DNA have revealed a lack of strong haplotype structure among populations across the Northern hemisphere ( Thalmann et al. 2013 ; Loog et al. 2020 ), whereas nuclear genomic analyses have identified distinct lineages in Eurasia and North America ( Fan et al. 2016 ; Gopalakrishnan et al. 2018 ). These studies have also revealed widespread admixture among domestic dogs, gray wolves, and other species in the genera Canis and Cuon ( Freedman et al. 2014 ; Fan et al. 2016 ; Gopalakrishnan et al. 2018 ; Pilot et al. 2019 ). This evolutionary history of dynamic long-distance dispersal, population replacements, and cross-species gene flow has complicated efforts to understand both how gray wolf populations are related to each other and the location, origin, and timing of dog domestication ( Koepfli et al. 2015 ; Perri et al. 2021 ).

Among the least studied populations of gray wolves are those that inhabit the Indian subcontinent. Early taxonomists described two species endemic to this region: the Himalayan wolf, Canis laniger ( Hodgson 1847 ), found in the highland regions of the Tibetan Plateau and eastern Kashmir, and the Indian wolf Canis pallipes ( Sykes 1831 ), distributed within the arid/semi-arid lowland plains of peninsular India. Since these first descriptions, Himalayan and Indian wolves have been reclassified as subspecies within the gray wolf complex, Canis lupus chanco and C. l. pallipes , respectively ( Allen 1938 ). The current range of C. l. pallipes extends from the eastern Mediterranean region of western Asia eastward to peninsular India, where several isolated populations are reported ( Nowak 1995 ; Jhala 2003 ).

Genetic studies using mitochondrial and nuclear markers have shown that the Himalayan wolf is distinct from other gray wolf populations ( Aggarwal et al. 2007 ; Ersmark et al. 2016 ; Werhahn et al. 2017 , 2018 ). Similarly, Indian wolves are morphologically, behaviorally, and genetically distinct from other wolf subspecies ( Aggarwal et al. 2003 ; Sharma et al. 2004 ; Aggarwal et al. 2007 ). Compared with other wolves, Indian wolves are smaller in size (18–22 kg) with less and relatively shorter hair that is light brown in color with black hair tips ( fig. 1A ). Indian wolves are also among the most threatened canid subspecies in the world, with an estimated population size of ∼2,000–3,000 individuals ( Jhala 2003 ).

Sampling location and mitochondrial phylogeny. ( A ) A photograph of an Indian wolf from peninsular India (provided by Y. Shah). ( B ) Map showing the distribution of samples used in this study. The red dot depicts the location where the Indian wolf IW01 ( supplementary fig. S1 , Supplementary Material online ) was sampled. ( C ) Maximum-likelihood tree estimated from mitochondrial genomes (15,462 bp). The Indian wolf (IW01) is a sister clade to domestic dogs and other gray wolves but inside the lineage of Tibetan wolf+Himalayan wolf. IDs in brackets are the GenBank accession numbers.

More recently, relationships among gray wolves have been analyzed using whole-genome sequences. In one study examining admixture among gray wolves and domestic dogs ( Fan et al. 2016 ), a wolf presumed to originate from India but lacking precise locality information (NCBI accession: SRS661487) clustered with wolves from Iran and Israel, which together were grouped within a larger cluster of gray wolves from Eurasia. This result was, however, at odds with earlier phylogenetic analyses based on mitochondrial sequences that suggested that Himalayan wolves and wolves from peninsular India are the earliest branchings and most divergent lineages among all gray wolf populations, with Indian wolves diverging from other lineages ∼270–400 thousand years ago (ka) ( Aggarwal et al. 2003 ; Sharma et al. 2004 ; Aggarwal et al. 2007 ). To resolve this inconsistency, additional analyses using wolf samples of unambiguous provenance are necessary, in particular as the complex history of admixture among canids can lead to discordance among individual gene trees (mitochondrial and nuclear) and the population/species tree ( Degnan and Rosenberg 2009 ; Toews and Brelsford 2012 ).

Here, we address this by generating and analyzing a high-coverage nuclear and mitochondrial genome from a male Indian wolf captured for a radio-telemetry study in Velavadar Blackbuck National Park, Gujarat State, in western India ( Jhala 2003 ), which we call IW01. IW01 had the morphological traits ( supplementary fig. S1 , Supplementary Material online ) and mitochondrial sequence of a typical Indian peninsular wolf ( Sharma et al. 2004 ). We analyze IW01 in conjunction with previously published genomic data from gray wolves sampled from Eurasia and North America ( supplementary table S1 , Supplementary Material online ), including SRS661487, the wolf mentioned above whose precise origins remain ambiguous. We find strong evidence that IW01, along with Himalayan/Tibetan wolves, comprise lineages that are basal to all other gray wolves in both mitochondrial and nuclear phylogenies. Reconstruction of demographic histories also reveals that IW01 has a distinct effective population size trajectory compared with other wolves. Finally, we uncover evidence of historical admixture between IW01 and several canid lineages from Africa despite their current geographical separation, as well as gene flow between the domestic dog + gray wolf clade and these African canids. Our analyses indicate, however, that despite this history of admixture, the Indian wolf lineage has been evolving in isolation from other gray wolf lineages for around 110 thousands years.

Genome Sequencing and Mitochondrial Phylogeny

We extracted genomic DNA from IW01 using a whole blood sample collected in 1995. We prepared four pair-end sequencing libraries from which we sequenced 93.5 G nucleotide bases. We mapped sequencing reads to the domestic dog CanFam3.1 reference genome assembly, which yielded a 30.7-fold coverage genome for IW01. In addition, we de novo assembled the mitochondrial genome from IW01 to 2,557-fold coverage. From this whole mitochondrial genome, we extracted the cytochrome b and 16S rRNA gene sequences, which we used to estimate a phylogeny including IW01 and previously published mitochondrial data from Indian and other gray wolves for which full mitochondrial genomes were unavailable. Maximum-likelihood trees based on these two genes place IW01 in a previously reported clade containing other wolves from peninsular India that, along with Himalayan/Tibetan wolves, is basal to Holarctic gray wolves and domestic dogs ( supplementary fig. S2 , Supplementary Material online ).

Using published raw read data, we also de novo assembled mitochondrial genomes of wolves putatively originating from India (SRS661487) and Iran (SRS661488), both of which lack precise locality information ( Fan et al. 2016 ). We aligned these to a data set of 36 previously published mitochondrial genomes representing different Eurasian and North American gray wolf populations, including one Tibetan wolf and one Himalayan wolf, domestic dogs, and other species belonging to the genera Canis , Cuon , and Lycaon . As with the single gene analyses, IW01 was basal to all Holarctic gray wolves but inside the clade containing the Himalayan and Tibetan wolves, and distant from the SRS661487 (India) and SRS661488 (Iran), which cluster within the clade comprising Holarctic wolves and domestic dogs ( fig. 1C ).

Phylogenetic Relationship between the Indian Wolf IW01 and Other Gray Wolves

We combined gene trees estimated from 5,000 randomly selected 20-kb regions across the nuclear genomes of IW01 and 18 other canids and reconstructed a species tree using ASTRAL-III ( Zhang et al. 2018 ). As observed previously ( Koepfli et al. 2015 ; Gopalakrishnan et al. 2018 ), the African wolf and golden jackal are basal to the coyote and gray wolf clades ( fig. 2A and B ), and the Ethiopian wolf is an outgroup to the golden jackal. Domestic dogs and East Asian gray wolves formed a clade sister to European gray wolves, but with low support ( fig. 2A and supplementary fig. S3A , Supplementary Material online ). Quartet frequencies of gene trees comprising domestic dog, East Asian wolf, and European wolf were similar ( supplementary fig. S3A , Supplementary Material online ). When IW01, SRS661487 (India), and SRS661488 (Iran) are included in the ASTRAL tree, these three lineages form a well-supported clade basal to North American and Eurasian wolves following the split of Himalayan and Tibetan wolves, the latter of which comprises the earliest diverging lineage in the gray wolf/domestic dog clade ( fig. 2A ). This result is inconsistent with the phylogenetic tree presented in Fan et al. (2016) , based on a supermatrix analysis of genome-wide SNP data that do not account for gene tree discordance. In Fan et al. (2016) , SRS661487 and SRS661488 fall in the clade with European wolves, as they do in our mitochondrial phylogeny ( fig. 1C ). When we estimated the ASTRAL tree excluding IW01, SRS661487, and SRS661488 cluster with European wolves ( supplementary fig. S4 , Supplementary Material online ) as in Fan et al. (2016) . When the ASTRAL tree includes IW01 but excludes SRS661487 and SRS661488, IW01 falls basal to all gray wolves and the domestic dog, including the Himalayan and Tibetan wolf clade, with strong support ( fig. 2C and D , panel 10). However, the placement of the Himalayan/Tibetan wolf clade has low support ( supplementary fig. S3B , Supplementary Material online ), suggesting that the phylogenetic relationship among IW01, Himalayan/Tibetan wolf, and the domestic dog + gray wolf clade is not well resolved, possibly due to incomplete sorting and/or gene flow among these lineages.

Phylogenetic analysis of Indian wolf IW01 and other canids based on nuclear genomes. ( A ) Consensus phylogenetic tree obtained using ASTRAL-III, estimated from 5,000 20-kb regions sampled across the nuclear genomes of the domestic dog, representative gray wolves, and other canid species. The blue-colored values at each node show the mean posterior probability of that node. White numbers with black squares denote branch numbers. As domestic dogs are a monophyletic group within the clade containing gray wolves from Eurasia ( Fan et al. 2016 ; Wang et al. 2016 ), we only chose one high-coverage domestic dog genome for this analysis. ( B ) Quartet frequencies of three possible topologies for branch 9 in ( A ). The format “15,16|6,8” indicates the quartet topology with branches 15 and 16 together on one side and branches 6 and 8 on the other side. Quartet topology frequencies for 16 branches in the underlying unrooted phylogeny are shown in supplementary figure S3A , Supplementary Material online . The red bar indicates the frequency of the topology shown in ( A ) and the other blue-colored bars represent frequencies of the two alternative topologies. The dotted line represents the one-third frequency cut-off of the true topology for each quartet ( Allman et al. 2011 ). ( C ) Phylogenetic tree estimated from the nuclear genome using ASTRAL-III but excluding the previously reported gray wolf genomes of SRS661487 (India) and SRS661488 (Iran). ( D ) The quartet frequencies of three possible topologies for branch 10 in ( C ). Quartet topology frequencies for 14 branches in the underlying unrooted phylogeny are shown in supplementary figure S3B , Supplementary Material online .

To further explore the placement of IW01, we aligned the high-coverage nuclear genomes from IW01, a Tibetan wolf, a Chinese wolf, and a dhole and divided the alignment into 250-kb, 500-kb, and 1-Mb nonoverlapping segments, and then estimated maximum-likelihood phylogenetic trees for each segment. The most commonly observed topology, which accounted for 48–57% of windows, placed IW01 as basal to the Tibetan wolf and the Chinese wolf ( supplementary fig. S5 , Supplementary Material online ).

Given that the most commonly observed topology placed IW01 as basal to Tibetan wolves, which previously estimated contained as much as 39% ancestry from a deeply divergent “ghost” lineage ( Wang et al. 2020 ), it is possible that all or some component of the ancestry of IW01 is also from this “ghost” lineage. To test this, we constructed a neighbor-joining tree using only genomic segments characterized as of “ghost” origin in Himalayan and Tibetan wolves ( Wang et al. 2020 ). Similar to the mitochondrial tree ( fig. 1C ), IW01 and Himalayan/Tibetan wolves formed two distinct clades in this analysis, with the latter clade basal to other gray wolves, including IW01, with high bootstrap support ( supplementary fig. S6 , Supplementary Material online ). These results suggest that IW01 is not the possible source of the “ghost” lineage ancestry. Instead, the “ghost” lineage is likely basal to IW01.

Finally, we modeled the genetic makeup and phylogenetic assignments of IW01 using admixture graphs. Because this analysis is based on genotype calls, we prioritized genomes with sequence coverage over 10-fold. In agreement with the above analyses, our data fit the graph models (no f4 outliers) in which IW01 is assigned to a lineage basal to Eurasian gray wolves and shows no signals of admixture with other gray wolf populations ( fig. 3 ). Our results also indicate Tibetan wolves have admixed ancestry that is perhaps derived from ancient hybridization between a lineage basal to IW01 and Eurasian gray wolves. Interestingly, this analysis suggests that the Mongolian wolf is also admixed ( fig. 3 ), with the majority of its ancestry coming from European wolves, and the remainder from a lineage connecting them to Himalayan/Tibetan wolves. Previous studies have suggested that the range of Himalayan/Tibetan wolves was probably expanded across much of Mongolia and Northwest China ( Sharma et al. 2004 ), although these wolves maintain different distributions and represent distinct genetic lineages today ( Werhahn et al. 2017 ).

Fitted admixture graphs (no f4 outliers) showing the genetic makeup for IW01, European wolves (represented here by the Spanish wolf), and two highland wolves (represented by the Mongolian wolf and Tibetan wolf). Dashed lines indicate inferred admixture events and the admixture proportions are reported next to the dashed lines. The likelihood is shown at the top of each graph. The first graph has the highest likelihood of support.

Gene Flow between Indian Wolf IW01 and Other Canids

The uncertainty of the phylogenetic placement of gray wolves SRS661487 (India) and SRS661488 (Iran), as well as previous reports of admixture among canid lineages ( Koepfli et al. 2015 ; Skoglund et al. 2015 ; Gopalakrishnan et al. 2018 ), suggest that one or more of the sampled Middle Eastern and Indian wolf lineages may have admixed ancestry. We explored genetic affinity and admixture between IW01 and other gray wolves using TreeMix ( Pickrell and Pritchard 2012 ) and D -statistics ( Green et al. 2010 ) by analyzing 32 nuclear genomes ( supplementary table S1 , Supplementary Material online ). Our results support IW01 as a diverged wolf lineage basal to other Eurasian gray wolves and that SRS661487 is closely related to Iranian and European gray wolves ( fig. 4 and supplementary figs. S4–S9 , Supplementary Material online ).

TreeMix tree graph allowing two migration edges. This configuration reveals IW01 is basal to other gray wolves and domestic dogs. Two admixture events are shown, one between the African wolf and Ethiopian wolf, and the other between IW01 and SRS661487 (India)+SRS661488 (Iran). Tree graphs and Treemix residuals inferred by allowing zero to five migration edges are shown in supplementary figures S7 and S8 , Supplementary Material online . The graph with two migrations has the lowest residual distance.

We also found evidence of gene flow between IW01 and the two wolves of suspect origin (SRS661487 [India] and SRS661488 [Iran]) ( figs. 4 and 5A and B ; supplementary figs. S7–S9 , Supplementary Material online ), as well as between IW01 and three more recently reported Iranian wolves ( supplementary fig. S10 , Supplementary Material online ) ( Amiri Ghanatsaman et al. 2020 ). Admixture among these lineages is expected, given the lack of reproductive barriers and any major geographic barriers separating these populations. We did not find evidence, however, of gene flow between IW01 and the Himalayan wolf. This is surprising, given the proximity of their ranges but consistent with previous findings based on mitochondrial sequences ( Sharma et al. 2004 ) and our phylogenetic results. It is possible that differences in local adaptation between highland wolves of the trans-Himalayan and Tibetan plateau ( Werhahn et al. 2018 ; Wang et al. 2020 ) versus lowland wolves of the semi-arid habitats in peninsular India, along with the small population sizes and fragmented habitat of Indian wolves may lessen chances for admixture between these lineages ( Owen et al. 2002 ; Blinkhorn and Petraglia 2017 ). However, given that our analyses are currently limited to a single Indian wolf sample of known origin, additional genomes from wolves sampled across peninsular India and the Himalayan region will be required to reveal the extent of gene flow among these lineages.

D-statistics testing the amount of allele sharing between IW01 and other canid species. ( A ) Schematic plot showing the topology used for calculating D -statistics. The calculation with | Z |≥3 was considered statistically significant. ( B ) D -statistics find no clear evidence of admixture between IW01 and Himalayan wolf. Most D values are positive, suggesting that IW01 shares more derived alleles with other gray wolves (H2) than with the Himalayan wolf. We note that D(IW01, H2; Himalayan wolf, Andean fox) calculated when H2 includes North American wolves showed significant negative values. However, this should not be taken to signify admixture between IW01 and the Himalayan wolf. Rather, this is likely due to North American wolves sharing ancestry from more divergent species like the coyote ( vonHoldt et al. 2016 ; Sinding et al. 2018 ). ( C ) D -statistics plot showing the amount of allele sharing between dhole and IW01 and other gray wolves. ( D ) D -statistics plot showing allele sharing between the African wolf and IW01. ( E ) D -statistics plot depicting allele sharing between the Ethiopian wolf and IW01. This test also finds evidence of admixture between the African wolf and the Ethiopian wolf. ( F ) D -statistics plot showing allele sharing between the African wild dog and IW01. ( G ) D -statistics plot showing the amount of allele sharing between the golden jackal and the ancestral clade of gray wolves. D -statistics were significantly positive when the domestic dog, East Asian wolves, and European wolves were in position H2, but became insignificant when North American wolves and Tibetan wolves were in the H2 position. These results suggest past gene flow between the Eurasian golden jackal and the ancestor of domestic dog and Eurasian gray wolves, supporting previous studies ( Gopalakrishnan et al. 2018 ; Chavez et al. 2019 ). This also suggests that the Eurasian golden jackal has no or less gene flow with IW01 compared with domestic dogs and Eurasian gray wolves, despite the overlapping distributions of the former two species.

Using D -statistics, we did not find any evidence of admixture between IW01 and the Asiatic dhole when the domestic dog, East Asian wolf, Croatian wolf, Spanish wolf, or North American wolf are in position H2 ( fig. 5C ). However, we detected significant gene flow between IW01 and Kenyan African wolf ( fig. 5D ), Ethiopian wolf ( fig. 5E ), and African wild dog ( fig. 5F ). This is consistent with the recent radiation of including Lycaon , Cuon , and Canis , which has been estimated at ∼1.72 Ma in models that include the possibility of gene flow among lineages ( Chavez et al. 2019 ). Such gene flow may have been mediated through an unknown, earlier diverging donor species ( Gopalakrishnan et al. 2018 ). We also found evidence of gene flow between IW01 and each of three recently reported northwestern African wolves (from Senegal, Morocco, and Algeria) ( Liu et al. 2018 ), although the proportion of shared ancestry varied among individuals sampled ( supplementary tables S2 and S3 , Supplementary Material online ). Moreover, past gene flow has been reported in other geographically distant canid species ( fig. 5E and F ; supplementary table S2 , Supplementary Material online ), such as between Ethiopian wolf and Eurasian gray wolves and golden jackals, and between Ethiopian wolf and lineage ancestral to northwestern and eastern African wolves ( Gopalakrishnan et al. 2018 ).

We constructed admixture graph models to further investigate admixture among IW01 and African canids. Because this analysis requires a specified graph topology for testing, it is challenging to implement this test with a large number of populations or species with histories involving complex admixture events. Following previous canid genomic studies ( Sinding et al. 2018 ), we simplified admixture graphs by beginning with a model that includes European wolf, Tibetan wolf, IW01, and Andean fox (as an out-group), and then adding African canid species to fit all possible f4-statistics ( Lipson 2020 ). In agreement with our D -statistics results, the fitted admixture graphs (no f4 outliers) indicated that IW01 had gene flow with the African wolf, Ethiopian wolf, and African wild dog ( fig. 6 and supplementary fig. S11 , Supplementary Material online ). We found more gene flow between IW01 and the African wolf and Ethiopian wolf than between IW01 and with the African wild dog. Because the admixture history among gray wolf and canid species is complex, these fitted graphs reflect a parsimonious summary of our data and may not reflect the complete admixture history for these lineages.

Admixture graph modeling the gene flow between IW01 and African canids. ( A ) Fitted model (no f4 outliers) showing the genetic makeup of the Ethiopian wolf. ( B ) Fitted model (no f4 outliers) showing that the African wild dog carries 1% ancestry from IW01, which supports the D -statistics analysis ( fig. 5F ). Admixture between IW01 and African wild dogs is also detected when African wolf or golden jackal are included ( supplementary fig. S11 , Supplementary Material online ). ( C ) Fitted model (no f4 outliers) showing admixture between IW01 and the African wolf and the Ethiopian wolf. This result is also supported by an alternative fitted model ( supplementary fig. S11 , Supplementary Material online ). Both support a previous conclusion that African wolves carry admixed ancestries ( Gopalakrishnan et al. 2018 ). Dashed lines indicate inferred admixture events and admixture proportions are reported beside the dashed lines. Because this analysis required genotype calls, we included only genomes with sequencing coverage >10-fold. As genome sequence coverage for the Himalayan wolf is 7-fold, we used the Tibetan wolf to represent the highland gray wolf for admixture graph construction.

Lastly, we applied PCAdmix ( Brisbin et al. 2012 ) to perform local ancestry inference for IW01, with African canids, Eurasian gray wolf, and domestic dog as source populations. Although this analysis has low power and resolution to infer small tracts reflecting anciently admixed ancestry, IW01 shared some potentially admixed tracts (posterior probabilities > 0.9) with each of the three African canid species, the African wolf, Ethiopian wolf, and African wild dog ( supplementary table S4 , Supplementary Material online ). The identified admixed tracts were short and few in number, indicative of ancient gene flow. IW01 shared the largest number and length of admixed blocks with the African wolf, followed by the Ethiopian wolf.

The above analyses support pervasive ancient gene flow between IW01 and African canids. Compared with the two wolves SRS661487 (India) and SRS661488 (Iran), IW01 shares less ancestry with African wolves and a comparable amount of ancestry with the Ethiopian wolf ( fig. 5D and E ). A possible explanation for this pattern is that gene flow between IW01 and African canids was mediated through Middle Eastern wolves. However, this model does not explain the shared ancestry between IW01 and African wild dogs ( figs. 5F and 6 ; supplementary fig. S11 , Supplementary Material online ).

Further, our results show that Iranian wolf genomes shared a large excess of genetic ancestry with IW01 ( fig. 4 and supplementary fig. S7 , Supplementary Material online ). This suggests that the lineage leading to IW01 may have been more widely distributed in the past, from the Indian subcontinent to the Arabian Peninsula ( Sharma et al. 2004 ), and overlapping in range and potentially hybridizing with Middle-Eastern gray wolves and African canid lineages in the past.

