Introduction
Certain fish groups have important functional roles in structuring coral reef ecosystems, including corallivory (Cole et al., Reference Cole, Lawton, Pratchett and Wilson2011), herbivory (Diaz-Pulido & McCook, Reference Diaz-Pulido and McCook2003; Hughes et al., Reference Hughes, Rodrigues, Bellwood, Ceccarelli, Hoegh-guldberg and McCook2007; Burkepile & Hay, Reference Burkepile and Hay2010) and piscivory (Jennings & Polunin, Reference Jennings and Polunin1997; Boaden & Kingsford, Reference Boaden and Kingsford2015). Conservation of such fish groups is a key priority for ecosystem managers (Nyström, Reference Nyström2006; Bellwood et al., Reference Bellwood, Hughes and Hoey2006). One such fish that is also of conservation concern is the bumphead parrotfish Bolbometopon muricatum, the largest herbivorous and corallivorous fish, categorized as Vulnerable on the IUCN Red List (Chan et al., Reference Chan, Sadovy and Donaldson2012). An individual can consume up to 5 t of structural reef carbonate per year (Bellwood et al., Reference Bellwood, Hoey and Choat2003), and the species promotes coral growth and recruitment by balancing coral erosion and calcification, preventing macroalgal growth and maintaining sediment flow in the reef ecosystem (Bellwood et al., Reference Bellwood, Hoey and Choat2003, Reference Bellwood, Hoey and Hughes2011; Kobayashi et al., Reference Kobayashi, Friedlander, Grimes, Nichols and Zgliczynski2011; McCauley et al., Reference Mccauley, Young, Guevara, Williams, Power and Dunbar2014; Roff et al., Reference Roff, Doropoulos, Mereb and Mumby2017).
A large body size, aggregating behavior and limited activity at night make B. muricatum an easy target for spearfishers (Bellwood et al., Reference Bellwood, Hoey and Choat2003; Donaldson & Dulvy, Reference Donaldson and Dulvy2004; Kobayashi et al., Reference Kobayashi, Friedlander, Grimes, Nichols and Zgliczynski2011; Chan et al., Reference Chan, Sadovy and Donaldson2012; Munoz et al., Reference Munoz, Zgliczynski, Teer and Laughlin2014; Hamilton et al., Reference Hamilton, Almany, Stevens, Bode, Pita, Peterson and Choat2016). Combined with slow growth and low replacement rates, this has resulted in population declines across the Indo-Pacific and Red Sea (Kobayashi et al., Reference Kobayashi, Friedlander, Grimes, Nichols and Zgliczynski2011; Chan et al., Reference Chan, Sadovy and Donaldson2012; Andrews et al., Reference Andrews, Choat, Hamilton and DeMartini2015). These declines have been particularly marked in the Solomon Islands, Papua New Guinea, Samoa, Guam, the Marshall Islands and parts of Malaysia and Fiji (Bellwood et al., Reference Bellwood, Hoey and Choat2003; Aswani & Hamilton, Reference Aswani and Hamilton2004; Dulvy & Polunin, Reference Dulvy and Polunin2004). Many countries have implemented conservation measures for the species, including establishment of marine protected areas and restrictions on spearfishing (Dulvy & Polunin, Reference Dulvy and Polunin2004; Weeks et al., Reference Weeks, Green, Joseph, Peterson and Terk2017).
In India B. muricatum occurs from the Gulf of Mannar in Tamil Nadu state (Varghese et al., Reference Varghese, Manisseri, Ramamurthy, Geetha, Thomas and Gandhi2011), to the Lakshadweep archipelago (Rohan Arthur, pers. comm.) and the Andaman and Nicobar Islands (Rajan et al., Reference Rajan, Sreeraj and Immanuel2013). Little is known about the species' distribution and conservation status around the Andaman and Nicobar Islands. In the recent past the islands have been affected by a series of coral bleaching events (in 1998, 2002, 2010 and 2016) and the Tsunami of 2004, which may have disturbed the species’ habitat (Jeyabaskaran and Rao Reference Jeyabaskaran and Rao2007; Patankar et al., Reference Patankar, D'Souza, Kumaraguru and Arthur2012; Mondal et al., Reference Mondal, Raghunathan and Venkataraman2013).
For species that are rare, threatened or Data Deficient, local ecological knowledge has been used to examine population trends (Aswani & Hamilton, Reference Aswani and Hamilton2004; Lavides et al., Reference Lavides, Polunin, Stead, Tabaranza, Comeros and Dongallo2009; Lavides et al., Reference Lavides, Molina, de la Rosa, Mill, Rushton, Stead and Polunin2016; Pan et al., Reference Pan, Wei, Cunningham, Li, Chen, Milner-Gulland and Turvey2016). Studies that have combined ecological data from underwater surveys with information from fishers have shown that both sources can provide reliable information (Iniesta-Arandia et al., Reference Iniesta-Arandia, del Amo, Garcia-Neto, Pineiro, Montes and Martin-Lopez2014; Zappes et al., Reference Zappes, Gatts, Lodi, Simões-Lopes, Laporta, Andriolo and Di Beneditto2014). In addition, the inclusion of local ecological knowledge in conservation research can increase the involvement