Our results support the hypothesis that the Sinai Peninsula and Southwest Levant are important hubs of canid evolution, where pervasive interspecific hybridization has been detected among gray wolves, African wolves, and Eurasian golden jackals ( Koepfli et al. 2015 ; Gopalakrishnan et al. 2018 ). Assemblages of Early Pleistocene mammalian fossils from the Pinjor Formation in India, including remains of at least two species of Canis , suggest paleobiogeographic linkages with African and Middle Eastern faunas ( Patnaik and Nanda 2010 ). The connections between the faunas of India and Africa are also supported by the vertebrate fossil records from Late Pleistocene deposits in Gujurat, which includes a Canis sp. that is larger and more robust than the present-day Indian wolf ( Costa 2017 ), and from other taxa, as Asiatic lions in India have experienced extensive gene flow with African lions ( de Manuel et al. 2020 ), and African leopards are known to have admixed with leopards from the Middle East (Palestine region) and Central Asia (Afghanistan) ( Paijmans et al. 2021 ). Our model is, of course, speculative, and additional data from both fossils and living animals will be helpful to understand the history of admixture among these canid lineages.

Intriguingly, D -statistics tests of allele sharing between IW01 and African canids revealed the Himalayan wolf as distinct from other wolf lineages ( fig. 5C–G ), leading us to hypothesize that the Himalayan wolf was less admixed. To test this, we computed D -statistics with the Himalayan wolf as H1, domestic dog and gray wolves as H2, and African wolf, Ethiopian wolf, African wild dog, or golden jackal as H3. All analyses resulted in significant positive D values ( Z > 3), suggesting that the domestic dog and gray wolves also shared excess derived alleles with African canids and golden jackals ( supplementary fig. S12 , Supplementary Material online ). This analysis provides support for the idea that present-day wolves and domestic dogs have admixed ancestries ( Fan et al. 2016 ; Frantz et al. 2016 ) and that the Himalayan wolf is relatively isolated (less or unadmixed with other canids) compared with other wolves ( fig. 5B ).

Demographic History and Divergence Time for the IW01 and Other Gray Wolves

To place the evolution of IW01 in a chronological context along with other gray wolves, we calculated relative cross-coalescence rates (CCR, the ratio between the cross- and the within-coalescence rates) for each pair of populations using the multiple sequentially Markovian coalescent (MSMC) model ( Schiffels and Durbin 2014 ), including genomes with a sequence coverage >20-fold. Using 50% CCR as a cutoff to estimate divergence time, these analyses suggest that IW01 diverged from domestic dogs and Chinese, Tibetan, European, and American wolves ∼110 ka ( fig. 7A ). This divergence date is much older than the previous estimates of ∼68–81 ka for divergence between the Tibetan wolf and domestic dog/East Asian gray wolves ( Wang et al. 2020 ) and supports our phylogenetic result that IW01 is basal to the Tibetan/Himalayan wolf and domestic dog + gray wolf clade ( figs. 2C and 4 ; supplementary figs. S5–S7 , Supplementary Material online ). This analysis also showed that IW01 split from SRS661487 (India) and SRS661488 (Iran) more recently, around 86 and 81 ka, respectively ( fig. 7A ), although these estimates will be impacted by the admixed ancestry of these three individuals ( fig. 4 ). We estimated that SRS661487 diverged from the domestic dog and Chinese wolf ∼68–85 ka and from European wolf ∼17 ka ( supplementary fig. S13 , Supplementary Material online ), and that SRS661487 separated from the Iranian wolf (SRS661488) ∼5.5 ka, consistent with these two samples clustering together in the phylogeny ( fig. 4 and supplementary figs. S6 and S7 , Supplementary Material online ). Therefore, SRS661487 likely represents a gray wolf that recently descended from Middle Eastern and European wolf lineages that then admixed with the IW01 lineage, whereas IW01 is a distinct and deeply diverged lineage.

Inferences of the time of divergence and demographic history of Indian wolf IW01 and other gray wolves. ( A ) MSMC estimation of splitting time for Indian wolves (IW01) from the domestic dog and other representative gray wolves. Gray-dashed vertical line indicates the estimated split time at ∼110 ka. ( B ) Results of PSMC analysis showing the demographic trajectories of seven representative gray wolves. For each sample, we performed 100 bootstrap replicates.

We used the pairwise sequentially Markovian coalescent (PSMC) model ( Li and Durbin 2011 ) to reconstruct historical patterns of effective population size over time for IW01 and other gray wolves with sequencing coverage ≥20-fold ( fig. 7B ). Generally, all gray wolves shared similar demographic trajectories up to ∼150 ka. Thereafter, IW01 and the Tibetan wolf diverged first around 110 ka and then experienced continuous contractions in population size. Generally consistent with MSMC and PSMC, we used Coal-HMM ( Mailund et al. 2011 ) and estimated that IW01 diverged from dogs and other gray wolves ∼130–140 ka ( supplementary fig. S14 , Supplementary Material online ). In contrast, SRS661487 shared a similar demographic trajectory with European, Iranian, and North American wolves whose population size expanded slightly between 100 and 50 ka, which was then followed by contraction ( fig. 7B ). These results corroborate that IW01 and SRS661487 represent two different gray wolf lineages.

To explore the recent history of the Indian wolf population, we examined nucleotide diversity and runs of homozygosity (ROH) for IW01 and compared this with the estimates from other gray wolves. Because such analyses are sensitive to genotyping errors, we focused on genomes with sequencing coverage ≥20-fold. IW01 had a nucleotide diversity of approximately 0.00104 ± 0.00098 (mean±SD), slightly higher than that of the Tibetan wolf, but lower than that estimated for European wolves, SRS661487, the Iranian wolf (SRS661488), the Mongolian wolf, and the North American wolf ( supplementary fig. S15 , Supplementary Material online ). IW01 had 11 blocks of ROH with a length >1 Mb, the longest of which was 1.57 Mb, whereas the Tibetan wolf had 48 blocks of ROH >1 Mb and 5 ROH >2 Mb ( supplementary fig. S16 , Supplementary Material online ). We found that 33% of the IW01 genome and 43% of the Tibetan wolf genome were homozygous, which was higher than that observed in other gray wolves except for the Chinese wolf ( supplementary fig. S16 , Supplementary Material online ). These results are consistent with the long-term small effective population sizes inferred in our PSMC analysis and with earlier ecological studies ( Aggarwal et al. 2003 , 2007 ; Sharma et al. 2004 ), and also suggest recent inbreeding.

Our results suggest that IW01 represents an evolutionarily distinct gray wolf lineage living in the semi-arid lowland region of the Indian subcontinent that diverged from other gray wolf populations ∼110 ka. IW01 shares ancestry with other gray wolves (SRS661487 and SRS661488) that fall within the geographic range described for C. l. pallipes . Consistent with our previous study, gray wolves from the Trans-Himalayan mountain range and Tibetan Plateau also carry deeply diverged ancestries ( Wang et al. 2020 ). The persistence of these ancient and diverged lineages in the Indian subcontinent may be due in part to the region’s unique topography and paleoenvironmental history. Similar patterns of locally divergent lineages have been observed in Trans-Himalayan red pandas ( Hu et al. 2020 ) and Chinese mountain cats ( Yu et al. 2021 ). Together, these findings point to the importance of the Indian subcontinent and Trans-Himalayan region as refugia during the Pleistocene ( Sharma et al. 2004 ; Costa 2017 ) that enabled the persistence of divergent lineages.

During the Pleistocene ice ages, the Indian subcontinent was dry and cold, and much of the Himalayan and Trans-Himalayan regions and southern Tibet ( Owen et al. 2002 ) were covered by ice. Regional unglaciated refugia persisted, however, within which small populations of gray wolves may have become isolated, leading to the evolution of distinct lineages ( Blinkhorn and Petraglia 2017 ). Our estimate of the timing of divergence between IW01 and other gray wolves coincides roughly with the end of the Last Interglacial period (Eemian), when warmer, wetter conditions occurred in the northern latitudes of Eurasia, whereas the Indian subcontinent and neighboring lower latitude regions experienced a cooler, drier climate ( Pedersen et al. 2017 ). These paleoclimatic differences, combined with geographic isolation, may have facilitated ecological and genetic divergence of the Indian wolf lineage.

Despite the relative isolation and small population size of Indian wolves today, we find that the IW01 lineage harbors evidence of a mosaic of past gene flow with the African wolf, Ethiopian wolf, African wild dog, and western Asian gray wolves. We also find that the Himalayan wolf shares significantly less admixed ancestry with modern-day African canids ( supplementary fig. S12 , Supplementary Material online ), which is consistent with its isolation and adaptation to the high-altitude arid environments of the Himalayan and Tibetan plateaus. It is possible that the distribution of gray wolves and African canids overlapped in the past, possibly in the Sinai Peninsula or Southwest Levant where several canid species are hypothesized to have hybridized ( Gopalakrishnan et al. 2018 ).

Our results present a scenario of pervasive gene flow between gray wolves and other canid species, adding to the growing evidence of the important role of interspecific hybridization in the evolution of canid species and populations specifically and the role of network-linked and reticulated evolution of species more generally. Although our study is based on a single sample of precisely known provenance, our analyses of IW01 bridge a data gap for gray wolves and provide an important resource for future studies. Additional sampling of Indian wolves from other regions of peninsular India, of other wolves from across the range of C. l. pallipes , and perhaps from ancient samples will be necessary to inform the conservation of this threatened and elusive gray wolf subspecies.

IW01: Origins and Sampling

The Indian wolf (IW01; fig. 1 and supplementary fig. S1 , Supplementary Material online ) sequenced for this study was captured in 1995 inside Velavadar Blackbuck National Park in Gujarat state, India (latitude = 22.0438°N, longitude = 72.0202°E), for a radio-telemetry-based ecological study of the species. The wolf was captured using a rubberized-jaw McBride foot-hold trap (Minnesota) and anesthetized using Telozol ( Kreeger et al. 1989 ). Whole blood was drawn from the brachial vein for DNA profiling and disease study. Permissions for capture and collaring were obtained from the Ministry of Environment and Forest, government of India, and from the Chief Wildlife Warden, Gujarat state. The whole blood sample was stored in alcohol at −20 °C until genomic DNA was extracted.

Genome Sequencing and Variant Calling

Four paired-end DNA sequencing libraries were prepared for IW01, resulting in a total of 311,789,040 paired-end 150-bp reads (corresponding to 93.5 Gb) generated by the M/s Xcelris Labs Ltd. Ahmedabad, Gujarat, India, using the Illumina HiSeq 2500 platform. We downloaded published genomic sequences from 30 other canid samples from the NCBI SRA (accession IDs are available in supplementary table S1 , Supplementary Material online ) including domestic dogs, African wild dog ( Lycaon pictus ), dhole (Cuon alpinus), coyote ( Canis latrans ), Eurasian golden jackal ( Canis aureus ), African wolf ( Canis lupaster ), Ethiopian wolf ( Canis simensis ), and Andean fox ( Lycalopex culpaeus ). We used Btrim ( Kong 2011 ) to remove low-quality bases. Because a highly contiguous chromosome-level reference genome assembly is not yet available for the gray wolf, we aligned the remaining reads to the domestic dog CanFam3.1 reference genome ( Lindblad-Toh et al. 2005 ) using the BWA-MEM algorithm ( Li 2014 ) with the settings “-t 4 –M.” We processed the bam alignment by coordinate sorting, marking duplicated reads, performed local realignment, and recalibrated base quality scores using the Picard (version 1.56; http://broadinstitute.github.io/picard/ , last accessed January 27, 2022) and GATK (version 3.7.0) packages ( McKenna et al. 2010 ). We called SNPs for all samples together using the UnifiedGenotyper function in GATK. To increase the reliability of the data, SNPs were further filtered as previously described ( Wang et al. 2020 ) using the VariantFiltration command in GATK with parameters: “QUAL < 40.0 MQ < 25.0 MQ0 ≥ 4 && ((MQ0/(1.0×DP)) > 0.1) cluster 3 -window 10.” Index, depth, and mapping statistics were computed using available tools in SAMtools v1.3.1 ( Li et al. 2009 ).

Mitochondrial Assembly and Phylogenetic Analysis

Because no complete mitochondrial genome is available in GenBank for the Indian wolf, we performed de novo assembly of the mitochondrial genome for IW01, SRS661487 (India), and SRS661488 (Iran) using NOVOPlasty v2.7.2 ( Dierckxsens et al. 2017 ) with a k-mer size of 31 based on whole-genome sequencing data. The domestic dog mitochondrial genome (GenBank accession: NC_002008.4 ) was used as a seed/reference sequence. We downloaded mitochondrial genomes for coyote, African dog, dhole, African wolf, and other gray wolves and domestic dogs from NCBI (GenBank accessions are shown in fig. 1C ) and included the Tibetan and Himalayan wolf sequences from a previous study ( Wang et al. 2020 ). A total of 39 mitogenomes were analyzed in this study. These sequences were aligned using MUSCLE v3.8.31 ( Edgar 2004 ) and the alignments were checked manually. After removing poorly aligned and control regions, an alignment file with a length of 15,462 bp was used for phylogenetic analysis. A maximum-likelihood tree was reconstructed using RAxML v8.2.12 ( Stamatakis 2014 ) with the GTR+G model of DNA substitution, and 1,000 bootstraps were run to assess node support.

We also downloaded previously reported mitochondrial cytochrome b and 16S rRNA sequences for Indian wolf, domestic dog, and other gray wolves from GenBank (accessions are shown in supplementary fig. S2 , Supplementary Material online ) and aligned and analyzed these data (554 bp for 16S rRNA and 332 bp for cytochrome b ) for phylogenetic analysis using the same methods described above.

Nuclear Phylogeny Construction

We constructed phylogenetic trees using nuclear genome sequences to explore the relationship of IW01 with other gray wolves and canid species. For each canid taxon, only one sample was used. Given that domestic dogs constitute a monophyletic clade ( Fan et al. 2016 ; Wang et al. 2016 ), we chose the high-coverage Dingo genome (31.3-fold; SRR7120191) to represent the domestic dog lineage. As a result, a total of 19 samples were used to construct phylogenetic trees ( fig. 2A and supplementary fig. S3 , Supplementary Material online ). We generated a consensus genome for each sample using ANGSD v0.931 ( Korneliussen et al. 2014 ) (-doFasta 1). Reads with a minimum mapping quality lower than 25 were discarded (-minMapQ 25). For genomes with an average sequencing depth of over or less than 10-fold, the minimum depth for each base was set to 4-fold (-setMinDepth 4) or to 3-fold (-setMinDepth 3), respectively. Additional filter parameters implemented were: -doCounts 1 -uniqueOnly 1 -nThreads 2. We selected 5,000 random regions with a length of 20 kb from across the genome of the domestic dog reference assembly and the other 18 canid taxa using the “random” function in BEDTools v2.28.0 ( Quinlan and Hall 2010 ) (-l 20,000 -n 5,000). Sequences for each region were retrieved using the “faidx” function of SAMtools v1.3.1 ( Li et al. 2009 ). For each region, a maximum-likelihood tree was constructed by RAxML v8.2.12 ( Stamatakis 2014 ) with 100 bootstrap replicates using the command: raxmlHPC-PTHREADS-SSE3 -x 12,345 -k -# 100 -p 321 -m GTRGAMMAI -T 4 -s myseq.fas -f a -n myseq.ml.tre. The 5,000 gene trees were then concatenated and used as input for ASTRAL-III v5.7.5 ( Zhang et al. 2018 ) to generate a species tree, using default parameters. We used DiscoVista ( Sayyari et al. 2018 ) to analyze the discordance frequencies between the ASTRAL species tree and the 5,000 gene trees.

We retrieved and concatenated genotypes for 31 samples (in VCF format) within regions containing the signal of diverged origin in high-altitude wolves (Himalayan and Tibetan wolves) ( Wang et al. 2020 ), and converted into .fas format files. A neighbor-joining tree was constructed using the mega-cc tool ( Kumar et al. 2012 ) in MEGA7 ( Kumar et al. 2016 ) and nodal support was evaluated with 1,000 bootstrap replicates. Lastly, following ( Wang et al. 2020 ), we split four high-coverage genomes from the Chinese wolf, IW01, Tibetan wolf, and dhole into 250-, 500-, and 1,000-kb windows across autosomes and constructed phylogenies for each window using TreeMix v1.13 ( Pickrell and Pritchard 2012 ) with dhole as the outgroup. The frequency of each topology was calculated using APE v5.5 ( Popescu et al. 2012 ).

PSMC Analysis

We used the PSMC model to infer historical demographic trajectories for the sampled gray wolves ( Li and Durbin 2011 ). We only analyzed genomes with coverage >20-fold to ensure the accurate calling of heterozygotes ( Nadachowska-Brzyska et al. 2016 ), although some studies used low coverage genomes with false negative rate corrections ( Kim et al. 2014 ; Hawkins et al. 2018 ). A diploid consensus sequence for each individual was generated using the “mpileup” command of the SAMtools package (v1.3.1) ( Li et al. 2009 ) with the option “-C50.” Variants with less than about 1/3 (“-d” option) or over two times (“-D” option) of average read depth were marked as missing and excluded from consensus sequence assignment. Sequences with consensus quality lower than 20 were also filtered out. The program “fq2psmcfa” from the PSMC package was used to convert the consensus sequences into 100-bp bin-input files for PSMC. We ran PSMC with parameters “-N25 -t15 -r5 -p 4 + 25×2 + 4 + 6.” A total of 100 bootstraps were analyzed for each sample. These PSMC estimates are scaled using a generation time ( g ) of 3 years and a mutation rate (µ) of 4e − 9 substitutions per site per generation as used previously ( Skoglund et al. 2015 ). This mutation rate was comparable to a recent estimation based on pedigree analysis ( Koch et al. 2019 ).

MSMC and Coal-HMM Inference of Splitting Time

We used the multiple sequential Markovian coalescent (MSMC2) model to infer the divergence time for the domestic dog and gray wolf population pairs ( Schiffels and Durbin 2014 ). Genotypes for all dogs and wolves were phased together using Beagle V.4.1 ( Browning and Browning 2016 ). The MSMC input files comprising four haplotypes (two individuals) were generated as suggested by the authors using available tools from the MSMC-tool package ( https://github.com/stschiff/msmc-tools , last accessed January 27, 2022). We ran MSMC for each pair of genomes using default settings and the time when the relative cross-coalescent rate was dropped to 50% as an approximate estimate of the splitting time ( Malaspinas et al. 2016 ). For each calculation, four haplotypes were analyzed, and estimations were scaled using a generation time ( g ) of 3 years and a mutation rate (µ) of 4e −9 substitutions per site per generation ( Skoglund et al. 2015 ). Similar to the PSMC analysis, we restricted this analysis to genomes with coverage >20-fold.

We also used Coal-HMM ( Mailund et al. 2011 ), a coalescent hidden Markov model-based approach, to measure the divergence time for the Indian wolf (IW01) and dog and other wolves. We performed estimation for each population pair using 1-Mb nonoverlapping sliding window segments across each chromosome. We filtered out windows with over 10% missing rate for such analysis. We also removed results for each segment where: 1) the recombination rate was lower 0.1 or over 10 cM/Mb, 2) the ancestral effective population size below 1,000 or above 1,000,000, and 3) the split time was below 1,000 years or above 1,000,000 years.

Nuclear Diversity and ROH Analysis

Nucleotide diversity (π) ( Nei and Li 1979 ) was calculated for each sample across the autosomes using VCFtools v0.1.13 ( Danecek et al. 2011 ) in 50-kb sliding windows with a step of 25 kb. ROH was calculated for each sample across the autosomes using the “roh” function in the BCFtools v1.4-7-g41827a3 ( Narasimhan et al. 2016 ) with default parameters.

TreeMix, ABBA-BABA, and AdmixtureGraph Analyses

To explore the phylogenetic relationships and admixture among gray wolves and other canid species, we also used TreeMix v1.13 ( Pickrell and Pritchard 2012 ) to construct maximum-likelihood tree graphs by allowing gene flow. TreeMix analysis was run for all variants located on autosomes using 1,000 variants per block (-k 1,000) and allowing zero to five migrations, with Andean fox used as the outgroup.

We used the ABBA-BABA test, also known as D -statistics ( Green et al. 2010 ) to detect the amount of allele sharing between gray wolf populations. This analysis is based on the topology (((H1, H2), H3), Outgroup) as shown in figure 5A . D = 0 suggests no gene flow between ingroup (H1 or H2) and H3; D > 0 suggests gene flow between H3 and H2; and D < 0 suggests gene flow between H3 and H1. We used the function “-doAbbababa 1” in ANGSD v0.931 ( Korneliussen et al. 2014 ) to perform this analysis with the additional settings “-doCounts 1 -minMapQ 25 -minQ 25 -uniqueOnly 1 -nThreads 6.”

To assess the genetic makeup and relationships among IW01, gray wolves, and three African canid species (African wolf, Ethiopian wolf, and African wild dog), we constructed admixture graph models using the qpGraph tool from AdmixTools package ( Patterson et al. 2012 ), the admixturegraph R package ( Leppala et al. 2017 ), and qpBrute ( Ni Leathlobhair et al. 2018 ; Liu et al. 2019 ). Because this analysis requires high-confidence genotype calls, we chose one sample with genome sequencing coverage over 10-fold from each population or species for constructing admixture graphs. To resolve the relationship between IW01, Himalayan/Tibetan wolves, and Eurasian gray wolves, we tested all possible graph models to fit all possible f4-statistics. The phylogenetic tree based on “ghost” admixed sequences and mitochondrial genomes from Himalayan or Tibetan wolves showed that the “ghost” lineage was basal to IW01. Therefore, we considered graphs in which Himalayan or Tibetan wolves were modeled as a product of admixture with one source from the lineage basal to IW01. To investigate the admixture between IW01 and African canids, we constructed admixture models starting with three populations (IW01, European wolf, and Himalayan or Tibetan wolf) and the fitted graph was then used as the base model in which we successively added each of the three African canid species.

Local Ancestry Inference

To identify potential admixed tracts along each chromosome in IW01, we performed local ancestry inference using PCAdmix ( Brisbin et al. 2012 ). We used phased genotypes as mentioned above as input, with IW01 designated as an admixed population and each of the African canid species, domestic dog, and Eurasian gray wolves as source populations. We performed two independent runs using 20 (default by the software) and 40 SNPs per window (“-w” parameter), respectively. The identified regions with posterior probabilities >0.9 were considered as potentially admixed.

Supplementary data are available at Genome Biology and Evolution online.

We thank the research communities for making their genomic data public, which makes this study possible. We also thank Y. Shah for the Indian wolf photos, and Robert Wayne for helpful discussions. This project was supported in part by DST-INSPIRE Faculty funding awarded to M.T. (04/2016/002246).

Author Contributions

M.T., Y.J., and B.S. conceived the idea and designed the research project. B.S., Y.J., K.-P.K., and R.E. G. supervised the analysis. M.-S.W., M.T., and S.W. performed the analysis with inputs from H.-M. C. and S.-S. D. Y.J., M.T., and Y.S. provided and coordinated genome sequencing of wolf sample IW01 within India. M.-S.W., M.T., and B.S. drafted the manuscript. B.S., Y.J., K.-P. K., M.-S.W., and M.T. revised the manuscript with input from all authors. Analysis of the Indian wolf sampled from western Gujarat, including sequencing and data analysis, was undertaken in India. Z.-X.L. submitted sequenced genome to NCBI. All authors read and improved the manuscript.

Data Availability

The genome sequencing raw reads were deposited in the NCBI-SRA database, under Bio-Project accession: PRJNA714797.

Aggarwal RK , Kivisild T , Ramadevi J , Singh L. 2007 . Mitochondrial DNA coding region sequences support the phylogenetic distinction of two Indian wolf species . J Zool Syst . 45 ( 2 ): 163 – 172 .

Google Scholar

Aggarwal RK , Ramadevi J , Singh L. 2003 . Ancient origin and evolution of the Indian wolf: evidence from mitochondrial DNA typing of wolves from trans-Himalayan region and Pennisular India . Genome Biol . 4 ( 6 ): P6 .

Allen GM. 1938 . The mammals of China and Mongolia. Natural history of Central Asia. XI, Part I . 1st ed. New York : The American Museum of Natural History . p. 342 – 345 .

Google Preview

Allman ES , Degnan JH , Rhodes JA. 2011 . Identifying the rooted species tree from the distribution of unrooted gene trees under the coalescent . J Math Biol . 62 ( 6 ): 833 – 862 .

Amiri Ghanatsaman Z , et al.  2020 . Whole genome resequencing of the Iranian native dogs and wolves to unravel variome during dog domestication . BMC Genomics 21 ( 1 ): 207 .

Blinkhorn J , Petraglia MD. 2017 . Environments and cultural change in the Indian Subcontinent: implications for the dispersal of Homo sapiens in the Late Pleistocene . Curr Anthropol . 58 ( S17 ): S463 – S479 .

Brisbin A , et al.  2012 . PCAdmix: principal components-based assignment of ancestry along each chromosome in individuals with admixed ancestry from two or more populations . Hum Biol . 84 ( 4 ): 343 – 364 .

Browning BL , Browning SR. 2016 . Genotype imputation with millions of reference samples . Am J Hum Genet . 98 ( 1 ): 116 – 126 .

Chavez DE , et al.  2019 . Comparative genomics provides new insights into the remarkable adaptations of the African wild dog ( Lycaon pictus ) . Sci Rep . 9 ( 1 ): 8329 .

Costa AG. 2017 . A new Late Pleistocene fauna from arid coastal India: implications for inundated coastal refugia and human dispersals . Quat Int . 436 : 253 – 269 .

Danecek P , et al.  2011 . The variant call format and VCFtools . Bioinformatics 27 ( 15 ): 2156 – 2158 .

Degnan JH , Rosenberg NA. 2009 . Gene tree discordance, phylogenetic inference and the multispecies coalescent . Trends Ecol Evol . 24 ( 6 ): 332 – 340 .

de Manuel M , et al.  2020 . The evolutionary history of extinct and living lions . Proc Natl Acad Sci U S A . 117 : 10927 – 10934 .

Dierckxsens N , Mardulyn P , Smits G. 2017 . NOVOPlasty: de novo assembly of organelle genomes from whole genome data . Nucleic Acids Res . 45 ( 4 ): e18 .

Edgar RC. 2004 . MUSCLE: multiple sequence alignment with high accuracy and high throughput . Nucleic Acids Res . 32 ( 5 ): 1792 – 1797 .

Ersmark E , et al.  2016 . From the past to the present: wolf phylogeography and demographic history based on the mitochondrial control region . Front Ecol Evol . 4 . doi: 10.3389/fevo.2016.00134 .

Fan Z , et al.  2016 . Worldwide patterns of genomic variation and admixture in gray wolves . Genome Res . 26 ( 2 ): 163 – 173 .

Frantz LA , et al.  2016 . Genomic and archaeological evidence suggest a dual origin of domestic dogs . Science 352 ( 6290 ): 1228 – 1231 .

Freedman AH , et al.  2014 . Genome sequencing highlights the dynamic early history of dogs . PLoS Genet . 10 ( 1 ): e1004016 .

Gopalakrishnan S , et al.  2018 . Interspecific gene flow shaped the evolution of the genus Canis . Curr Biol . 28 ( 21 ): 3441 – 3449.e3445 .

Green RE , et al.  2010 . A draft sequence of the Neandertal genome . Science 328 ( 5979 ): 710 – 722 .

Hawkins MTR , et al.  2018 . Genome sequence and population declines in the critically endangered greater bamboo lemur ( Prolemur simus ) and implications for conservation . BMC Genomics 19 ( 1 ): 445 .

Hodgson B. 1847 . Description of the wild ass and wolf of Tibet . Calcutta J Nat Hist . 7 : 469 – 477 .

Hu Y , et al.  2020 . Genomic evidence for two phylogenetic species and long-term population bottlenecks in red pandas . Sci Adv . 6 ( 9 ): eaax5751 .

Jhala Y. 2003 . Status, ecology and conservation of the Indian wolf Canis lupus pallipes sykes . J Bombay Nat Hist Soc . 100 : 293 – 307 .

Kim HL , et al.  2014 . Khoisan hunter-gatherers have been the largest population throughout most of modern-human demographic history . Nat Commun . 5 : 5692 .

Koch E , et al.  2019 . De novo mutation rate estimation in wolves of known pedigree . Mol Biol Evol . 36 ( 11 ): 2536 – 2547 .

Koepfli KP , et al.  2015 . Genome-wide evidence reveals that African and Eurasian golden jackals are distinct species . Curr Biol . 25 ( 16 ): 2158 – 2165 .

Kong Y. 2011 . Btrim: a fast, lightweight adapter and quality trimming program for next-generation sequencing technologies . Genomics 98 ( 2 ): 152 – 153 .

Korneliussen TS , Albrechtsen A , Nielsen R. 2014 . ANGSD: analysis of next generation sequencing data . BMC Bioinformatics 15 : 356 .

Kreeger TJ , Mandsager RE , Seal US , Callahan M , Beckel M. 1989 . Physiological response of gray wolves to butorphanol-xylazine immobilization and antagonism by naloxone and yohimbine . J Wildl Dis . 25 ( 1 ): 89 – 94 .

Kumar S , Stecher G , Peterson D , Tamura K. 2012 . MEGA-CC: computing core of molecular evolutionary genetics analysis program for automated and iterative data analysis . Bioinformatics 28 ( 20 ): 2685 – 2686 .

Kumar S , Stecher G , Tamura K. 2016 . MEGA7: molecular evolutionary genetics analysis version 7.0 for bigger datasets . Mol Biol Evol . 33 ( 7 ): 1870 – 1874 .

Leonard JA , et al.  2007 . Megafaunal extinctions and the disappearance of a specialized wolf ecomorph . Curr Biol . 17 ( 13 ): 1146 – 1150 .

Leppala K , Nielsen SV , Mailund T. 2017 . admixturegraph: an R package for admixture graph manipulation and fitting . Bioinformatics 33 ( 11 ): 1738 – 1740 .

Li H. 2014 . Toward better understanding of artifacts in variant calling from high-coverage samples . Bioinformatics 30 ( 20 ): 2843 – 2851 .

Li H , Durbin R. 2011 . Inference of human population history from individual whole-genome sequences . Nature 475 ( 7357 ): 493 – 496 .

Li H , et al.  2009 . The Sequence Alignment/Map format and SAMtools . Bioinformatics 25 ( 16 ): 2078 – 2079 .

Lindblad-Toh K , et al.  2005 . Genome sequence, comparative analysis and haplotype structure of the domestic dog . Nature 438 ( 7069 ): 803 – 819 .

Lipson M. 2020 . Applying f4-statistics and admixture graphs: theory and examples . Mol Ecol Resour . 20 ( 6 ): 1658 – 1667 .

Liu L , et al.  2019 . Genomic analysis on pygmy hog reveals extensive interbreeding during wild boar expansion . Nat Commun . 10 : 1992 .

Liu Y-H , et al.  2018 . Whole-genome sequencing of African dogs provides insights into adaptations against tropical parasites . Mol Biol Evol . 35 ( 2 ): 287 – 298 .

Loog L , et al.  2020 . Ancient DNA suggests modern wolves trace their origin to a late Pleistocene expansion from Beringia . Mol Ecol . 29 ( 9 ): 1596 – 1610 .

Mailund T , Dutheil JY , Hobolth A , Lunter G , Schierup MH. 2011 . Estimating divergence time and ancestral effective population size of Bornean and Sumatran orangutan subspecies using a coalescent hidden Markov model . PLoS Genet . 7 ( 3 ): e1001319 .

Malaspinas AS , et al.  2016 . A genomic history of Aboriginal Australia . Nature 538 ( 7624 ): 207 – 214 .

McKenna A , et al.  2010 . The Genome Analysis Toolkit: a MapReduce framework for analyzing next-generation DNA sequencing data . Genome Res . 20 ( 9 ): 1297 – 1303 .

Nadachowska-Brzyska K , Burri R , Smeds L , Ellegren H. 2016 . PSMC analysis of effective population sizes in molecular ecology and its application to black-and-white Ficedula flycatchers . Mol Ecol . 25 ( 5 ): 1058 – 1072 .

Narasimhan V , et al.  2016 . BCFtools/RoH: a hidden Markov model approach for detecting autozygosity from next-generation sequencing data . Bioinformatics 32 ( 11 ): 1749 – 1751 .

Nei M , Li WH. 1979 . Mathematical model for studying genetic variation in terms of restriction endonucleases . Proc Natl Acad Sci U S A . 76 ( 10 ): 5269 – 5273 .

Ni Leathlobhair M , et al.  2018 . The evolutionary history of dogs in the Americas . Science 361 ( 6397 ): 81 – 85 .

Nowak RM. 1979 . North American quaternary Canis . Lawrence (KS) : Museum of Natural History, University of Kansas .

Nowak RM. 1995 . Another look at wolf taxonomy. In: Carbyn LN , Fritts SH , Seip DR , editors. Ecology and conservation of wolves in a changing world . Edmonton (AB) : Canadian Circumpolar Institute . p. 375 – 398 .

Owen LA , Finkel RC , Caffee MW. 2002 . A note on the extent of glaciation throughout the Himalaya during the global Last Glacial Maximum . Quat Sci Rev . 21 ( 1–3 ): 147 – 157 .

Paijmans JLA , et al.  2021 . African and Asian leopards are highly differentiated at the genomic level . Curr Biol . 31 ( 9 ): 1872 – 1882 .

Patnaik R , Nanda AC. 2010 . Early Pleistocene mammalian faunas of India and evidence of connections with other parts of the world. In: Fleagle JG , Shea JJ , Grine FE , Baden AL , Leakey RE , editors. Out of Africa I . Dordrecht (The Netherlands) : Springer Netherlands . p. 129 – 143 .

Patterson N , et al.  2012 . Ancient admixture in human history . Genetics 192 ( 3 ): 1065 – 1093 .

Pedersen RA , Langen PL , Vinther BM. 2017 . The last interglacial climate: comparing direct and indirect impacts of insolation changes . Clim Dyn . 48 ( 9–10 ): 3391 – 3407 .

Perri AR , et al.  2021 . Dire wolves were the last of an ancient New World canid lineage . Nature 591 ( 7848 ): 87 – 91 .

Pickrell JK , Pritchard JK. 2012 . Inference of population splits and mixtures from genome-wide allele frequency data . PLoS Genet . 8 ( 11 ): e1002967 .

Pilot M , et al.  2019 . Global phylogeographic and admixture patterns in grey wolves and genetic legacy of an ancient Siberian lineage . Sci Rep . 9 ( 1 ): 17328 .

Popescu AA , Huber KT , Paradis E. 2012 . ape 3.0: new tools for distance-based phylogenetics and evolutionary analysis in R . Bioinformatics 28 ( 11 ): 1536 – 1537 .

Quinlan AR , Hall IM. 2010 . BEDTools: a flexible suite of utilities for comparing genomic features . Bioinformatics 26 ( 6 ): 841 – 842 .

Sayyari E , Whitfield JB , Mirarab S. 2018 . DiscoVista: interpretable visualizations of gene tree discordance . Mol Phylogenet Evol . 122 : 110 – 115 .

Schiffels S , Durbin R. 2014 . Inferring human population size and separation history from multiple genome sequences . Nat Genet . 46 ( 8 ): 919 – 925 .

Sharma DK , Maldonado JE , Jhala YV , Fleischer RC. 2004 . Ancient wolf lineages in India . Proc Biol Sci . 271 ( Suppl 3 ): S1 – S4 .

Sinding MS , et al.  2018 . Population genomics of grey wolves and wolf-like canids in North America . PLoS Genet . 14 ( 11 ): e1007745 .

Skoglund P , Ersmark E , Palkopoulou E , Dalen L. 2015 . Ancient wolf genome reveals an early divergence of domestic dog ancestors and admixture into high-latitude breeds . Curr Biol . 25 ( 11 ): 1515 – 1519 .

Smeds L , Kojola I , Ellegren H. 2019 . The evolutionary history of grey wolf Y chromosomes . Mol Ecol . 28 ( 9 ): 2173 – 2191 .

Stamatakis A. 2014 . RAxML version 8: a tool for phylogenetic analysis and post-analysis of large phylogenies . Bioinformatics 30 ( 9 ): 1312 – 1313 .

Sykes WH. 1831 . Catalogue of the Mammalia of Dukun (Deccan); with observations on the habits, etc., and characters of new species . Proc Zool Soc Lond . 1830–1831 : 99 – 106 .

Thalmann O , et al.  2013 . Complete mitochondrial genomes of ancient canids suggest a European origin of domestic dogs . Science 342 ( 6160 ): 871 – 874 .

Toews DP , Brelsford A. 2012 . The biogeography of mitochondrial and nuclear discordance in animals . Mol Ecol . 21 ( 16 ): 3907 – 3930 .

vonHoldt BM , et al.  2016 . Whole-genome sequence analysis shows that two endemic species of North American wolf are admixtures of the coyote and gray wolf . Sci Adv . 2 ( 7 ): e1501714 .

Wang GD , et al.  2016 . Out of southern East Asia: the natural history of domestic dogs across the world . Cell Res . 26 ( 1 ): 21 – 33 .

Wang MS , et al.  2020 . Ancient hybridization with an unknown population facilitated high-altitude adaptation of canids . Mol Biol Evol . 37 ( 9 ): 2616 – 2629 .

Werhahn G , et al.  2018 . The unique genetic adaptation of the Himalayan wolf to high-altitudes and consequences for conservation . Glob Ecol Conserv . 16 : e00455 .

Werhahn G , et al.  2017 . Phylogenetic evidence for the ancient Himalayan wolf: towards a clarification of its taxonomic status based on genetic sampling from western . R Soc Open Sci . 4 ( 6 ): 170186 .

Wilson DE , Reeder DM. 2005 . Mammal species of the world: a taxonomic and geographic reference . Baltimore (MD): Johns Hopkins University Press .

Yu H , et al.  2021 . Genomic evidence for the Chinese mountain cat as a wildcat conspecific ( Felis silvestris bieti ) and its introgression to domestic cats . Sci Adv . 7 : eabg0221 .

Zhang C , Rabiee M , Sayyari E , Mirarab S. 2018 . ASTRAL-III: polynomial time species tree reconstruction from partially resolved gene trees . BMC Bioinformatics 19 ( S6 ): 153 .

Author notes

Supplementary data.

Email alerts

Citing articles via, affiliations.

• Online ISSN 1759-6653
• Copyright © 2024 Society for Molecular Biology and Evolution
• About Oxford Academic
• Publish journals with us
• University press partners
• What we publish
• New features  
• Open access
• Institutional account management
• Rights and permissions
• Get help with access
• Accessibility
• Advertising
• Media enquiries
• Oxford University Press
• Oxford Languages
• University of Oxford

Oxford University Press is a department of the University of Oxford. It furthers the University's objective of excellence in research, scholarship, and education by publishing worldwide

• Copyright © 2024 Oxford University Press
• Cookie settings
• Cookie policy
• Privacy policy
• Legal notice

This Feature Is Available To Subscribers Only

Sign In or Create an Account

This PDF is available to Subscribers Only

For full access to this pdf, sign in to an existing account, or purchase an annual subscription.

ORIGINAL RESEARCH article

Distribution, status, and conservation of the indian peninsular wolf.

A correction has been applied to this article in:

Corrigendum: Distribution, status, and conservation of the Indian peninsular wolf

• Read correction

• 1 Wildlife Institute of India, Dehradun, India
• 2 Department of Wildlife Sciences, Aligarh Muslim University, Aligarh, India

An understanding of the distribution range and status of a species is paramount for its conservation. We used photo captures from 26,838 camera traps deployed over 121,337 km 2 along with data from radio-telemetry, published, and authenticated wolf sightings to infer wolf locations. A total of 3,324 presence locations were obtained and after accounting for spatial redundancy 574 locations were used for modeling in maximum entropy framework (MaxEnt) with ecologically relevant covariates to infer potentially occupied habitats. Relationships of wolf occurrence with eco-geographical variables were interpreted based on response curves. Wolves avoided dense wet forests, human disturbances beyond a threshold, arid deserts, and areas with high top-carnivore density, but occurred in semi-arid scrub, grassland, open forests systems with moderate winter temperatures. The potential habitat that can support wolf occupancy was 364,425 km 2 with the largest wolf habitat available in western India (Saurashtra-Kachchh-Thar landscape 102,837 km 2 ). Wolf habitats across all landscapes were connected with no barriers to dispersal. Breeding packs likely occurred in ≈89,000 km 2 . Using an average territory size of 188 (SE 23) km 2 , India could potentially hold 423–540 wolf packs. With an average adult pack size of 3 (SE 0.24), and a wolf density < 1 per 100 km 2 in occupied but non-breeding habitats, a wolf population of 3,170 (SE range 2,568–3,847) adults was estimated. The states of Madhya Pradesh, Rajasthan, Gujarat, and Maharashtra were major strongholds for the species. Within forested landscapes, wolves tended to avoid top-carnivores but were more sympatric with leopards and dhole compared to tigers and lions. This ancient wolf lineage is threatened by habitat loss to development, hybridization with dogs, fast-traffic roads, diseases, and severe persecution by pastoralists. Their status is as precarious as that of the tiger, yet focused conservation efforts are lacking. Breeding habitat patches within each landscape identified in this study should be made safe from human persecution and free of feral dogs so as to permit packs to breed and successfully recruit individuals to ensure wolf persistence in the larger landscape for the long term.

Introduction

Reliable information on the status, that is the distribution, population size, extent, and habitat contiguity between populations, are essential for the management of any endangered species ( Sousa-Silva et al., 2014 ). This basic information is not available for many species, and conservation management is often based on educated guesses that can have direr consequences ( Blake and Hedges, 2004 ) and is especially relevant for threatened species that occur outside of protected areas ( Maron et al., 2018 ; Simmonds and Watson, 2019 ). Carnivores, due to their wide-ranging behavior, low density, and elusive nature, are one of the most difficult taxa to study ( Garshelis, 1992 ). The status of many carnivores was assessed from indices, such as pug-marks for tigers and lions ( Wynter-Blyth and Dharmakumarsinhji, 1949 ; Choudhary, 1970 ), simulated howls for wolves ( Harrington and Mech, 1982 ), and golden jackals ( Graf and Hatlauf, 2021 ), questionnaire surveys, and interactions with the local community ( Jhala and Giles, 1991 , Karanth et al., 2009 ). In the absence of any better approach, the information generated by these methods was often used for policy decisions and management actions. However, now with the advent of cost-effective modern technologies, such as camera traps and radio-telemetry, and analytical approaches, i.e., species distribution models ( Sousa-Silva et al., 2014 ), better insights on species distribution and abundance and their determining factors are possible.

Indian peninsular wolves ( Canis lupus pallipes ) are an ancient lineage of wolves endemic to the Indian sub-continent ( Sharma L. K. et al., 2004 ; Hennelly et al., 2021 ). They are considered endangered and are listed on Schedule 1 of the Wildlife Protection Act (1972) . Several attempts have been made to evaluate their status locally ( Jhala and Giles, 1991 ; Kumar and Rahmani, 1997 ; Singh and Kumara, 2006 ) and at the country scale ( Shahi, 1982 ; Jhala, 2003 ; Karanth et al., 2009 ; Srivathsa et al., 2020 ). Earlier range maps and population estimates were based on ground surveys, information from local pastoralists, and knowledge of wolf ecology and their habitat ( Shahi, 1982 ; Jhala and Giles, 1991 ; Kumar and Rahmani, 1997 ; Kumar, 1998 ; Kumar and Rahmani, 2000 ; Jethva and Jhala, 2004 ; Singh and Kumara, 2006 ; Kumar and Rahmani, 2008 ; Agarwala et al., 2010 ). Karanth et al. (2009) used expert knowledge, while Srivathsa et al. (2020) used a combination of data from field surveys, citizen science, and authenticated reports, while both studies used occupancy framework with eco-geographical and human footprint covariates to model wolf distribution across India.

In this study, we used data generated from the largest camera trap survey to date covering 121,337 km 2 ( Jhala et al., 2020 ) in combination with wolf locations obtained from radio-telemetry and authenticated records as presence data to model species distribution. We subsequently estimate population size based on territory size and pack size estimates in occupied and breeding habitats. We evaluate wolf distribution and relative abundance with respect to other large competing carnivores and identify wolf stronghold populations that should be targeted for conservation to ensure wolf persistence in the larger landscape for the long term.

Materials and Methods

The geographical extent of our study covered the entire range of Indian wolves within India. We modeled wolf distribution using the maximum entropy approach in maximum entropy framework (MaxEnt; version 3.4.1, Phillips et al., 2006 ) that uses machine learning from occurrence locations of the target species and background points along with ecologically relevant spatial environmental variables to develop statistical relationships ( Elith et al., 2011 ). These relationships are then used to predict species occurrence across modeled space ( Elith et al., 2011 ). We used a combination of methods to infer wolf presence locations. These were (a) extensive coverage of forested habitats across 20 Indian states by camera traps carried out by State Forest Department personnel and research biologists of the Wildlife Institute of India ( Jhala et al., 2020 ). Camera traps with heat and motion detectors were deployed at 26,838 locations in 2018–2019 to cover a forested area of 121,337 km 2 ( Figure 1 ). All photo captures of wildlife were geotagged and subsequently segregated into species. Camera trap locations that recorded wolf captures were used for modeling wolf distribution. (b) Since Indian peninsular wolves were known to use agro-pastoral landscapes (outside of forest habitats; Jhala, 1993 ) and since these areas were not camera trapped, we obtained records of wolf presence from Shahi (1982) , Jhala (1993 , 2003 , 2007) , Jhala and Sharma (1997) , Kumar and Rahmani (1997) , Jethva (2003) , Habib (2007) , Lokhande and Bajaru (2013) , Saren et al. (2019) , Ghaskadbi et al. (2021) , Mahajan and Khandal (2021) , Maurya et al. (2021) , Sadhukhan et al. (2021) , Sharma (2021) , and Trivedi et al. (2021) , and from radio-telemetry ( Jhala, 2007 ) and geotagged records from Jhala Y.V. et al. (2021) to augment the camera trap data.

Figure 1. Wolf distribution modeled in MaxEnt using presence locations with eco-geographical variables of human modification, climate, habitat, competing carnivores, and prey. Map inset shows region within India where wolf distribution was modeled.

Since many of the radio-telemetry-based locations and other locations were clumped, we picked only one location for approximately every 5 km 2 . This reduced the spatial redundancy of information in location data and we were left with 571 locations that were used for model building. Based on knowledge of wolf ecology and behavior ( Mech, 1970 ; Jhala, 1993 ; Mech and Boitani, 2007 ), we hypothesized a priori that Peninsular Indian wolves would occur in semi-arid grasslands, scrub, and open forests with high ambient temperatures, would avoid areas of high human density but occur in rural areas with livestock husbandry, and would avoid areas having a high density of competing carnivores. The eco-geographical variables used in MaxEnt were as follows: (a) habitat characteristics (land use land cover, Normalized Difference Vegetation Index (NDVI), elevation, and ruggedness; (b) climatic factors (temperatures of coldest and hottest months, rainfall, and aridity); (c) human footprint indices (distance to night light, distance to roads, road density, and human modification index; (d) prey indices as livestock density, goat and sheep density, and cattle density, and (e) top-carnivore density (tiger and lion density across their range of occurrence) ( Supplementary Table 1 ). Linear, quadratic, and product features available in MaxEnt were used in combination with representative variables from each of the above-mentioned eco-geographical variable categories. The models were assessed based on area under the curve (AUC) of receiver operator curves (ROC), specificity and sensitivity of the models, and testing the model classification accuracy on 30% of the data that were not used for model building ( Jiménez-Valverde, 2011 ). Best models were selected on the basis of model fit and parsimonious use of relevant ecological covariates that made ecological sense based on our a priori expectations ( Supplementary Table 1 ). We used clog-log analysis ( Phillips et al., 2017 ) to determine the probability value beyond which pixels had high wolf occurrence classification and below which wolves were likely absent to determine the area occupied by wolves. We also determined the pixel probabilities for 16 known breeding packs from 14 different areas spread across India and used one SD on the mean pixel values to address uncertainty in the cutoff values to determine occupied and breeding habitats.

Wolves are known to be territorial where neighboring territory areas overlap minimally ( Jhala, 2003 ; Habib, 2007 ). Since 100% Minimum Convex Polygon territories of four wolf packs reported by Habib (2007) did not differ from 95% fixed kernel estimates of another eight radio-collared packs from three different sites ( Jhala, 2007 ) ( t -test, p = 0.9) we combined these estimates for our analysis to get better coverage of territory sizes from across India ( Supplementary Table 2 ). We removed isolated wolf occurrence habitat patches that were <100 km 2 from further analysis as these would be too small to harbor wolves. We used data from 35 wolf packs for estimating adult pack size ( Supplementary Table 3 ) to estimate the potential wolf population within areas of breeding habitat. Occupied areas outside of breeding habitats would hold dispersing individuals, old ousted pack members, and sub-adults biding their time to join packs or form their own packs ( Packard and Mech, 1980 ). For areas that were above the MaxEnt clog-log probability value of occurrence but below the threshold of breeding packs, we used a conservative estimate of wolf density of less than one wolf per 100 km 2 (range between 0.75 and 0.5 wolves per 100 km 2 ).

To get a better understanding of species interactions within forested habitats, we computed relative abundance index (RAI, Carbone et al., 2001 ) as the number of photo captures per 100 trap days of wolves, dhole, leopards, and tigers and averaged these for all camera traps in 25 km 2 grids. We plotted wolf RAI against dhole RAI, leopard density, and tiger density from Jhala et al. (2020) and Jhala Y.V. et al. (2021) and inspected scatterplots, fitted models, and tested for linear correlations to better understand species interactions.

We obtained 34,858,623 photographs of wildlife from which 2,812 were of wolves from 313 camera locations. Published (34), other geo-tagged records (365), and radio-telemetry (2,612) contributed to a total of 3,324 wolf presence locations from across the range of the species in India ( Figure 1 ). The best MaxEnt model was a good fit with an AUC of 0.83 and performed well in classifying 30% of the test data ( Figure 2 ). Wolf occurrence was best explained by (1) climatic variables: (a) average rainfall, (b) average temperature of the coldest quarter; (2) habitat characteristics: (a) pre-monsoon NDVI, (c) land use and land cover; (3) Human Modification Index (maximum contribution to the model 40%); (4) prey availability in the form of livestock density; and (5) density of top-carnivores ( Figure 2 ). As per our a priori predictions, wolves were tolerant of higher temperatures ( Figure 2 and Supplementary Figure 1 ), they preferentially occurred at semi-arid sites that had lower rainfall, higher temperatures, lower values of canopy cover (NDVI), avoided high human densities but their occurrence coincided with moderate livestock densities. As expected, the response of wolves to top-carnivore density was a right-skewed bell-shaped function, with wolves occurring in areas of low top-carnivore densities but declining at high top-carnivore densities ( Figure 2 and Supplementary Figure 1 ).

Figure 2. Response curves of wolf occurrence with eco-geographical variables, their contributions, and model fit assessment obtained from 100 bootstrap runs of the best MaxEnt model. (A) Variation in the omission of model data and predicted area with increasing MaxEnt cumulative threshold values. (B) Receiver operating curve of test and training data. (D) Land use land cover classes were (1) arid scrub, (3) grassland, (4) agriculture, (5) settlement, (6) open, (8) water, (9) riparian, (10) evergreen open, (11) evergreen broadleaf, (12) deciduous broadleaf, (13) deciduous open, (14) mixed open, (15) evergreen broadleaf open, (16) deciduous broadleaf open, (17) scrub, and (18) coastal marsh. (E) Normalized Difference Vegetation Index (NDVI). (F) Carnivore density (density of tigers and lions) across their range in India.

Wolf territory size was estimated at 189 (SE 23) km 2 ( Supplementary Table 2 ). The total area above the threshold value obtained from clog-log analysis ( p = 0.47 SE 0.0094) that could potentially be occupied by wolves after removing isolated areas that were smaller than 100 km 2 was 364,425 km 2 in India. The largest potential for wolf occupancy was in the contiguous Saurashtra-Kachchh-Thar landscape (102,837 km 2 , Figure 3 ). Area suitable for breeding packs was estimated at 89,138 km 2 with the largest contiguous breeding habitats available in the Central Indian landscape (37,323 km 2 , Figure 3 ). Considering an average adult pack size of 3 (SE 0.24) adult wolves ( Supplementary Table 3 ) for breeding habitat and a density range from 0.75 to 0.5 wolves per 100 km 2 for occupied areas outside of the breeding habitat, the potential number of wolves in India was estimated at 3,170 (SE range 2,568–3,847). Besides the Saurashtra-Kachchh-Thar landscape, the other habitat patch that could potentially hold a population of > 150 wolves was Udanti Sitanadi-Indravati-Kawal-Tadoba ( Figure 3 ). Shivpuri-Mukundara-Gandhi Sagar, Satpura-Betul-Melghat, Bandhavgarh-Sanjay, and Panna-Nauradehi were other areas that support good wolf populations. Madhya Pradesh supported the largest wolf population followed by the states of Rajasthan, Gujarat, and Maharashtra ( Table 1 ).

Figure 3. Wolf-occupied landscapes and breeding habitats across India inferred from the MaxEnt models. MaxEnt, maximum entropy framework.

Table 1. State-wise estimated wolf population based on the MaxEnt model estimate of potential occupied, breeding habitat, average pack size of 3 (SE 0.24), and territory size of 188 (SE 23) km 2 .

Scatter plots of wolf RAI against dhole RAI, leopard, and tiger density categories in forested habitats ( Supplementary Figure 2 ) showed that wolf relative abundance declined with an increase in competing for carnivore relative and absolute abundances. Declines in wolf photo-capture rates were sharper and statistically significant with an increase in tigers compared to that of leopard and dhole.

Assessing the status of widespread, low density, and elusive species, such as the wolf, is a difficult task ( Kunkel et al., 2005 ). Shahi (1982) estimated the Indian wolf population at ≈800 individuals, while subsequent estimates were higher (2,000–3,000; Jhala, 2003 ) due to a better understanding of wolf distribution and ecology. The current assessment uses robust quantitative information of occurrence data (from large-scale geo-tagged camera trap, telemetry, and authenticated sightings) in combination with species distribution models with relevant eco-geographic covariates to evaluate wolf status. We use clog-log models with 100 bootstrap runs in MaxEnt ( Phillips et al., 2017 ) to determine the threshold probability below which wolf occurrence was unlikely, to determine wolf-occupied area. Estimates based on models are only as good as the data used to build these models; with an extensive coverage of wolf location data from across their range, from varied habitats, and eco-climatic conditions, we believe that our model predictions are good (as also shown by model evaluation statistics). However, we caution that due to the clog-log threshold used to determine wolf occupancy, there will be some areas where wolves may be present and our model threshold failed to predict them or predicated wolf occupancy in areas of known absence. We believe that at the country scale, these small errors would not matter, but at local scales where conservation measures need to be implemented, deviations from the truth would make a large difference. Therefore, the wolf habitat suitability map provided in this article should be used as a first cut and subsequent ground validation of the model results eventually used for conservation investments and management. The current distribution ( Figures 1 , 2 ) and population estimate ( Table 1 ) are similar to earlier estimates and validate Jhala (2003) with better information and formal model-based analysis. In the past two decades, wolf populations seem to have colonized new areas while losing out in some of their strongholds. Wolves have been recently recorded from several areas from where they had been exterminated or were not known to exist in the recent past [e.g., Rajaji Tiger Reserve ( Sharma, 2021 ), Bangladesh ( Muntasir et al., 2021 ), Indian Sundarbans ( Ghai, 2017 ), Valmiki Tiger Reserve ( Maurya et al., 2021 ), and Kaveri Wildlife Sanctuary ( Gubbi et al., 2020 )]. While wolves have declined from their stronghold of Kachchh and parts of Rajasthan primarily due to persecution, hybridization with dogs, and development of fast traffic roads. The easternmost limit of the Indian wolf was the Sundarban mangrove forest ( Ghai, 2017 ; Muntasir et al., 2021 ), there were no records of the Indian wolf from Assam and the North East States. No suitable occupied habitat was predicted in the states of Haryana and Punjab, possibly due to extensive and intensive agriculture, yet it is possible that wolves can also sporadically occur in these two states. It was believed that Indian peninsular wolves rarely used forested habitats ( Jhala, 2003 ), however, as evidenced from the extensive camera trap data, wolves have been recorded from several forested areas of India ( Figure 1 ). Notably, the tiger reserves of Mukundara, Kawal, Udanti Sitanadi, Melghat, Panna, Palamau, Bor, Kanha, Satpura, and Pench had a good number of wolf photo captures. Wolf photo captures from these tiger reserves were either from the buffer zone or from parts of the reserve that had relatively open canopied forests and scrubland habitats, and these parts had a relatively low density of tigers. Conserving a large carnivore outside of the realms of a protected area, especially when it has the propensity of predation on livestock, is a formidable task despite being protected by law ( Woodroffe et al., 2006 ). Protected areas targeting wolves as a focal species for conservation were few (e.g., Mahuadanr, Hazaribagh, Gandhi Sagar, and Nauradehi wildlife sanctuaries). Therefore, documenting breeding wolf populations in some well-protected areas of India heralds well for the long-term conservation of Indian wolves. Earlier estimates of wolves from Gujarat and Rajasthan ( Jhala and Giles, 1991 ) mapped their distribution and abundance based on extensive ground surveys and expert knowledge of local pastoral communities. These estimates were lower than the estimates reported herein. The MaxEnt-based analysis identifies habitats that meet the requirements for wolf occupancy based on the covariates used to build the model, human persecution can severely deplete wolf populations within suitable habitats as has been observed in Kachchh in recent times. Therefore, detailed ground surveys and radio-telemetry-based estimates of pack size, territory configurations, and sizes in selected sites are required to validate the population estimates obtained by model-based inference and for monitoring long-term population trends. Telemetry studies from mid-1990s to 2005 in the Bhal and Kachchh regions of Gujarat and Nashik ( Jhala, 2007 ) and Sholapur in Maharashtra ( Kumar and Rahmani, 1997 ; Habib, 2007 ; Habib et al., 2021 ) have shown that wolf populations were vulnerable to disease and persecution and fluctuated substantially ( Jhala, 2003 ). Unfortunately, no long-term telemetry-based studies are being implemented on the Indian wolves at specific sites to monitor population dynamics. Source populations of wolves within each of the identified landscapes need to be monitored continuously through radio-telemetry to keep the pulse of the population , i.e., ensure that these populations are not declining, and if declining, identify site-specific threats so as to address them in a timely manner. As long as these source populations are secure within each landscape, they will recruit wolves that will disperse and occupy the larger landscapes. Efforts to reintroduce wolves from captive-bred zoo populations should only be considered after appropriate rewilding, evaluation of their behavior, and skills of hunting wild prey. Such wolves (if habituated to humans) can become a major cause of human-wolf conflict ( Jhala and Sharma, 1997 ; Rajpurohit, 1999 ) and compromise the conservation of the entire species due to community backlash ( Treves et al., 2006 ).

Response curves of wolf occurrence to eco-geographical covariates were in consonance with our hypothesis conforming to their behavioral ecology. Besides climatic and habitat characteristics, top carnivore densities contributed (12.6%) to explaining wolf occurrence. It has long been speculated that Indian wolves have likely been out-competed by other large carnivores that dwell in forested habitats ( Jhala, 1993 ). The alternative hypothesis could be that Indian wolves evolved at a time when India was undergoing a dry spell ( Sharma D. K. et al., 2004 ; Hennelly et al., 2021 ) and adapted to open semi-arid habitats and therefore now avoid thick forests. Wolves often occurred in the buffer zones of protected areas, but were rarely seen within the core areas of PA’s that have high large carnivore densities even though habitats were suitable. For example, the habitats of Gir Protected Area and that of Ranthambore National Park were suitable for wolves (dry open canopied deciduous and thorn forests) and wolves occurred in the periphery of these reserves, but they were rarely seen in the core areas that have high lion and tiger densities, though these core areas abound in prey. While in protected areas, namely, Nauradehi, Gandhi Sagar, and Mukundara, that have similar habitats but do not have tigers or lions and dhole, wolves use most parts of these protected areas. These observations suggest that though Indian wolves may have specialized for open habitats, they were also likely limited by direct competition with other large carnivores. Since we had density estimates of only tigers and lions covering the full extent of these carnivores’ range across India, we could use these for modeling wolf occurrence in MaxEnt ( Figure 2 ). However, wolves were also likely limited by leopards and dhole. Leopards occur outside of forests as well ( Daniel, 1996 ), while dholes are primarily forest dwellers ( Johnsingh and Acharya, 2013 ) in tropical India. Since leopard, dhole, and wolf photo capture rates were available only from forested habitats, we restricted our analysis on their interactions to this habitat that was extensively camera trapped across India ( Jhala et al., 2020 ). Wolves tended to avoid all three competing large carnivores but were more tolerant of leopards and dhole compared to tigers ( Supplementary Figure 2 ).

The peninsular Indian wolf is an ancient lineage endemic to the Indian sub-continent ( Sharma L. K. et al., 2004 ; Hennelly et al., 2021 ), its status is precarious and with only ≈3,100 adult individuals their population is as big as that of the tiger in India ( Jhala Y. et al., 2021 ). Wolves are persecuted by pastoralists, threatened by diseases spread by dogs, and genetically swamped by a large feral dog population ( Jhala, 2003 ; Vanak and Gompper, 2009 ; Srivathsa et al., 2019 ). Conserving wolves is a more formidable task compared to tigers, since the majority of their population resides outside the realm of protected areas and there are currently no focused efforts for conserving the species. For successful recruitment, all that wolves require, within the larger occupied landscapes that include several types of land use and cover, are small patches (5–15 km 2 ) of safe habitat for denning and rendezvous sites between December to March ( Jhala, 2003 ). Besides the use of poison, the new multi-lane fast-traffic motorways being built through wolf habitats are a death knell for wolves and other threatened species and need careful mitigation to provide safe passage ( Dennehy et al., 2021 ). Ensuring that breeding habitats are well protected would enable wolves to continue to persist in the larger occupied landscape. This study provides the required information for focused efforts to target and assist in their long-term conservation.

Data Availability Statement

The data analyzed in this study is subject to the following licenses/restrictions: Data on pack size, territory size are included in the Supplementary Material , location data is provided in the figure. Since the precise location of Schedule 1 species under the Wildlife Protection Act is not possible to be provided in the public domain, therefore, wolf location data will be provided for genuine users based on reasonable requests to the corresponding author. Requests to access these datasets should be directed to corresponding author.

Ethics Statement

Ethical review and approval was not required for the animal study because the manuscript does not involve capture or handling of any animal and depends on secondary data that was generated with appropriate legal approvals as per the wildlife protection act.

Author Contributions

YJ conceived the study, collected field data, did the data analysis, and wrote the manuscript. SS conducted data analysis and wrote the manuscript. QQ and SK contributed field data. All authors reviewed and commented on the manuscript.

Extensive camera trap survey was funded by the National Tiger Conservation Authority and State Forest Departments as part of the National tiger status estimation exercise 2018.

Conflict of Interest

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

Publisher’s Note

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.

Acknowledgments

We thank Indrajeet Ghorpade for providing wolf location data from northern Karnataka.

Supplementary Material

The Supplementary Material for this article can be found online at: https://www.frontiersin.org/articles/10.3389/fevo.2022.814966/full#supplementary-material

Supplementary Figure 1 | Relationships of wolf occurrence with eco-geographical variables when all variables were considered together in the model and with variables that were considered in explaining wolf occurrence but not used in the final model.

Supplementary Figure 2 | Three-dimensional and two-dimensional scatter plots of wolf relative abundance index (RAI) against tiger (B,C) , leopard (A,B,D) , and dhole (A,E) . Two-dimensional scatter plots show intensity and 95% ellipses of data distribution. Wolf RAI was negatively correlated with all three large carnivores but was statistically significant ( p < 0.01) only for tigers.

Agarwala, M., Kumar, S., and Treves, A. Naughton-Treves, L. (2010). Paying for wolves in Solapur, India and Wisconsin, USA: comparing compensation rules and practice to understand the goals and politics of wolf conservation. Biol. Conserv. 143, 2945–2955. doi: 10.1016/j.biocon.2010.05.003

CrossRef Full Text | Google Scholar

Blake, S., and Hedges, S. (2004). Sinking the flagship: the case of forest elephants in Asia and Africa. Conserv. Biol. 18, 1191–1202. doi: 10.1111/j.1523-1739.2004.01860.x

Carbone, C., Christie, S., Conforti, K., Coulson, T., Franklin, N., Ginsberg, J. R., et al. (2001). The use of photographic rates to estimate densities of tigers and other cryptic mammals. Anim. Conserv. 4, 75–79. doi: 10.1017/S1367943001001081

Choudhary, S. R. (1970). Let us count our tiger. Cheetal 14, 41–51.

Google Scholar

Daniel, J. C. (1996). The leopard in India – A Natural History. Dehradun: Natraj Publishers, 228.

Dennehy, E., Llaneza, L., and López-Bao, J. V. (2021). Contrasting wolf responses to different paved roads and traffic volume levels. Biodivers. Conserv. 30, 3133–3150. doi: 10.1007/s10531-021-02239-y

Elith, J., Phillips, S. J., Hastie, T., Dudík, M., Chee, Y. E., and Yates, C. J. (2011). A statistical explanation of MaxEnt for ecologists. Divers. Distrib. 17, 43–57. doi: 10.1111/j.1472-4642.2010.00725.x

Garshelis, D. L. (1992). “). Mark-recapture density estimation for animals with large home ranges,” in Wildlife 2001: Populations , eds D. R. McCullough and R. H. Barrett (Dordrecht: Springer), 1098–1111. doi: 10.1890/1051-0761(2006)016[0584:donfsi]2.0.co;2

PubMed Abstract | CrossRef Full Text | Google Scholar

Ghai, R. (2017). Wolf Snapped in the Sundarbans for the First Time. Wildlife & Biodiversity. New Delhi: DownToEarth.

Ghaskadbi, P., Das, J., Mahadev, V., and Habib, B. (2021). First record of mixed species association between dholes and a wolf from Satpura Tiger Reserve, India. Canid Biol. Conserv. 23, 15–17.

Graf, L., and Hatlauf, J. (2021). Distance estimation of howling golden jackals ( Canis aureus ) using relative sound level. Mamm. Res. 66, 567–572. doi: 10.1007/s13364-021-00587-2

Gubbi, S., Ramesh, S., Menon, A. M., Girish, M. N., and Poomesha, H. C. (2020). TheLone wolf: new distribution update of the Indian wolf ( Canis lupus pallipes ) in southern India. Canid Biol. Conserv. 22, 21–24.

Habib, B. (2007). Ecology of the Indian wolf (Canis lupus pallipes Sykes 1831) and modelling its potential in the Great Indian Bustard Sanctuary, Maharashtra, India. Ph.D. Thesis. Aligarh: Aligarh Muslim University, 274.

Habib, B., Ghaskadbi, P., Khan, S., Hussain, Z., and Nigam, P. (2021). Not a cakewalk: insights into movement of large carnivores in human-dominated landscapes in India. Ecol. Evol. 11, 1653–1666. doi: 10.1002/ece3.7156

Harrington, F. H., and Mech, L. D. (1982). An analysis of howling response parameters useful for wolf pack censusing. J. Wildl. Manage. 46, 686–693. doi: 10.2307/3808560

Hennelly, L. M., Habib, B., Modi, S., Rueness, E. K., Gaubert, P., and Sacks, B. N. (2021). Ancient divergence of Indian and Tibetan wolves revealed by recombination-aware phylogenomics. Mol. Ecol. 30, 6687–6700. doi: 10.1111/mec.16127

Jethva, B. (2003). Feeding Ecology and Habitat Needs of Wolves (Canis lupus pallipes) in the Bhal area of Gujarat. Ph.D. Thesis. Dehradun: FRI Deemed University.

Jethva, B. D., and Jhala, Y. V. (2004). Foraging ecology, economics and conservation of Indian wolves in the Bhal region of Gujarat, Western India. Biol. Conserv. 116, 351–357. doi: 10.1016/s0006-3207(03)00218-0

Jhala, Y. V. (1993). Predation on blackbuck by wolves in Velavadar National Park, Gujarat, India. Conserv. Biol. 7, 874–881.

Jhala, Y. V. (2003). Status, ecology and conservation of the Indian Wolf ( Canis lupus pallipes Sykes). J. Bombay Nat. Hist. Soc. 100, 293–307.

Jhala, Y. V. (2007). Conservation of the Indian Wolf. Final Report to the United States Fish and Wildlife Service. Dehradun: Wildlife Institute of India, 531.

Jhala, Y. V., and Giles, R. H. Jr. (1991). The status and conservation of the wolf in Gujarat and Rajasthan, India. Conserv. Biol. 5, 476–483. doi: 10.1111/j.1523-1739.1991.tb00354.x

Jhala, Y. V., and Sharma, D. K. (1997). Child-lifting by wolves in eastern Uttar Pradesh, India. J. Wildl. Res. 2, 94–101.

Jhala, Y. V., Qureshi, Q., and Nayak, A. (eds) (2020). The Status of Tigers, Co-Predators and Prey in India 2018. Dehradun: Wildlife Institute of India.

Jhala, Y. V., Qureshi, Q., and Yadav, S. P. (2021). Status of Leopards, Co-Predators, and Megaherbivores in India 2018. Dehradun: Wildlife Institute of India.

Jhala, Y., Gopal, R., Mathur, V., Ghosh, P., Negi, H. S., Narain, S. S., et al. (2021). Recovery of tigers in India: critical introspection and potential lessons. People Nat. 3, 281–293. doi: 10.1002/pan3.10177

Jiménez-Valverde, A. (2011). Insights into the area under the receiver operating characteristic curve (AUC) as a discrimination measure in species distribution modelling. Glob. Ecol. Biogeogr. 21, 498–507. doi: 10.1111/j.1466-8238.2011.00683.x

Johnsingh, A. J. T., and Acharya, B. (2013). “Asiatic wild dog ( Cuon alpinus ),” in Mammals of South Asia, Vol. I , eds A. J. T. Johnsingh and N. Manjrekar (Hyderabad: Universities Press (India) Private Ltd), 392–415.

Karanth, K. K., Nichols, J. D., Hines, J. E., Karanth, K. U., and Christensen, N. L. (2009). Patterns and determinants of mammal species occurrence in India. J. Appl. Ecol. 46, 1189–1200. doi: 10.1111/j.1365-2664.2009.01710.x

Kumar, S. (1998). Ecology and Behaviour of Indian Grey Wolf (Canis lupus pallipes Sykes, 1831) in the Deccan Grasslands of Solapur, Maharashtra, India . Ph.D. Thesis. Aligarh: Aligarh Muslim University, 215.

Kumar, S., and Rahmani, A. R. (1997). Status of Indian Grey Wolf Canis lupus pallipes and Its conservation in marginal agricultural areas of Solapur district, Maharashtra. J. Bombay Nat. Hist. Soc. 94, 466–472.

Kumar, S., and Rahmani, A. R. (2000). Livestock depredation by wolves in the Great Indian Bustard Sanctuary, Nannaj (Maharashtra), India. J. Bombay Nat. Hist. Soc. 97, 340–348.

Kumar, S., and Rahmani, A. R. (2008). Predation by wolves ( Canis lupus pallipes ) on blackbuck ( Antilope cervicapra ) in the Great Indian Bustard Sanctuary, Nannaj, Maharashtra, India. Int. J. Ecol. Environ. Sci. 34, 99–112.

Kunkel, K., Mack, C. M., and Melquist, W. E. (2005). An Assessment of Current Methods for Surveying and Monitoring Wolves. Idaho: Nez Perce Tribe.

Lokhande, A. S., and Bajaru, S. B. (2013). First record of melanistic Indian wolf Canis lupus pallipes from the Indian subcontinent. J. Bombay Nat. Hist. Soc. 110, 220–221.

Mahajan, P., and Khandal, D. (2021). Assessing the ecological status of Indian Grey Wolf ( Canis lupus pallipes ) with the focus on anthropological interactions in Kailadevi Wildlife Sanctuary. Canid Biol. Conserv. 23, 8–14.

Maron, M., Simmonds, J. S., and Watson, J. E. M. (2018). Bold nature retention targets are essential for the global environment agenda. Nat. Ecol. Evol. 2, 1194–1195. doi: 10.1038/s41559-018-0595-2

Maurya, K. K., Shafi, S., Mall, A., Ojha, G., and Roy, H. K. (2021). First photographic record of Indian wolf Canis lupus pallipes in Valmiki Tiger Reserve, Bihar, India. Canid Biol. Conserv. 23, 1–4.

Mech, L. D. (1970). The Wolf: Ecology and Behavior of an Endangered Species . Minneapolis, MN: University of Minnesota Press, 384.

Mech, L. D., and Boitani, L. (eds) (2007). Wolves: Behavior, Ecology, and Conservation. Chicago: University of Chicago Press.

Muntasir, A., Khan, H., Chowdhury, U., Islam, R., Nehleen, R., Mila, F., et al. (2021). On the reappearance of the Indian grey wolf in Bangladesh after 70 years: what do we know? Mamm. Biol. 101, 163–171. doi: 10.1007/s42991-020-00064-4

Packard, J. M., and Mech, L. D. (1980). Population Regulation in Wolves. Jamestown: USGS Northern Prairie Wildlife Research Center, 372.

Phillips, S. J., Anderson, R. P., and Schapire, R. E. (2006). Maximum entropy modeling of species geographic distributions. Ecol. Model. 190, 231–259. doi: 10.1016/j.ecolmodel.2005.03.026

Phillips, S. J., Anderson, R. P., Dudík, M., Schapire, R. E., and Blair, M. E. (2017). Opening the black box: an open-source release of Maxent. Ecography 40, 887–893. doi: 10.1111/ecog.03049

Rajpurohit, K. S. (1999). Child lifting: wolves in Hazaribagh, India. Ambio 28, 162–166.

Sadhukhan, S., Root-Gutteridge, H., and Habib, B. (2021). Identifying unknown Indian wolves by their distinctive howls: its potential as a non-invasive survey method. Sci. Rep. 11:7309. doi: 10.1038/s41598-021-86718-w

Saren, P. C., Basu, D., and Mukherjee, T. (2019). Status Survey of Indian Grey Wolf ( Canis lupus pallipes ) in West Bengal and some part of Jharkhand. Rec. Zool. Surv. India 119, 103–110.

Shahi, S. P. (1982). Report of grey wolf ( Canis lupus pallipes Sykes) in India-a preliminary survey. J. Bombay Nat. Hist. Soc. 79, 493–502.

Sharma, D. K., Maldonado, J. E., Jhala, Y. V., and Fleischer, R. C. (2004). Ancient wolf lineages in India. Proc. R. Soc. Lond. B Biol. Sci. 271, S1–S4. doi: 10.1098/rsbl.2003.0071

Sharma, L. K., Mukherjee, T., Saren, P. C., and Chandra, K. (2004). Identifying suitable habitat and corridors for Indian Grey Wolf ( Canis lupus pallipes ) in Chotta Nagpur Plateau and Lower Gangetic Planes: a species with differential management needs. PLoS One 14:e0215019. doi: 10.1371/journal.pone.0215019

Sharma, S. (2021). Photo Captures Indian Grey Wolf in Northern Himalayas. New Delhi: Newsclick. Available online at: https://www.newsclick.in/Photo-captures-Indian-grey-wolf-northern-Himalayas

Simmonds, J. S., and Watson, J. E. M. (2019). All threatened species habitat is important. Anim. Conserv. 22, 324–325. doi: 10.1111/acv.12518

Singh, M., and Kumara, H. N. (2006). Distribution, status and conservation of Indian gray wolf ( Canis lupus pallipes ) in Karnataka, India. J. Zool. 270, 164–169. doi: 10.1111/j.1469-7998.2006.00103.x

Sousa-Silva, R., Alves, P., Honrado, J., and Lomba, A. (2014). Improving the assessment and reporting on rare and endangered species through species distribution models. Glob. Ecol. Conserv. 2, 226–237. doi: 10.1016/j.gecco.2014.09.011

Srivathsa, A., Majgaonkar, I., Sharma, S., Singh, P., Punjabi, G. A., Chawla, M. M., et al. (2020). Opportunities for prioritizing and expanding conservation enterprise in India using a guild of carnivores as flagships. Environ. Res. Lett. 15:064009. doi: 10.1088/1748-9326/ab7e50

Srivathsa, A., Puri, M., Karanth, K. K., Patel, I., and Kumar, N. S. (2019). Examining human–carnivore interactions using a socio-ecological framework: sympatric wild canids in India as a case study. R. Soc. Open Sci. 6:182008. doi: 10.1098/rsos.182008

Treves, A., Wallace, R. B., Naughton-Treves, L., and Morales, A. (2006). Co-managing human–wildlife conflicts: a review. Hum. Dimens. Wildl. 11, 383–396. doi: 10.1080/10871200600984265

Trivedi, K., Parmar, P., and Solanki, C. (2021). Recent records of Indian wolves from Bharuch and Surat districts, Gujarat, India. Canid Biol. Conserv. 23, 5–7.

Vanak, A. T., and Gompper, M. E. (2009). Dogs Canis familiaris as carnivores: their role and function in intraguild competition. Mamm. Rev. 39, 265–283. doi: 10.1111/j.1365-2907.2009.00148.x

Wildlife Protection Act (1972). Wildlife Protection Act Number 53 of 1972. 9th September 1972 . New Delhi: Ministry of Law Justice and Company.

Woodroffe, R., Frank, L. G., Lindsey, P. A., Ole Ranah, S. M. K., and Romañach, S. S. (2006). Tools for conserving large carnivores in Africa’s community rangelands: a case-control study of livestock husbandry. Biodivers. Conserv. 16, 1245–1260. doi: 10.1007/s10531-006-9124-8

Wynter-Blyth, M. A., and Dharmakumarsinhji, R. S. (1949). The Gir Forest and its lions. J. Bombay Nat. Hist. Soc. 48, 493–513.

Keywords : Canis lupus pallipes , camera traps, radio telemetry, MaxEnt, home range, pack size, population estimate, wolf-large carnivore Interaction

Citation: Jhala Y, Saini S, Kumar S and Qureshi Q (2022) Distribution, Status, and Conservation of the Indian Peninsular Wolf. Front. Ecol. Evol. 10:814966. doi: 10.3389/fevo.2022.814966

Received: 14 November 2021; Accepted: 10 January 2022; Published: 04 March 2022.

Reviewed by:

Copyright © 2022 Jhala, Saini, Kumar and Qureshi. This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY) . The use, distribution or reproduction in other forums is permitted, provided the original author(s) and the copyright owner(s) are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.

*Correspondence: Yadvendradev Jhala, [email protected]

Disclaimer: All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article or claim that may be made by its manufacturer is not guaranteed or endorsed by the publisher.

The challenge and the opportunity to recover wolf populations

This article is reprinted with permission from Conservation Biology; 1995. 9(2): 1-9.

By L. David Mech, Biological Resources Division, U.S. Geological Survey

Introduction The gray wolf ( Canis lupus ) was one of the first highly visible animals to be included on the U. S. Endangered Species list. The creature now symbolizes endangered species and has become the cause celebre of numerous animal-interest groups. Probably because of the affinity of the wolf to the dog ( Canis lupus familiaris ) and certainly because the species has historically been so persecuted (Young and Goldman 1944), a new mythology about the wolf has evolved; the vile wolf has been replaced by the unjustly persecuted wolf.

As this wolf deification took place, remnant wolf populations were relegated to only the most pristine wilderness of North America and the least developed parts of the rest of the world. Thus, both laypeople and resource managers widely believed that wolves preferred wilderness. The animal came to symbolize wilderness, “for wolves and wilderness are inseparable . . . (Theberge 1975:152).

However, the wolf survived only in wildernesses mostly because it was exterminated everywhere else. After the U. S. Endangered Species Act of 1973 protected the wolf in the 48 contiguous United States as of 1974 and public attitudes about wolves improved, wolves began to colonize a wide variety of habitats and to demonstrate that they did not require wilderness. The wolf has now begun to recover in the northern U. S. and in several parts of Europe. The question of the next decade will not be how to save the wolf, but rather how best to manage the animal. This paper traces the history of the wolf’s status and recovery and explores the dilemma of its management.

History and Persecution Originally, gray wolves were distributed throughout the northern hemisphere in every habitat where large ungulates were found. Saturating most of the region between 20° N latitude (mid-Mexico and India) and the North Pole, in temperatures from -40° to +40° C, the wolf inhabited areas as diverse as Israel and Greenland.

Every kind of northern ungulate, as well as beavers ( Castor canadensis ) and arctic hares ( Lepus arcticus ), can serve as prey for wolves, and wolves easily switch their prey from wild to domestic species. Conflict between wolves and humans over domestic animals probably became an issue soon after ungulates were domesticated.

As firearms, poisons, and traps were developed, they were used ruthlessly against wolves with devastating effectiveness (Young &amp; Goldman 1944). In Eurasia, most wolf populations reached their lowest point between the 1930s and the 1960s (Pimlott 1973; Delibes 1990; Promberger & Bibikov 1993). In the more-developed regions of Eurasia, wolves disappeared except in the central Appenine Mountains of Italy, the Cantabrian mountains of northern Spain, the Carpathians of Eastern Europe, the northern parts of the former Soviet Union, and the central plains and mountainous regions of Asia. Some populations also remained in the deserts of the Middle East. In North America, wolf numbers were lowest in the late 1950s. Populations survived primarily in Canada and Alaska (Mech 1970). In the 48 contiguous United States only the wilderness of northern Minnesota and nearby Isle Royale National Park in Lake Superior held wolves.

The Environmental Revolution The environmental revolution ushered in the first endangered species legislation in the U.S, the Endangered Species Act of 1966. This act did not protect endangered species but only encouraged federal agencies to give them special consideration and to promote their recovery.

At this time, about the only information available on wolves was anecdotal and hearsay. Historical notes by Young and Goldman (1944) and Murie’s (1944) field study on Mt. McKinley wolves were practically the only available published information. A few more studies followed. After the considerable publicity produced by Durward Allen’s seminal investigation of the wolves and moose of Isle Royale National Park, published in National Geographic (Allen & Mech 1963) wolf studies proliferated. In 1967, the first wolf symposium was held by the American Society of Zoologists, culminating in the publication of the proceedings in the May 1967 issue of American Zoologist. By then the full force of the environmental movement could be felt. Private wolf organizations sprang up in many areas, and the wolf quickly gained a popular constituency in the U.S. and abroad.

In Italy, Luigi Boitani and Eric Zimen pioneered a study of the wolf in the Abruzzo Mountains east of Rome (Zimen 1981; Boitani 1986). The World Wildlife Fund and the International Union for the Conservation of Nature and Natural Resources (IUCN), now the World Conservation Union, took great interest in the wolf, and the animal was listed in IUCN’s Red Data Book of endangered species. The IUCN Wolf Specialist Group was formed in 1973 (Pimlott 1975).

Meanwhile radio-tracking was developed in the early 1960s (Cochran &amp; Lord 1963), a technique especially valuable to wolf research. Wolves were difficult to study with traditional methods because they were restricted to wilderness areas, highly elusive, and low in population density. Kolenosky and Johnston (1967) first radio-tracked wolves in Ontario. Mech and Frenzel (1971) then combined that technique with aerial tracking and observation, and numerous studies using these techniques followed.

The second U. S. Endangered Species Act passed in 1973 and protected the wolf in the contiguous 48 United States beginning in August 1974. Recovery teams were appointed by the U. S. Fish and Wildlife Service for three wolf subspecies, the eastern timber wolf, the northern Rocky Mountain wolf, and the Mexican wolf, as well as the red wolf (U.S. Fish and Wildlife Service 1975, 1982a, 1982b, 1987). At first many wolves were killed illegally (Mech 1977), but eventually that number dropped (Fuller 1989), and wolf reservoir populations in less accessible areas expanded (Fuller et al. 1992) They first recolonized the more remote areas in their surroundings, reinforcing the view that they were creatures of the wilderness.

Much of the public misinterpreted the wolf’s endangered status in the 48 contiguous states, thinking it meant no wolves were left anywhere else in the world. Private groups began to raise wolves to help restore populations, not realizing that Canada alone supported 50,000 of them. The wolf’s apparent dependence on the wilderness was quantified in the 1970s and 1980s using road density as a measure. Roads were the routes by which the public and the government had been able to reach wolves to kill them. Thiel (1985) found that recolonizing wolves in Wisconsin lived only where the road density was 0.6 km per km2, a figure corroborated for Michigan (Jensen et al. 1986) and Minnesota (Mech et al. 1988). The wolf then officially became a wilderness animal, and road densities became the yardstick by which wolf habitat suitability was measured by agencies and recovery teams.

Wolf Recovery As more was learned about the wolf, the increasingly urbanized public continued to favor wolf recovery. Even though illegal taking of wolves persists in local areas of North America and Europe, it has not been sufficient to prevent wolf population growth. In Minnesota, some 75 percent of the public viewed the wolf favorably (Kellert 1986), a statistic that may be mirrored in much of the northern hemisphere.

Minnesota’s wolf population, now probably about 2000 based on trend estimates by Fuller et al. (1992), proliferated into neighboring Wisconsin and Michigan (Thiel 1978; Mech et al. 1995b), where they currently number over 100 (Mech et al. 1995a). Other Minnesota wolves eventually spread into the Dakotas (Licht & Fritts 1994). Canadian wolves were no longer killed when they reached Montana, and they began to recolonize the Glacier Park National Park area (Ream & Mattson 1982). One pair even raised pups amongst a herd of cattle on the prairies of the Rockies’ eastern front (Diamond 1994). Montana now supports an estimated 70 wolves, and additional animals from Canada are entering Idaho and Washington state (Mech et al. 1995a).

Europe has seen the same trend. In Italy the wolf population responded to the protection resulting from the research and educational efforts of Boitani (1986) and increased to 300 individuals that inhabit even areas around the outskirts of Rome. In Spain wolf numbers reached 1500-2000 (Blanco et al. 1990), and in Poland, about 850 (Bobek et al. 1993). Overflow to develop in Finland (Pulliainen 1993) and eventually a nascent population developed that straddles Norway and Sweden, currently numbering 20-25 (Promberger et al. 1993a). Wolves also are spreading from northern Italy into France and from Poland into eastern Germany (Promberger et al. 1993b).

The much improved public attitude toward wolves, coupled with publicity and law enforcement have allowed the burgeoning wolf populations to use areas that had not been wolf habitat for decades, thus demonstrating the wolf’s inherent adaptability. The wolf’s new range includes areas of higher road density (Fuller et al. 1992) and much more open, accessible, and populated areas. Breeding packs now live less than 90 km from Minneapolis and St. Paul, Minnesota. One wolf was radio-tracked out of the forests in which it had been raised and into farm fields within 30 km of St. Paul’s center (Wydeven 1994). The animal roamed the farmlands for several weeks before returning to forest. Other wolves making their way south of Minneapolis and St. Paul are being killed by cars or shot when mistaken for coyotes ( Canis latrans ). Wolves dispersing into North and South Dakota have been crossing great expanses of open areas (Licht & Fritts 1994).

In Spain wolves live like coyotes in wheat and sunflower fields in regions with human densities of up to 200 people per km2 (Vila et al. 1993). The animals scavenge garbage and livestock remains and hunt smaller mammals. In Canada, Alaska, Scandinavia, the Mideast, and much of Asia, wolf numbers are stable or increasing (Ginsberg &amp; Macdonald 1990).

Given protection, wolves can expand their range rapidly (Fuller et al. 1992). Average litter sizes reach five to six (Mech 1970). The territorial packs produce young each year, and maturing individuals disperse (Fritts and Mech 1981, Gese and Mech 1991) distances that may exceed 800 km straightline (Fritts 1983). They search out mates and begin new packs (Rothman and Mech 1979) in new areas (Ream et al. 1991).

As wolves dispersed from wildernesses, they successfully contended with more highways, traffic, residences, habitat fragmentation, and other human disturbances (Mech et al. unpublished data). Some probably were unable to adapt, especially the first waves. Nevertheless, wolves that did settle semi-wilderness areas probably became more habituated to the increased disturbances, and as a population then adapted more to increasing disturbance.

In Italy, Spain, and Portugal, where much of the wolf’s food is comprised of garbage, wolves have long inhabited the wooded mountains during the day and made their way into rural villages to scavenge at night (Zimen & Boitani 1979). In North America, ungulate population densities are high close to population centers. Thus, wolves have plentiful natural prey when they move to new, nonwilderness areas.

As wolves show up in new regions they gather new constituencies that support their recovery. In Europe the European Wolf Network dedicated to the recovery of the wolf in central Europe (Promberger & Schroder 1993) became a branch of the IUCN Wolf Specialist Group in 1992. Other organizations have formed in North America that call for the reintroduction of wolves into such places as Arizona, Colorado, northern New York, and New England.

Problems of Wolf Recovery As wolves move into agricultural areas, conflicts with humans greatly increase. For example, when Minnesota wolves increased from 1988 through 1993 by an estimated 15 percent, the number of wolves killed by the U.S. Department of Agriculture Animal Damage Control Program increased from 59 to 139, or 223 percent (Paul 1994). In Spain estimated damage by wolves now exceeds $1 million per year (Vila et al. 1993).

With these conflicts come a distinct danger of public backlash. Not only will wolves in semi-agricultural areas take increasing numbers of livestock and incur the wrath of the livestock industry, which often has strong political support, but they also will kill pets. In Minnesota wolves killing dogs has caused considerable public animosity (Fritts &amp; Paul 1989). As the media begin publicizing such issues, the public gains an exaggerated impression of the problem. A strong backlash of antiwolf sentiment could result in management practices that would again restrict wolves to wilderness areas. Poland has experienced three such cycles of wolf protection and persecution (Okarma 1992). How can these problems be avoided and the wolf be restored to as many places as possible? Until some nonlethal method of controlling wolf populations is discovered, it appears that lethal control will remain the ultimate means of curbing wolf damage to livestock and pets.

Several non-lethal methods of preventing livestock losses to wolves have been tried and abandoned. In Italy and other European countries, for example, traditional husbandry techniques relied on guard dogs and shepherds tending small flocks of livestock; such techniques today are uneconomical. Use of guard dogs alone has been tried against wolves in Minnesota with only limited success (Fritts et al. 1992), although the method may be useful in specific cases. Wolves have also been translocated to other areas, but many either returned to where they were caught or became problems elsewhere (Fritts et al. 1984, 1985). Aversive conditioning (Gustavson & Nicolaus 1987) has not yet proven effective with wild wolves (Fritts et al. 1992). Currently an electric fence in use in Sweden seems to hold some promise for protecting livestock from wolves, but it has not yet been subject to controlled testing (Eles 1986). Furthermore, such fences tested for coyotes have generally been expensive, high-maintenance, and better suited for smaller areas (Dorrance & Bourne 1980, Nass &amp; Theade 1988).

Compensation for livestock losses is useful for minimizing public animosity toward wolves, especially when wolf populations are low and each wolf is important to the population. In Italy, compensation was important in changing public attitudes toward acceptance of wolves in agricultural areas. But as wolf populations proliferate, compensation payments must also increase, sometimes disproportionately. At some point compensation payments will become politically unpopular as the public learns it is subsidizing wolves via payments to farmers for their wolf-killed livestock. Thus many government agencies are wary of even initiating such payments.

An innovative alternative to public payment for livestock killed by wolves was instituted by the Defenders of Wildlife in the U.S. This private, nonprofit organization established a fund to reimburse ranchers in the western U.S. and even encouraged ranchers to allow wolves to raise pups on their private land via a payment of $5000 per den (Fischer et al. 1994). The public may well begin demanding that animal organizations assume these burdens from the government as the costs increase. In any case, without wolf population control, people would eventually object to the payments or the damages caused by wolves.

Wolf Management Zoning With natural habitat in so many areas greatly fragmented and wolves adapting to travel through relatively settled and open areas, some disjunct wolf populations are developing where wolves can live without causing livestock damages. For example, about 90 km northwest of Minneapolis and St. Paul, Minnesota, a pack has lived and bred for at least two years on a wildlife management area surrounded by agricultural land without killing local livestock. Similar instances are known in Montana (Diamond 1994) and other parts of Minnesota (Fritts & Mech 1981, Fritts et al. 1992). This suggests that management zoning could allow wolves to inhabit areas where they can feed on natural prey while they are kept out of agricultural areas.

The approach is to designate zones of potential wolf habitat and distinguish them from areas that should be kept wolf-free. Zoning is common in regulating wildlife harvesting and has been applied on a large scale in wolf recovery plans (U. S. Fish and Wildlife Service 1975, 1987). If public attitudes continue to lean toward protectionism, pressure may develop to apply zoning on local levels such that small sanctuaries are maintained and control is applied only outside these areas.

The scale of zoning is important. Wolves could be zoned out of entire states or zoned into only large national parks or nature preserves. Or they could be allowed to inhabit any area they naturally colonize as long as their sole prey is wild species. For example, in a wildlife refuge of only 100 km2 surrounded by farmland including livestock, wolves could be protected in the refuge but destroyed immediately outside it. This is similar to the situation in Riding Mountain National Park, Manitoba, which, although a much larger area, is an island of wilderness in a sea of agricultural land (Carbyn 1982).

The main advantage of large-scale zoning is simplification and efficiency of management. Any wolf in a designated no-wolf state or outside any large wolf refuge would be subject to legal taking, while those inside would be protected or managed through regulated taking. This scenario could allow wolf populations to remain in the Lake Superior states and much of the mountainous regions of the western U.S., depending on how large the zones are.

The main disadvantage of large-scale zoning is the need to protect livestock that would inevitably live inside some of the larger zones. In Minnesota this would perpetuate the current situation in which close to 150 wolves are killed by government controllers annually for about $1225 each. A second disadvantage is that wolves would probably not be allowed in many areas where they really could live. This might mean banishing wolves two packs have been living without causing livestock depredations. Furthermore, in most of Europe where there are few if any large, remote regions left, large-scale zoning would be very difficult.

With small-scale zoning, the main disadvantage for management agencies is complexity. At one extreme even single wolf packs in areas without livestock would be protected, while immediately outside wolves could be taken. This could present difficult law enforcement problems, although such problems are not unlike those that currently exist for other species in wildlife refuges, national parks and other protected areas. A small-scale zoning proposal in Italy (Boitani & Fabbri 1983) was opposed by wolf protectionists because of the difficulty of law enforcement and the feeling that wolves would be relegated to areas too small to maintain viable populations.

Such a fine-grained approach would probably require management agencies to identify possible wolf areas so that when colonized they would be recognized as wolf sanctuaries. Geographic information systems would greatly simplify this task. Furthermore, identification of such sanctuaries could be incorporated into ecosystem management plans, biodiversity initiatives, and similar strategies as they are developed for other reasons.

The main advantage of small-scale zoning would be to allow wolves to live in enclaves throughout much of Europe and the United States similar to the way they currently inhabit Wisconsin and Michigan (Hammill 1993; Wydeven et al. 1994). For several reasons, this approach would not require the very large-scale land and habitat protection visualized by the Wildlands Project (Mann and Plummer 1993). Although dispersing wolves would be subject to persecution while passing through nonprotected areas, those moving primarily at night or outside of hunting seasons would stand a reasonable chance of survival. With enough small enclaves of wolves, there should be large numbers of such dispersers to colonize new areas, resupply reduced populations, provide sufficient outbreeding, and thus comprise regional metapopulations. Furthermore, inbreeding depression, while a problem among some captive wolves (Laikre & Ryman 1991), probably is not in most wild populations because of the high natural turnover and ensuing selection. Deleterious alleles should get cleansed from the population quickly.

The Isle Royale wolf population is instructive. Isle Royale is a 538-km2 national park in Lake Superior some 25 km from Ontario. It was colonized by wolves about 1949 (Mech 1966), probably by only two unrelated wolves (Rothman & Mech 1979). Genetic testing after 40 years indicated a single female founder (Wayne et al. 1991). Nevertheless, the population stabilized at about 23 for a long period and increased to 50 in 1980, the highest wolf density on record (Peterson & Page 1988). Although the population then crashed, raising concerns about inbreeding depression and disease (Peterson & Krumenaker 1989), the wolves survive. In 1994, eight 1993 offspring survived (Peterson 1994). Thus, with just two founders and 50 percent loss of genetic variability (Wayne et al. 1991), this population has survived for 45 years. Had it been on the mainland, chances are good that some outbreeding would have occurred.

Biologically, wolves could inhabit parts of almost all regions of the U. S. and many of the European countries. Since protection, they have been recorded in nine and possibly ten U. S. states. If biology were the only relevant factor, however, wolves would never have had to be declared endangered. Throughout the wolf’s former range, it has been persecuted because of its tendency to prey on livestock and pets. Even though it is currently on the endangered species list in the U.S., control has been applied in Minnesota, Wisconsin, and Montana. Thus there is every reason to believe that wolf control will parallel wolf recovery wherever it takes place (Mech 1979, Fritts 1993).

The Dilemma of Wolf Management The inevitability of wolf control, however, introduces a new, complex element into the equation governing the wolf’s future in all but the remotest areas of the world: wolf protectionism. The same cultural attitudes that fostered wolf recovery also encouraged an extreme degree of wolf protectionism. Those of us professionally involved with wolf recovery have traditionally been maligned by antiwolf people (Haubner 1990). Now we are vilified by many wolf lovers as wolf enemies because of our acknowledgement that wolves often require control.

Wolves are revered for several reasons. Because they tend to kill prey that are old, sick or weak (Murie 1944; Mech 1970), many laypeople mistakenly believe that, without wolves, prey would automatically die out from disease. Wolves are also hailed as good models for the human race because of their alleged monogamy and family allegiances. A book was even titled The Soul of the Wolf (Fox 1980). Other misconceptions about wolves encourage wolf protectionism. Because of the book Never Cry Wolf by Farley Mowat (1963) and the popular movie made from the book, many people believe wolves live primarily on mice rather than ungulates. Both are fiction (Banfield 1964, Pimlott 1966), but both purport to be true and are sold and shown by museums and other unsuspecting educational organizations. Other misconceptions, half truths, and outdated views that many protectionists hold include the following: wolves only prey on livestock when no natural prey is available; the loss of pack members fost ers disastrous social chaos in the wolf population; wolves socially limit their own population; because the wolf is on the U. S. endangered species list, this means that there are very few left anywhere in the world; and wolves are so shy of humans that they will move out of areas of high activity or avoid settling in them, and they will maintain dens and pups only many kilometers from such activity.

Because of these misconceptions and the power of animal rights groups, wolf control is resisted by much of the public (cf. Garrott et al. 1993). This attitude has three major negative implications for wolf recovery. First, some people revere wolves so much that, rather than having wolves face control, these people would rather not restore wolves to areas where they would have to be controlled. Because wolves will probably have to be controlled almost everywhere they are restored, this sentiment translates into political pressure against wolf recovery. Second, the antiwolf public, such as some livestock owners and organizations, intensify their antiwolf attitudes in reaction to the extremism of the other side. They also fear the possibility of road closures and other restrictions on land use that are often fostered by protectionists using the wolf to prevent logging, mining, snowmobiling or other human uses of semiwilderness and wilderness. Third, some wolf advocates resort to terrorism (Hayes, personal communication) and deceptive advertisements (Anonymous 1992). This zealotry intimidates public officials, who might otherwise be predisposed toward wolf recovery, to shun it.

Of course, the prowolf contingent holds a wide spectrum of attitudes. Thus, some people will accept control against livestock depredations but oppose control prescribed for increasing game herds. Some will accept control by government agencies but not by the public. Many people will accept indirect methods of control such as fencing, guard dogs, or aversive conditioning. These indirect methods are more acceptable because they do not involve humans killing wolves directly. Few proponents of these methods seem to realize, however, that keeping wolves from prey ultimately reduces the carrying capacity of wolf range, and thus fosters starvation and increased deaths from intraspecific strife (Mech 1994). This is particularly true in countries such as Italy, Spain, Israel, where a high percentage of the total carrying capacity for wolves is comprised of livestock, but it applies on a smaller scale to North America as well. As long as wolf deaths are either indirect (and thus not so obvious) or natural, many people accept these deaths who would not tolerate direct or human-caused deaths.

Direct lethal control is still usually the only practical course under most conditions. There are several ways to apply this control. Control by government agency, usually the Department of Agriculture in the U. S., is the type generally most acceptable to wolf advocates, but it is by far the most expensive and time-consuming. Control by landowners or their agents is the one most favored by landowners, but it is difficult to police, and most landowners lack the time and expertise for it, except by poisoning. Open taking of wolves year-around in no-wolf zones similar to the taking of coyotes in most areas of the U. S., and regulated taking by the public, could be applied in no-wolf zones or in wolf sanctuaries to hold the population down such as is done in many suburban areas for white-tailed deer ( Odocoileus virginianus ), geese ( Anser sp. ), and beavers. A modification of this type of control is public taking by special permit.

All of the non-government approaches to control are much less expensive but also less precise to the area or to specific wolves taken and generally are the most disliked by wolf advocates. A notable exception is the government control of wolves to increase herds of big game in areas of Alaska and Canada. A public take of 1200-1500 wolves per year in Alaska brings little or no protest, but the state’s controlling of 150 wolves to increase big game herds is protested vehemently (Anonymous 1993). While biologically this seems illogical, politically such state control allows animal-rights groups to portray this control as a dastardly government program that must be stopped.

The wolf’s high reproductive potential and its tendency to disperse hundreds of kilometers insure that there are few places where wolves could be restored without some form of control being necessary. But the very people most enthusiastically promoting wolf recovery are generally those who want no control, so this dilemma makes public officials reluctant to promote recovery.

Because wolf-taking by landowners or the public is the least expensive and most acceptable by people who do not regard the wolf as special, there will be greater local acceptance for wolf recovery in more areas where such control is allowed. Thus, if wolf advocates could accept effective control, wolves could live in far more places.

The Need for Public Education It appears that the best way to promote wolf recovery is to encourage public education about wolf management issues so that a significant proportion of the public would support wolf recovery while tolerating some form of control. Public education programs must include the message that any restoration of wolves will ultimately result in a need to control wolves (Fritts et al. 1994). Of course, there will always be animal-rights advocates who never will accept any wolf control. If their views are seen by most of the public as counterproductive to wolf recovery, however, officials can probably be persuaded to allow wolves to live in far more of their former range.

Literature Cited

Allen, D. L., and L. D. Mech. 1963. Wolves versus moose on Isle Royale. National Geographic 123:200-219.

Anonymous. 1992. Vaguely worded ads force Alaska to delay wolf kill. Star Tribune, December 22:20A.

Anonymous. 1993. Alaska wolf plan defeated by communication failures. International Wolf 3(2):10-15.

Banfield, A. W. F. 1964. Review of F. Mowat’s Never Cry Wolf. Canadian Field Naturalist 78:52-54.

Blanco, J. C., L. Cuesta, and S. Reig, editors. 1990. El lobo (Canis Lupus) en Espana situacion, problematica y apuntes sobre su ecologia. Ministero de Agricultura Pesca y Alimentacion; Icona; Publicaciones del Instituto Nacional para la Conservacion de la Naturaleza; Madrid.

Bobek, B., M. Kosobucka, K. Perzanowski, K. Plodzien. 1993. Distribution and wolf numbers in Poland. Pages 26-29 in C. Promberger and W. Schroder, editors. Wolves in Europe: Status and perspectives. Proceedings of Wolves in Europe – current status and prospects. Munich Wildlife Society, Ettal, Germany.

Boitani, L. 1986. Dalla parte del lupu. L’Airone di Giorgio Mondadori e Associati Spa, Milano. Italy.

Boitani, L. and M. L. Fabbri. 1983. Strategia nationale di conservationi per il lupo. Ric Biol. Selv. 72:1-31.

Carbyn, L. N. 1982. Incidence of disease and its potential role in the population dynamics of wolves in Riding Mountain National Park Manitoba. Pages 106-116 in F. H. Harrington and P. C. Paquet, editors. Wolves of the world: Perspectives of behavior, ecology and conservation. Noyes Publications, Park Ridge, New Jersey.

Cochran, W. W., and R. D. Lord, Jr. 1963. A radio-tracking system for wild animals. Journal of Wildlife Management 27:9-24.

Delibes, M. 1990. Status and conservation needs of the wolf (Canis lupus) in the Council of Europe member states. Nature and Environment Series 47. Council of Europe. Strasbourg, France.

Diamond, S. 1994. The prairie wolf returns. International Wolf 4(1):3-7.

Dorrance, M. J. and J. Bourne. 1980. An evaluation of anti-coyote electric fencing. Journal of Range Management 33:385-387.

Eles, H. 1986. Vargen. Arsbok fran Varmlands Museum, Argang 84. Andra upplagan, AB Ystads Centraltryckeri 74013, Sweden.

Fischer, H., B. Snape, and W. Hudson. 1994. Building economic incentive into the Endangered Species Act. Endangered Species Technical Bulletin 19(2):4-5.

Fox, M. W. 1980. The soul of the wolf. Little, Brown and Company, Boston.

Fritts, S. H. 1983. Record dispersal by a wolf from Minnesota. Journal of Mammalogy 64:166-167.

Fritts, S. H. 1993. Controlling wolves in the greater Yellowstone area. Pages 173-233 in Robert S. Cook, editor. Ecological issues on reintroducing wolves into Yellowstone National Park. Scientific Monograph NPS/NRYELL/ NRSM-93/22, U.S. Department of the Interior, National Park Service. Washington, D. C.

Fritts, S. H., and L. D. Mech. 1981. Dynamics, movements, and feeding ecology of a newly protected wolf population in northwestern Minnesota. Wildlife Monographs No. 80.

Fritts, S. H., and W. J. Paul. 1989. Interactions of wolves and dogs in Minnesota. Wildlife Society Bulletin 17:121-123.

Fritts, S. H., W. J. Paul, and L. D. Mech. 1984. Movements of translocated wolves in Minnesota. Journal of Wildlife Management 48:709-721.

Fritts, S. H., W. J. Paul, and L. D. Mech. 1985. Can relocated wolves survive? Wildlife Society Bulletin 13:459-463.

Fritts, S. H., W. J. Paul, L. D. Mech, and D. P. Scott. 1992. Trends and management of wolf-livestock conflicts in Minnesota. Resource Publication 181. U.S. Department of the Interior Fish and Wildlife Service, Washington, D. C.

Fritts, S. H., E. E. Bangs, J. A. Fontaine, W. G. Brewster, and J. F. Gore. In press. Restoring wolves to the northern Rock Mountains of the United States. In L. D. Carbyn, S. H. Fritts, and D. R. Seip, editors. Ecology and conservation of wolves in a changing world. Canadian Circumpolar Institute, Edmonton, Alberta.

Fuller, T. K. 1989. Population Dynamics of wolves in north central Minnesota. Wildlife Monograph 105:1-41.

Fuller, T. K., W. E. Berg, G. L. Radde, M. S. Lenarz and G. B. Joselyn. 1992. A history and current estimate of wolf distribution and numbers in Minnesota. Wildlife Society Bulletin 20:42-55.

Garrott, R. A., P. J. White, and C. A. Vanderbilt White. 1993. Overabundance: an issue for conservation biology. Conservation Biology 7:946-949.

Gese, E. M., and L. D. Mech. 1991. Dispersal of wolves (Canis lupus) in northeastern Minnesota, 1969-1989. Canadian Journal of Zoology 69: 2946-2955.

Ginsberg, J. R. and D. W. Macdonald. 1990. Foxes, wolves, jackals, and dogs. World Conservation Union, Gland, Switzerland.

Gustavson, C. R. and L. K. Nicolaus. 1987. Taste aversion conditioning in wolves, coyotes and other canids: Retrospect and prospect. Pages 169-200 in H. Frank, editor. Man and Wolf: Advances, issues, and problems in captive wolf research. Dr. W. Junk Publishers, Dordrect, The Netherlands.

Hammill, J. 1993. Wolves in Michigan: A historical perspective. International Wolf 3(1):22-23.

Haubner, M. H. 1990. Our wildlife crisis – the wolf saga. Nancy Finberg, Carlton, MN 55718.

Jensen, W. F., T. K. Fuller, and W. L. Robinson. 1986. Wolf (Canis lupus) distribution on the Ontario-Michigan border near Sault Ste. Marie. Canadian Field Naturalist 100:363-366.

Kellert, S. R. 1986. The public and the timber wolf in Minnesota. Pages 193-200 in R. E. McCabe, editor. Transactions of North American Wildlife and Natural Resource Conference, vol. 51. Reno, Nevada.

Kolenosky, G. B., and D. H. Johnston. 1967. Radio-tracking timber wolves in Ontario. American Zoologist 7:289-303.

Laikre, L., and N. Ryman. 1991. Inbreeding depression in a captive wolf (Canis lupus) population. Conservation Biology 5:33-40.

Licht, D. S., and S. H. Fritts. 1994. Gray wolf (Canis lupus) occurrences in the Dakotas. American Midland Naturalist 132(1):74-81.

Mann, C. C., and M. L. Plummer. 1993. The high cost of biodiversity. Science 260:1868-1871.

Mech, L. D. 1966. The wolves of Isle Royale. Fauna Series No. 7. U.S. National Park Service, Washington, DC.

Mech, L. D. 1970. The Wolf: The Ecology and Behavior of an Endangered Species. Natural History Press, Doubleday Publishing Co., NY.

Mech, L. D. 1977. Productivity, mortality and population trends of wolves in northeastern Minnesota. Journal of Mammalogy 58:559-574.

Mech, L. D. 1979. Some considerations in re-establishing wolves in the wild. pp. 445-457 in E. Klinghammer, editor. The behavior and ecology of wolves. Garland STPM Press, New York.

Mech, L. D. 1994. Buffer zones of territories of gray wolves as regions of intraspecific strife. Journal of Mammalogy 75:199-202.

Mech, L. D., and L. D. Frenzel, Jr. 1971. Ecological studies of the timber wolf in northeastern Minnesota. U.S. Department of Agriculture Forest Service Research Paper NC-52. North Central Forest Experiment Station St. Paul, Minnesota.

Mech, L. D., S. H. Fritts, G. Radde, and W. J. Paul. 1988. Wolf distribution in Minnesota relative to road density. Wildlife Society Bulletin 16:85-88.

Mech, L. D., D. H. Pletscher, and C. J. Martinka. 1995a. Gray wolf status and trends in the contiguous United States. In E. T. LaRoe, G. S. Farris, C. E. Puckett, and P. D. Doran, editors. Our Living Resources 1994: A report to the nation on the distribution, abundance and health of U. S. plants, animals and ecosystems. National Biological Survey, Washington, D. C.

Mech, L. D., S. H. Fritts, and D. Wagner. 1995b. Minnesota wolf dispersal to Wisconsin and Michigan. American Midland Naturalist 133(2), in press.

Mowat, F. 1963. Never cry wolf. Dell Publishing Company, New York.

Murie, A. 1944. The Wolves of Mt. McKinley. Fauna Series 5. U.S. National Park Service, Washington, DC.

Nass, R. D., and J. Theade. 1988. Electric fences for reducing sheep losses to predators. Journal of Range Management 41:251-252.

Okarma, H. 1992. Wilk in Poland. Polish Academy of Sciences, Mammal Research Institute. 17-236 Bialowieza, Poland.

Paul, W. J. 1994. Wolf depredation on livestock in Minnesota annual update of statistics – 1993. U.S. Animal and Plant Health Inspection Service, Animal Damage Control, Grand Rapids, Minnesota.

Peterson, R.O. 1994. Out of the doldrums for Isle Royale wolves? International Wolf 4(2):19.

Peterson, R. O., and R. E. Page. 1988. The rise and fall of the Isle Royale wolves. Journal of Mammalogy 69:89-99.

Peterson, R. O., and R. J. Krumenaker. 1989. Wolves approach extinction on Isle Royale: A biological and policy conundrum. George Wright Forum 6:10-15.

Pimlott, D. H. 1966. Review of F. Mowat’s Never Cry Wolf. Journal of Wildlife Management 30:236-37.

Pimlott, D. H. 1975. Wolves. D. H. Pimlott, editor. Proceedings of the first working meeting of wolf specialists and of the first international conference on the conservation of the wolf. International Union for Conservation of Nature and Natural Resources, Morges, Switzerland.

Promberger, C., and D. I. Bibikov. 1993. The wolf in the Community of Independent States (former Soviet Union). Pages 23-24 in C. Promberger and W. Schroder, editors. Wolves in Europe: Status and perspectives. Munich Wildlife Society, Ettal, Germany.

Promberger, C., and W. Schroder, editors. 1993. Wolves in Europe: Status and perspectives. Munich Wildlife Society, Ettal, Germany.

Promberger, C., M. Dahlstrom, U. Wotschikowsky, E. Zimen, and P. Wabbakken. 1993a. Wolves in Sweden and Norway. Pages 8-13 in C. Promberger and W. Schroder, editors. Wolves in Europe – Status and perspectives. Munich Wildlife Society, Ettal, Germany.

Promberger, C., C. Vogel, and M. Von Loeper. 1993b. Wolves in Germany. Pages 30-34 in C. Promberger and W. Schroder, editors. Wolves in Europe: Status and perspectives. Munich Wildlife Society, Ettal, Germany. Pulliainen, E. 1993. The wolf in Finland. Pages 14-20 in C. Promberger and W. Schroder, editors. Wolves in Europe: status and perspectives. Wolves in Europe: Status and perspectives. Munich Wildlife Society, Ettal, Germany.

Ream, R. R., and U. I. Mattson. 1982. Wolf status in the Northern Rockies. Pages 362-381 in F. H. Harrington and P. C. Paquet, editors. Wolves of the world: Perspectives of behavior, ecology and conservation. Noyes Publications, Park Ridge, New Jersey.

Ream, R. R., M. W. Fairchild, D. K. Boyd, and D. H. Pletscher. 1991. Population dynamics and home range changes in a colonizing wolf population. Pages 349-366 in R. B. Keiter and M. S. Boyce, editors. The Greater Yellowstone ecosystem, redefining America’s wilderness heritage. Yale University Press, New Haven, Connecticut.

Rothman, R. J., and L. D. Mech. 1979. Scent-marking in lone wolves in and newly formed pairs. Animal Behavior 27:750-760.

Theberge, J. B. 1975. Wolves and wilderness. J. M. Dent and Sons, Toronto, Canada.

Thiel, R. P. 1978. The status of the timber wolf in Wisconsin 1975. Transactions Wisconsin Academy Science, Arts. amd Letters 66:186- 194.

Thiel, R. P. 1985. Relationship between road densities and wolf habitat suitability in Wisconsin. American Midland Naturalist 113: 404-407.

U.S. Fish and Wildlife Service. 1975. Recovery plan for the eastern timber wolf. USFWS, Twin Cities, Minnesota.

U.S. Fish and Wildlife Service. 1982a. Mexican wolf recovery plan. USFWS, Albuquerque, New Mexico.

U.S. Fish and Wildlife Service. 1982b. Red wolf recovery plan. USFWS, Atlanta, Georgia.

U.S. Fish and Wildlife Service. 1987. North Rocky Mountain wolf recovery plan USFWS, Denver, Colorado.

Vila, C., J. Castroviejo, and V. Urios. 1993. The Iberian wolf in Spain. Pages 104-109 in C. Promberger and W. Schroder, editors. Wolves in Europe: Status and perspectives. Munich Wildlife Society, Ettal, Germany.

Wayne, R. K., D. A. Gilbert, N. Lehman, K. Hansen, A. Eisenhawer, D. Girman, L. D. Mech, P. J. P. Gogan, U. S. Seal, and R. J. Krumenaker. 1991. Conservation genetics of the endangered Isle Royale gray wolf.

Conservation Biology 5:41-51.

Wydeven, A. P. 1994. Travels of a Midwestern disperser. International Wolf 4(1):20-22.

Wydeven, A. P., R. N. Schultz, and R. P. Thiel. 1994. Gray wolf monitoring in Wisconsin. L. D. Carolyn, S. H. Fritts, and D. R. Seip, editors. Ecology and conservation of wolves in a changing world. Canadian Circumpolar Institute, Edmonton, Alberta.

Young, S. P., and E. A. Goldman, editors. 1944. The wolves of North America. American Wildlife Institute, Washington, D.C., and Dover Publishers, New York.

Zimen, E. 1981. The Wolf: A species in danger. Delacourt, New York.

Zimen, E., and L. Boitani. 1979. Status of the wolf in Europe and the possibilities of conservation and reintroduction. Pages 43-83 in E. Klinghammer, editor. The behavior and ecology of wolves. Garland STPM Press, New York.

The International Wolf Center advances the survival of wolf populations by teaching about wolves, their relationship to wildlands and the human role in their future.

Click through the PLOS taxonomy to find articles in your field.

For more information about PLOS Subject Areas, click here .

Loading metrics

Open Access

Peer-reviewed

Research Article

Hybrid Tabu-Grey wolf optimizer algorithm for enhancing fresh cold-chain logistics distribution

Roles Conceptualization

Affiliation School of Business, Beijing Technology and Business University, Beijing, China

Roles Writing – review & editing

Roles Visualization, Writing – original draft

* E-mail: [email protected]

• Hao Zhang, 
• Jianing Yan, 
• Liling Wang

• Published: August 29, 2024
• https://doi.org/10.1371/journal.pone.0306166
• Peer Review
• Reader Comments

The increasing public demand for fresh products has catalyzed the requirement for cold chain logistics distribution systems. However, challenges such as temperature control and delivery delays have led a significant product loss and increased costs. To improve the current situation, a novel approach to optimize cold chain logistics distribution for fresh products will be presented in the paper, utilizing a hybrid Tabu-Grey wolf optimizer (TGWO) algorithm. The proposed hybrid approach combines Tabu search (TS) and Grey wolf optimizer (GWO), employing TS for exploration and GWO for exploitation, aiming to minimize distribution costs in total and establish efficient vehicle scheduling schemes considering various constraints. The effectiveness of the TGWO algorithm is demonstrated through experiments and case studies compared to other heuristic algorithms. Comparative analysis against traditional optimization methods, including Particle swarm optimization (PSO), Whale optimization algorithm (WOA), and original GWO, highlights its superior efficiency and solution quality. This study contributes theories by demonstrating the efficacy of hybrid optimization techniques in complex supply chain networks and dynamic market environments. The practical implication lies in the implementation of TGWO to bolster distribution efficiency, cost reduction, and product quality maintenance throughout the logistics process, offering valuable insights for operational and strategic improvements by decision-makers. However, the study has limitations in generalizability and assumptions, suggesting future research areas including exploring new search operators, applying additional parameters, and using the algorithm in diverse real-life scenarios to improve its effectiveness and applicability.

Citation: Zhang H, Yan J, Wang L (2024) Hybrid Tabu-Grey wolf optimizer algorithm for enhancing fresh cold-chain logistics distribution. PLoS ONE 19(8): e0306166. https://doi.org/10.1371/journal.pone.0306166

Editor: Mazyar Ghadiri Nejad, Cyprus International University Faculty of Engineering: Uluslararasi Kibris Universitesi Muhendislik Fakultesi, TURKEY

Received: January 31, 2024; Accepted: June 12, 2024; Published: August 29, 2024

Copyright: © 2024 Zhang et al. This is an open access article distributed under the terms of the Creative Commons Attribution License , which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.

Data Availability: All relevant data are within the manuscript and its Supporting information files.

Funding: This paper was supported by Beijing Natural Science Foundation (9222007).

Competing interests: The authors declare that there are no Competitive financial benefits and no conflict of interest.

1. Introduction

With the improvement of people’s living standards and the change in consumer attitudes, there is a growing demand for fresh products, which is effectively driving consumption and creating new growth opportunities for the fresh e-commerce industry. However, the perishable nature of fresh food means that it needs to be kept fresh for a certain period of time, otherwise, the quality and taste will be impacted, and it even poses a risk to human health. This is where cold chain logistics distribution comes in, which refers to the use of professional refrigeration equipment and technical means in the logistics process to maintain suitable temperature and humidity for fresh food throughout the distribution process. Despite its importance, the cold chain logistics distribution process faces many challenges due to the unique nature of fresh food, such as unstable temperature control, cargo damage and delayed delivery time. These problems not only impact the quality and taste of fresh food, but also increase logistics costs and delivery risks. Shockingly, despite 15% of the world’s total energy consumption already being dedicated to cold chain logistics and cooling systems, the annual loss of fresh products ranges from 5%-15% of the total production due to inadequate infrastructure and inefficient resource allocation [ 1 ]. In developing countries, the rate of loss can even reach up to 50% [ 2 ]. Furthermore, the expensive costs involved in the food cold chain limit the capacity for fresh products e-commerce to make a profit [ 3 ]. Therefore, logistics distribution optimization is necessary for the development of a fresh cold chain.

In the cold chain logistics distribution network for fresh products in cities, the distribution requirements (primarily temperature) vary depending on the type of product being transported. Additionally, each customer has a designated distribution time, which may differ from other customers’ times. Some customers are flexible with their availability and may accept compensation for late deliveries, while others have strict time windows and will reject products delivered outside their allotted slot. This creates a unique challenge in the form of a Vehicle Routing Problem with Time Windows (VRPTW), where each customer has their own soft and hard time window, which was first proposed by Dantzig and Ramser [ 4 ]. Therefore, efficient vehicle schedules play a critical role in optimizing the old chain logistics distribution, with extensive scholars identifying the challenge, trying to improve this situation, and applying various methodologies, including mixed integer programme models and evolutional algorithms [ 5 ]. Followed by Amorim and Almada-Lobo’s [ 6 ] theory of cost-freshness trade-offs, which will be explained in detail in Section 2. Although it can be seen that both soft and hard time windows exist in fresh cold chain logistics distribution. The key challenge that needs to be tackled is how to ensure that each customer is reached within their specified time window to their satisfaction.

Motivated by optimizing cold chain logistics distribution of fresh e-commerce, a mathematical optimization model is applied in this paper, aiming at optimizing cold chain logistics distribution for fresh e-commerce. The objective is to lower supply chain costs and ensure on-time delivery by optimizing food delivery routes while considering constraints, such as vehicle capacity, maximum travel range, minimum acceptable freshness level of food, and coexisting soft and hard time windows. In addition, a hybrid Tabu-Grey Wolf Optimizer (TGWO) algorithm was designed to solve the above problem, in which Tabu Search (TS) was used to increase the local search ability of the Grey Wolf Optimizer (GWO) and improve the quality of the initial solution by further exploring the search space. The results of the experiments proved the validity and potential application of TGWO in providing superior results compared to other common heuristic algorithms.

The remaining sections of this paper are organized as follows. The related literature will be reviewed in Section 2. In Section 3, the research methodology and solution approach will be presented, followed by Section 5, in which an overview of the results obtained will be provided. In the end, discussion and research limitations will be concluded.

2. Literature review

Many scholars have made significant efforts to enhance the optimization of vehicle routing of the fresh cold chain. To design delivery routes, some focused on the Vehicle Routing Problems (VRP), which was first introduced in 1959 by Dantzig and Ramser under the title “The Truck Dispatching Problem” [ 4 ]. The study of the VRP and its extensions have given rise to major developments in the field of exact algorithms and heuristics [ 7 ]. Wang [ 5 ] proposed a mixed integer programming model and developed a branch-and-price algorithm to solve vehicle routing problems with drones. Amorim and Almada-Lobo [ 6 ] examined the relation between distribution scenarios and the cost-freshness trade-off. A∈ constrained method is used for small-size instances for VRPTW, and for large-size instances a multi-objective evolutionary algorithm is implemented. Wang [ 8 ] constructed a green and low-carbon cold-chain logistics distribution route optimization model and Cycle Evolutionary Genetics Algorithm (CEGA) was proposed to contribute to the model. Davila [ 9 ] used tabu search, chaotic search and general algebraic modeling to solve the vehicle routing problem for distributing refrigerated products. The VRPTW commonly employs two types of time window constraints: hard and soft. In the cold chain logistics distribution problem of fresh products, hard time window constraints require fresh products to be delivered during the exact designated time window, forbidding both earliness and tardiness [ 10 ]. Nonetheless, in reality, fresh products remain edible even if the hard constraint is violated, and the distribution network may become over-constrained, resulting in feasible solutions [ 11 , 12 ]. Soft time window constraints handle the earliness and tardiness with penalties, which are often assumed to follow a linear relationship with the amount of variation from the time window [ 13 ]. Therefore, numerous exceptional works have been developed by combining VRPTW with various constraints and objectives to solve more practical instances in cold chain logistics. This paper mainly considers the constraints of vehicle capacity, maximum travel range, minimum acceptable freshness level of food and coexisting soft and hard time windows, aiming to optimize the food delivery routes to minimize total costs and ensure on-time delivery. As for solution method, instead of exact algorithm, more and more metaheuristic algorithms are proven to be suitable in cold chain distribution.

Currently, swarm intelligence-based algorithms are commonly used for optimizing cold chain logistics distribution, like ant colony optimization (ACO) [ 14 – 16 ], particle swarm optimization (PSO) [ 17 – 19 ], whale optimization algorithm (WOA) [ 20 , 21 ], and grey wolf optimizer (GWO) [ 22 , 23 ]. GWO was proposed by Mirjalili [ 24 ] in 2014. It simulates the hunting behaviour and social hierarchy of grey wolves in nature. The existing studies have shown that GWO is competitive with some common optimization algorithms [ 24 , 25 ]. Because of the advantage of fewer control parameters, a highly stochastic nature and derivation-free mechanism, numerous variants of GWO have been proposed to solve complex constrained practical optimization problems [ 26 , 27 ], for example, optimizing the power dispatch problem [ 28 ], the assembly flow shop scheduling problem [ 25 ], the capital goods scheduling problem [ 29 ] and the path planning problem [ 30 ].

However, as search space dimensions grow, GWO tends to perform poorly in exploitation [ 31 ]. Scholars have proposed various approaches to improve the original GWO algorithm. For example, to increase the diversity of the grey wolf population, Zhu [ 32 ] combined the GWO algorithm with the differential evolution (DE) algorithm to improve its performance and global exploration capability. Singh and Hachimi [ 33 ] balanced the developmental and explorational abilities of GWO by introducing the spiral equation of the whale optimization algorithm (WOA). Zhang [ 34 ] combined the GWO algorithm with the biogeographic optimization (BA) algorithm based on antagonistic learning, which can prevent the algorithm from quickly choosing the local optimal solution. Gupta and Deep [ 35 ] modified algorithm RW-GWO based on random walk and its performance is exhibited in comparison with GWO and state of art algorithms GSA, CS, BBO and SOS on IEEE CEC 2014 benchmark problems. The results demonstrated that the algorithm provides a better leadership to search a prey by grey wolves.

All these existing literature lays a theoretical foundation for this paper. GWO can effectively avoid local optimal solutions but its optimization accuracy needs to be improved, while tabu search algorithm has strong global search capability but is prone to local optimal solutions. To address these limitations, this paper proposes a hybrid Tabu-Grey wolf optimizer (TGWO) to enhance the optimization model as an extension of VRPTW. The proposed algorithm combines the exploration capability of the GWO algorithm with the local search capability of the Tabu search algorithm to achieve a better balance between exploration and exploitation. The specific enlightenment can be reflected in the following aspects: Firstly, as for shortcomings of the GWO algorithm in generating the initial solution, the tabu search algorithm with strong global search ability is introduced to expand the types of solutions and enhance the optimization ability of the algorithm. Secondly, although GWO algorithms are rarely used in similar discrete optimization problems, some scholars have improved algorithms to apply in vehicle routing problems and achieved great success. Therefore, this paper analyzed these improvement ideas, combined with the characteristics of the improved grey wolf optimization algorithm, proposed a solution method, and applied it to the VRPTW of fresh e-commerce cold chain distribution. Finally, based on the characteristics of fresh e-commerce cold chain distribution, when constructing the VRP model, the paper considered the coexistence of soft and hard time windows, and established a model with the maximum vehicle capacity and the maximum travel range constraints. The results show that it can provide better results compared to traditional methods and highly improve the efficiency and effectiveness of the cold chain logistics distribution process.

3. Methodology

3.1 model assumption.

To develop the mathematical model, the assumptions are summarized as: 1) The central depot has sufficient commodity supplies, and only one type of capacitated distribution vehicle is considered; 2) Each customer’s demand must be satisfied, and each customer is only visited once by a distribution vehicle; 3) Each vehicle departs from and returns to the same central depot; 4) The distribution vehicles travel at a constant speed, which means that the transportation time is only associated with the accumulated transport mileage; 5) The distribution vehicles are only responsible for delivering commodities and temporarily preserving the rejected goods, if any, until returning to the central depot. They are not responsible for receiving new freight; 6) The cargo damage rate ∂ 2 when unloading is higher than ∂ 1 when in transport.

3.2 Sets, parameters, and variables

The parameters used in the model are shown in Table 1 .

• PPT PowerPoint slide
• PNG larger image
• TIFF original image

https://doi.org/10.1371/journal.pone.0306166.t001

3.3 Cost analysis

The mathematical model aims to minimize the total cost, which encompasses the sum of the fixed cost, transportation cost of distribution vehicles, cargo damage cost, and penalty for violating time window constraints.

Cargo damage cost. The quality of perishable products during cold chain logistics distribution is a crucial factor affecting customer satisfaction and repeat purchase in e-commerce again [ 36 ]. This study assumes that the temperature inside the transportation compartment remains stable during transit, resulting in a constant spoilage rate of the perishable products. However, when unloading the goods, the opening door causes heat exchange with the environment, leading to a higher spoilage rate than during transit. As a result, the total cost of cargo damage comprises two parts: the cost incurred during transportation due to time extension and the cost of damage when unloading the goods at the customer’s location.

3.4 Objective function

Eq (9) is the objective function, which minimizes the total cost Z in the vehicle routing problem. Constraint (10) indicates that distribution vehicles start from the distribution centre and return to the centre after finishing the deliveries to the given customers. Constraint (11) indicates that the demand of all customers in each delivery route should not exceed the maximum vehicle capacity. Constraint (12) indicates that the summation of the distribution distance in each delivery route should not exceed the maximum transport range. Constraint (13) ensures that each customer is only accessible by one delivery vehicle. Constraint (14) ensures the continuity of the delivery services. Constraints (15) and (16) ensure that vehicles start and finish the deliveries to customers within the hard time windows as required. Constraint (17) indicates that the penalty coefficient is a nonnegative number.

3.5 Conversion of TSP to VRP

In order to propose the solution idea, this paper introduces the concept of “Traveling Salesman Problem (TSP)”, a special form of VRP, which given a set of cities and a cost to travel from one city to another, seeks to identify the tour that will allow a salesman to visit each city only once, starting and ending in the same city, at the minimum cost [ 37 ]. Although TSP is a simple version of VRP, it has long been proven to be an NP-Hard problem in combinatorial optimization.

Considering that the TGWO studied in this paper is an algorithm commonly used to solve continuous function optimization problems, and the settings of the upper and lower bounds of variables are limited when applied to discrete optimization problems such as VRP, this paper uses the following solutions: (1) to define each grey wolf as a TSP sequence starting from the distribution centre; (2) to define the total cost of the cold chain logistics as the fitness function. As in Fig 1 , (b) traverses through each TSP sequence in ascending order of their indices, and inserts the starting point (distribution centre) if necessary, which involves assigning the rest of the distribution tasks to a new driver if the previous route violates any constraints (capacity, transport mileage or food freshness). After converting a set of TSP routes into one feasible solution to the proposed model, the fitness value is calculated for further comparison with other candidate solutions; (3) to select define this as the optimal solution in the current schemes by comparing the current solutions with the candidates.

https://doi.org/10.1371/journal.pone.0306166.g001

4. Solution approach

4.1 grey wolf optimizer (gwo).

The inspiration for GWO originates from the social hierarchy and hunting behaviour of gray wolves. There are four types of society members in a gray wolf pack, namely, alpha (α), beta (β), delta (δ) and omega (ω), according to the descending order of their social hierarchy. Each wolf performs its own duties and cooperates with the other wolves. Accordingly, the best solution obtained by GWO is considered the alpha, the second and third solutions are called the beta and delta, respectively, and the remaining solutions are referred to as the omegas, which are also assumed to be inferior solutions. The principal steps are as follows.

• Step 1 : Initialization

It should be noted that in the vehicle routing model of this paper, each TSP route should start at the distribution centre. Thus, the first elements of each gray wolf’s position vector will be maintained as the smallest number of iterations by setting the upper and lower bounds of the first column to the same infinitesimal real number, lb 1 , and the lower bounds of the other dimensions are set to be the same number lb 2 ( lb 2 > lb 1 ). An example of the operation mode is shown in Fig 2 . Therefore, the random numbers in the first dimension would always be the first numbers when the elements of each gray wolf’s position vector are arranged in ascending order. Then, based on the decoding mechanism introduced in the next step, the obtained position vectors can be transformed into several sets of routes, which all start at the distribution centre, pass through all the customer locations and remain at the location of the last customer served.

https://doi.org/10.1371/journal.pone.0306166.g002

• Step 2 : Encoding and Decoding

After the grey wolf population is randomly initialized, the random numbers obtained in each dimension of each wolf are arranged in ascending order. Then, several sequences of integers arranged from 1 to n + 1 are obtained. The original customer code numbers corresponding to these ascending sequences are the candidate solutions of TSP.

• Step 3 : Solution Generating

Each non-holonomic TSP route needs to be transformed into one VRP route based on customer orders and several different constraints. The transformation is as follows.

First, the corresponding demands of the customers are summed up in sequence according to the obtained TSP route. Once the current vehicle load or distribution distance is about to or has already met the constraint of maximum capacity Q or the maximum transport mileage D , adding one more unscheduled customer to this route will result in an overload or running out of fuel on the return to the centre. The algorithm terminates the process of adding new customers and inserts number 1 (code number of the distribution centre) into the TSP route immediately before the code number of the following unscheduled customer. This indicates that the delivery route of the first vehicle has been finished. Then, consider the next delivery route arrangement. Continue the previous steps until all given customers are arranged. Finally, a set of vehicle routing schemes is obtained, in which 1 can appear more than once each time for a new route, and each customer order can only occur once.

• Step 4 : Fitness Function

After each iteration, the gray wolf with the smallest fitness value is defined as the alpha, and the wolves with the second and third smallest fitness values are defined as the beta and delta, respectively. In this way, the first three optimal gray wolves and their position vectors are obtained, as well as the corresponding candidate VRP routes.

• Step 5 : Update the positions of all gray wolves based on those of the first three optimal gray wolves, α, β and δ.

This updating process can be summarized in three sub-steps:

• Step 6: Termination Criterion

Repeat the above iterative search process. Terminate the algorithm when the pre-set maximum iteration is reached, and then, output the optimal solution obtained so far as well as the corresponding delivery vehicle route and fitness value (total delivery costs). With the use of the mentioned algorithms, the solution to the vehicle routing problem can be determined.

The GWO procedure is illustrated in Fig 3 .

https://doi.org/10.1371/journal.pone.0306166.g003

4.2 Tabu-grey wolf optimization algorithm

GWO tends to provide poor convergence performance with increasing search space dimensions. Thus, it is necessary to enhance the local search ability of the GWO algorithm. Therefore, the neighbourhood searching process of TS is introduced into TGWO to strengthen its exploitation ability, and the tabu list limits the solution returned to recently visited solutions in a given move and increases the diversity of the candidate solutions to vehicle routing optimization problems.

The main steps of the TGWO algorithm are as follows.

• Steps 1–4 are the same as those of GWO.
• Step 5: Neighbourhood Search

This is part of the section where TGWO differs from the standard GWO. Once a new feasible solution is generated in the iterative process, the fitness value needs to be calculated and compared with that of the best solution obtained so far. If the fitness value shows superiority, then the optimal solution should be replaced by the current one and a neighbourhood transformation should be performed to determine whether there is a better solution in the neighbourhood. This searching process is performed by running the neighbourhood searching process Iter neighbor times. N neighbor initial neighbourhood transformed solutions would be generated each time. Otherwise, the iterative process should be continued. The neighbourhood transformation further explores the search space and contributes to creating more feasible TSP routes and a better VRP solution. It improves the quality of the current solution.

A multi-neighbourhood structure is designed for TGWO, where three search operators are randomly adopted: insert, swap and 2-opt. This paper discusses the neighbourhood strategies in details by simulating a delivery route with seven customers: 1→3→5→4→6→2→7.

4.2.1 Insert operator.

A customer is randomly selected and placed (inserted) after another random position in the delivery route. The customers after that position move further. The operation procedure is presented in Fig 4 : a selected customer (framed by a dotted line) is inserted into a randomly selected position (marked by a positive triangle). Then, the original route is transformed into 1→3→2→5→4→6→7.

https://doi.org/10.1371/journal.pone.0306166.g004

4.2.2 Swap operator.

Interchange the positions of two randomly chosen customers to change the order through which the delivery vehicles pass in the TSP scheme. The operation procedure is presented in Fig 5 : a selected customer (framed by a dotted line) is randomly selected position, after swap transformation, the delivery order becomes 1→2→5→4→6→3 →7.

https://doi.org/10.1371/journal.pone.0306166.g005

4.2.3 2-opt operator.

This operator is also known as 2-exchange. Randomly select two customers, and invert all the customers between them. The operation procedure is presented in Fig 6 : a selected customer (framed by a dotted line) is randomly selected position, and the solid box marks the position where the inverted change occurs, and the delivery order becomes 1→3→6→4→5→2→7 after the 2-opt transformation.

https://doi.org/10.1371/journal.pone.0306166.g006

In order to increase the exploration ability of the algorithm on the neighborhood, this paper defines insert, swap and 2-opt as mutation mode 1, mutation mode 2 and mutation mode 3, respectively, and more possible TSP schemes are generated by randomly selecting the types of variants to increase their diversity.

• Step 6: Tabu List

One of the three mutation modes is randomly operated during one iteration in the neighbourhood searching process. This mechanism serves the purpose of increasing the diversity of the solutions and improving the robustness of the algorithms. Meanwhile, several recently visited solutions are forbidden for a number of iterations to prevent cycling. This function is realized by placing the solutions in a tabu list.

For the TSP/VRP problem, we set a tabu list of TGWO with a length as length and a width of 3. It has the ability to record the neighbourhood transform strategies and prohibit them from being performed repeatedly in a given period. The three columns of the tabu list are the code numbers for the mutation mode and two randomly chosen customers. The functional mechanism of the tabu list is shown in Fig 7 , and the sample route is 1→3→5→4→6→2→7.

https://doi.org/10.1371/journal.pone.0306166.g007

After generating several candidate solutions in the neighbourhood, their fitness values are stored for further comparison. If the fitness value of one mutated solution is proven to be better than that of the current solution, the TGWO algorithm verifies whether the mutation mode and corresponding customer position already exist in the tabu list. If so, this solution is abandoned and the searching process in the neighbourhood is continued. If not, the current fitter solution better – so – far is replaced by this one and its corresponding mutation mode is added into the tabu list, so it is not being visited again for a number of iterations.

• Step 7: Termination

According to the maximum number of iterations set in advance, when the number of cycles of the entire calculation process of the TGWO is greater than this value, the calculation is terminated, and the optimal scheme of this algorithm and the total distance are output.

5.1 Validity of TGWO

To evaluate the TGWO algorithms proposed in this paper, TGWO is benchmarked on 6 benchmark functions. These benchmark functions are chosen from the 23 classical benchmark functions utilized by many researchers [ 24 , 38 ] to compare our results to those of the current meta-heuristics. The benchmark functions used are minimization functions and divided into two groups, including three unimodal functions ( F 1 − F 3 ) and three multimodal functions ( F 4 − F 6 ). These benchmark functions are listed in Table 2 , where Dim indicates the dimension of the function, Range is the boundary of the function’s search space, and f min is the theoretical optimum. Fig 8 illustrates the 2-D versions of the unimodal and multimodal benchmark functions used. To verify the results, the TGWO algorithm is compared with the PSO, WOA and the original GWO algorithm for verification. Taking F1 as an example, the convergence curves of each algorithm in two different dimensions are shown in Fig 9 . And PSO and WOA are designed according to the paper of [ 39 ] and [ 40 ], except that the fitness function is the same as that of GWO and TGWO. The lower the fitness function value is, the better the result. The parameters of the algorithms are set according to Table 3 . In order to reduce the influence of randomness on the algorithm comparison results, all algorithms are run 200 times on each function for 10 and 30 dimensions and are tested in MATLAB R2020a. The operating system is Windows 10.

https://doi.org/10.1371/journal.pone.0306166.g008

https://doi.org/10.1371/journal.pone.0306166.g009

https://doi.org/10.1371/journal.pone.0306166.t002

https://doi.org/10.1371/journal.pone.0306166.t003

5.1.1 Exploitation analysis.

According to the Fig 9 , in the process of grey wolf search for prey, the neighborhood search and tabu list functions added by the improved grey wolf algorithm greatly enhance the ability of the original grey wolf population to explore and optimize the local solution, and increase the diversity of the obtained solution in the calculation process, and expand the search range in the search space, so the optimization effect of the algorithm has been greatly improved, and the TGWO algorithm provides superior results compared to those of PSO, WOA and GWO. It is especially evident in the optimization test of the unimodal functions ( F 1 , F 2 and F 3 ). The results show the superior performance in terms of exploiting the optimum.

5.1.2 Exploration analysis.

Differ from unimodal functions, multimodal functions have multiple local optima, and the number of these optima increases exponentially with the dimensionality of the function. This characteristic makes them ideal for evaluating an algorithm’s exploration capabilities as a benchmark. According to the Fig 9 , TGWO can provide highly competitive results on the multimodal benchmark functions. With the increase of problem dimensions, the GWO algorithm provides relatively poor convergence behaviour for exploitation. The TGWO algorithm, conversely, shows its superior performance to that of GWO. It even outperforms PSO on the results of experiments on multimodal functions F 5 and F 6 . All these results show that the TGWO algorithm has merit in terms of exploration ability in search space. This also shows that the combination of the GWO algorithm and the TS algorithm is reasonable and effective, which can greatly improve the reliability of the algorithm.

5.2 Case study

Since fresh product e-commerce is an online trading platform, the customer’s location, time window constraints and demands are constantly changing when planning the vehicle delivery route each time, so this paper selects one of the distribution centres of Company J, a fresh e-commerce platform in City B, as the research object. Besides, the distribution centre currently has 10 delivery vehicles waiting to be called, assuming that the speed of each delivery vehicle is 1, the maximum capacity is 20, and the maximum driving distance is 200, in order to complete the distribution task, each vehicle needs to start from the distribution centre 1 and finally return to it. The fixed cost of a delivery vehicle C 0 is 150 and the unit transportation cost is 10. (Other parameters setting: M = 100000000). In this section, the GWO, TS and TGWO are used to solve the above VRPTW problem. Among them, the solution software is MATLAB R2020a, the number of iterations is 800, the number of wolves is 80, the lower and upper bounds are set to 0.01 and 10 respectively (the upper and lower bounds of the first dimension are 0.001), the tabu list length is 10, the neighborhood search number is 80, and the tabu list length in TGWO is 30, the neighborhood search is 50 times, and the neighborhood iteration is 100 times. Finally, the optimal costs calculated by the three algorithms are 17314.8, 7754.9 and 7112.5, respectively.

The iteration processes of three algorithm are shown in Fig 10 .

https://doi.org/10.1371/journal.pone.0306166.g010

The VRP route obtained by the three algorithms is as follows.

TS . As shown in first figure of Fig 11 , a total of 4 delivery vehicles are required to provide services to the customers in this example. Among them, the first car has a route of 1→12→14→4→16→1; the second 1→15→5→7→11→6→2→1; the third 1→10→3→9→8→13→17→1, and the fourth vehicle has a route of 1→18→1.

https://doi.org/10.1371/journal.pone.0306166.g011

GWO . As shown in second figure of Fig 11 , a total of 4 delivery vehicles are required to provide services to the customers in this example. Among them, the first car has a route of 1→10→8→17→7→9→1; the second 1→4→3→6→14→2→1; the third 1→13→11→12→5→18→1, and the fourth vehicle has a route of 1→15→16→1.

TGWO . As shown in last figure of Fig 11 , a total of 4 delivery vehicles are required to provide services to the customers in this example. Among them, the first car has a route of 1→8→15→18→9→2→16→1; the second 1→10→11→13→7→17→1; the third 1→4→3→14→12→5→1, and the fourth vehicle has a route of 1→6→1.

The TGWO algorithm enhances the variety of available solutions by utilizing neighborhood search and tabu list functions in the solution process. This results in a more efficient and improved VRP scheme compared to the other two algorithms, with a reduced total driving distance and cost. Specifically, the TGWO algorithm saved 50.34% and 30.66% in terms of total travel distance compared to TS and GWO algorithms, respectively. In addition, it saved 143.44% and 9.03% in terms of total distribution cost, respectively. Furthermore, the VRP scheme obtained by the TGWO algorithm is clearer and more concise than the schemes obtained by TS and GWO. It arranges only one delivery person to provide services to customers in one direction as much as possible, resulting in reduced total driving distance and higher rationality. In conclusion, this paper presents a superior set of vehicle route planning schemes for the cold chain distribution process of fresh e-commerce. Comparison with the schemes obtained by different algorithms reveals that the TGWO algorithm provides a more rational, cost-effective and faster solution to the problem of cold chain distribution for fresh products. Thus, the improved grey wolf optimizer algorithm TGWO demonstrates its efficacy in solving discrete problems.

6. Discussion

The cold chain logistics distribution of fresh products plays a critical role in ensuring that the quality and safety of perishable goods are maintained during transportation and storage. However, optimizing the cold chain logistics distribution process can be a challenging task due to the complexity of the supply chain network, the varying demands and constraints, and the dynamic nature of the market. To address these challenges, recent studies focused more on various optimization techniques and have shown that hybridizing different optimization algorithms can lead to better performance and faster convergence. Therefore, this paper proposed a hybrid tabu-grey wolf optimizer algorithm for analyzing the cold chain distribution of fresh e-commerce and establishing a vehicle route optimization model to minimize the total distribution cost of the vehicle scheduling scheme. The improved algorithm integrates Tabu Search and Grey wolf optimizer algorithms, named Tabu-Grey wolf optimizer (TGWO) algorithm, which considers multiple constraints such as vehicle capacity, maximum driving distance, and coexistence of soft and hard time windows.

6.1 Theoretical implications

This study makes two aspects of theoretical contributions. Firstly, by proposing a hybrid TGWO algorithm for optimizing cold chain logistics distribution. By combining the two algorithms, the TGWO algorithm offers enhanced performance and faster convergence. This finding emphasizes the significance of employing hybrid optimization techniques to tackle the challenges posed by the complex supply chain network, dynamic market dynamics, and diverse demands and constraints. Secondly, the study specifically examines the cold chain distribution of fresh products within the e-commerce domain. Through the establishment of a vehicle route optimization model, the proposed TGWO algorithm aims to minimize the total distribution cost while considering various constraints, such as vehicle capacity, maximum driving distance, and soft and hard time windows. The effectiveness of the TGWO algorithm is demonstrated through benchmark tests and comparisons with classical optimization algorithms, including PSO, WOA and GWO. It validates the superior performance of TGWO in optimizing cold chain distribution routes for fresh products using actual distribution data from a fresh e-commerce platform. The findings contribute to the theoretical understanding of hybrid optimization algorithms in addressing the complexities of cold chain logistics distribution, particularly within the dynamic and time-sensitive environment of fresh e-commerce.

6.2 Managerial implications

The findings of this study have several managerial implications. The proposed hybrid TGWO offers a practical solution for optimizing the cold chain logistics distribution of fresh products. This approach enables managers to minimize the total distribution cost of vehicle scheduling schemes. The findings of this study highlight the importance of employing advanced optimization techniques to enhance the efficiency and effectiveness of cold chain logistics in preserving the quality and safety of perishable goods. Additionally, the study compares the performance of the TGWO algorithm with some classical optimization algorithms and demonstrate the superior performance of the TGWO algorithm in optimizing cold chain distribution routes for fresh products. This comparative analysis provides valuable insights for managers in selecting the most suitable algorithm for their specific distribution challenges. Managers can consider adopting the TGWO algorithm as a decision support tool to improve their cold chain logistics operations, reduce costs, and enhance overall supply chain performance.

6.3 Limitations and future research

This study acknowledges certain limitations that should be taken into account. This study focuses on the cold chain logistics distribution of fresh products, it is important to acknowledge the potential limitations when applying the findings to other industries or different types of products. Each industry and product category may have unique characteristics, constraints, and operational requirements that can significantly impact the effectiveness and feasibility of the proposed algorithm. Therefore, managers should exercise caution and adapt the algorithm accordingly to ensure its suitability and effectiveness in different operational settings. Additionally, the study is based on certain assumptions and constraints, which should be taken into account when applying the research findings to real-world scenarios. These assumptions may limit the generalizability of the results and require careful consideration of their applicability in specific practical situations. Managers should assess the compatibility of these assumptions and constraints with their own operational contexts to ensure the feasibility of implementing the proposed algorithm.

This study also needs some improvement in future research, such as introducing new search operators to enhance the algorithm’s performance, like inter-route search operators, or route destruction and repair operators, and other novel meta-heuristic algorithms can be designed; developing new parameters, for example, fuel consumption and vehicle speed can be considered to reduce carbon emissions, or the objective of maximizing customer satisfaction can be included; comparison with other algorithms for VRPTW, such as the improved ant colony algorithm [ 41 ] or the chaotic genetic algorithm with variable neighborhood search [ 42 ], to determine the best-performing algorithm for optimizing the cold chain logistics distribution of fresh products; application to more real-world scenarios, further research can be conducted to apply the algorithm to real-world scenarios and test its effectiveness in optimizing the cold chain logistics distribution of fresh products in practice. Overall, the research on optimizing the cold chain logistics distribution of fresh products based on a hybrid Tabu-Grey wolf optimizer algorithm has great potential for further development and improvement in the future.

Supporting information

S1 data set..

https://doi.org/10.1371/journal.pone.0306166.s001

• View Article
• Google Scholar
• PubMed/NCBI

Academia.edu no longer supports Internet Explorer.

To browse Academia.edu and the wider internet faster and more securely, please take a few seconds to  upgrade your browser .

Enter the email address you signed up with and we'll email you a reset link.

• We're Hiring!
• Help Center

Status of wolves in India

Related Papers

Canid Biology and Conservation

Prashant Mahajan , Dharmendra Khandal

The Indian grey wolf (Canis lupus pallipes) is the apex predator of the semi-arid landscapes of India. They have large home ranges and mostly thrive outside the protected areas, feeding on livestock to fulfil dietary needs, thus bringing them into direct conflict with humans, making it imperative to identify and conserve wolf-occupied areas. We used questionnaire surveys and field methods to estimate the number and status of wolves in Kailadevi Wildlife Sanctuary, Rajasthan. We estimated 19-45 wolves occurring at a density of 0.02-0.06 wolves/km 2 in 672.82 km 2 of Kailadevi Wildlife Sanctuary. The maximum number was estimated from the Nainyaki range. The presence of wolves was significantly positively related to the presence of sheep and goats. Due to low availability of natural prey in the study area, wolves depend on livestock, causing high economic loss to the resident people. Our study suggests that if strict conservation measures are taken, Kailadevi Wildlife Sanctuary holds the potential to act as a source population for the conservation of the Indian grey wolf in the larger landscape surrounding the study area. However, due to high anthropogenic pressure, the landscape is severely degraded and requires immediate attention to restore the existing scrubland for denning and rendezvous sites. Effective compensation schemes and awareness through outreach and education are required to reduce negative attitudes among the resident people and to prevent wolf persecution. Future research should make use of modern radio-telemetry techniques to better understand the ecology of the wolves in this landscape.

Krunal Trivedi

The Indian wolf (Canis lupus pallipes) is widely distributed in India, but is listed as data deficient in Gujarat. No proper study has been conducted on their current status, distribution, or threat assessment in the state. As per the current distribution, they are recorded from Northern and Central Gujarat. We present recent records of Indian wolves from Bharuch and Surat districts. We assume it to be southernmost distribution of the Indian wolf in Gujarat.

Shariq Shafi

The Indian wolf (Canis lupus pallipes) is an endangered species and known to be distributed in the vast areas of the Indian subcontinent. Intensive annual camera trapping surveys started in Valmiki Tiger Reserve in 2013. The objective of these surveys was to monitor tigers and co-predators. The study area was systematically sampled by deploying a set of two camera traps per location within a 2 km grid. The camera trap sites were determined by extensive search through sign surveys to find the ideal locations along animal trails, dirt tracks, and dry streambeds. Every year, camera traps were placed on average at 425 locations spread across 901 km. An Indian wolf was captured only once since 2013 at two different locations in the 2016-17 survey. This is the first photographic evidence of Indian wolf from Valmiki Tiger Reserve, Bihar, and might be an extension of the known eastern distribution range of the Indian wolf population from the Chhotanagpur plateau in Jharkhand and part of the...

Frontiers in Ecology and Evolution

Swati Saini

An understanding of the distribution range and status of a species is paramount for its conservation. We used photo captures from 26,838 camera traps deployed over 121,337 km2 along with data from radio-telemetry, published, and authenticated wolf sightings to infer wolf locations. A total of 3,324 presence locations were obtained and after accounting for spatial redundancy 574 locations were used for modeling in maximum entropy framework (MaxEnt) with ecologically relevant covariates to infer potentially occupied habitats. Relationships of wolf occurrence with eco-geographical variables were interpreted based on response curves. Wolves avoided dense wet forests, human disturbances beyond a threshold, arid deserts, and areas with high top-carnivore density, but occurred in semi-arid scrub, grassland, open forests systems with moderate winter temperatures. The potential habitat that can support wolf occupancy was 364,425 km2 with the largest wolf habitat available in western India (S...

Himanshu Shekhar Palei

The Indian grey wolf (Canis lupus pallipes) is a rare and lesser-known top predator in India. A rapid camera trapping survey was conducted to assess the large carnivores and their preys in the Sundargarh forest division, Odisha, India. Two individuals of Indian grey wolf were recorded during the survey offering the first photographic evidence of the Indian grey wolf outside protected areas of Odisha. This record increases knowledge on the distribution of the species. More extensive surveys are needed to understand the distribution and population dynamics of Indian grey wolf in the area. We provide photographic evidence of Indian grey wolves and highlight the importance of Odisha forest for species conservation.

Biological Conservation

Bharat Jethva

gokarna thapa

The grey wolf Canis lupus lupus is Critically Endangered in Nepal, and is a protected species there. Understanding the species’ status and distribution is critical for its conservation in the Nepalese Himalaya. We assessed the distribution of the grey wolf in the Himalayan and Trans-Himalayan regions using data from faecal and camera trap surveys and published data sources. We recorded 40 instances of wolf presence. Using these data we estimated a distribution of 28,553 km2, which includes potential as well as known habitat and comprises 73% of the Nepalese Himalaya. There is evidence of recovery of the grey wolf population in Kanchenjunga Conservation Area in the eastern portion of the species’ range. A livestock insurance scheme has been shown to be a viable option to reduce retaliatory killing of wolves as a result of livestock depredation. The wolf plays an important ecological role in the Himalaya, and its conservation should not be delayed by the ongoing taxonomic debate about...

Karma Sonam

We describe a pilot community-based conservation initiative for wolves Canis lupus that involves (i) voluntary deactivation of traditional trapping pits called Shandong, (ii) commitment to wildlife conservation by the local community, and (iii) collaborative construction and consecration of a Stupa (Buddhist shrine) in the vicinity of the Shandong as a symbol of conservation and repentance for past hunting. People and wolves have a complex relationship, in part shaped by predation on livestock, which can have severe impacts on livelihoods in pastoral societies. Consequently, wolf conservation often evokes strong and polarizing reactions. To control wolf populations, livestock herders across the Trans- Himalayan and Tibetan regions use different types of traps. Shandong is a relatively large, widely used traditional trapping pit with inverted funnel-shaped stone walls, usually built near villages or herder camps. Typically, a live domestic animal is placed in the pit to attract the w...

shaileshkumar desai

The Indian grey wolf is a crucial apex predator in India's semi-arid region, but their wide range and elusive behavior make population estimation challenging. Accurate population estimation is essential for effective management and conservation efforts. Our study focused on the bhal region of Gujarat and used local community information validated by ground surveys. Our estimates suggest a population of 64-88 wolves with a density of 0.021-0.029 wolf/km 2. However, despite the region's continued distribution, it faces numerous threats. To protect the Indian wolf, compensation schemes, community participation, and protection of traditional breeding areas are necessary.

Honnavalli Kumara

Abstract The Indian gray wolf Canis lupus pallipes is the major large carnivore in the plains of Karnataka, India. We carried out a study on its distribution and status from November 2001 to July 2004. We estimated 555 wolves occupying about 123 330 km 2 of the state. In the past 40 years, wolves have disappeared from the southern plateau from an area of about 31 801 km 2. Their distribution is now largely restricted to the north-eastern dry plains.

Loading Preview

Sorry, preview is currently unavailable. You can download the paper by clicking the button above.

RELATED PAPERS

sikenvis.nic.in

S Sathyakumar

Journal of Threatened Taxa

Bishnu Pandey

Ashwin Naidu , Kenneth Walker

Nishith A Dharaiya

SANDEEP R A N J A N MISHRA

Aparajita Datta

Current Science

Naresh Kusi , Geraldine Werhahn

S. Sathyakumar

Genome Biology

Jetty Ramadevi

Adrian Lyngdoh

Indian Forester

Dr. Aazad P Ojha

The Initiation

Buddi Poudel

Talk at National Seminar, 29 Oct 2021

Lala Aswini Kumar Singh

Nazimur Rahman Talukdar

Orus Ilyas , Shahzada Iqbal

Shiva Venkat

Dr. Vipul Maurya

Kamlesh Maurya

•   We're Hiring!
•   Help Center
• Find new research papers in:
• Health Sciences
• Earth Sciences
• Cognitive Science
• Mathematics
• Computer Science
• Academia ©2024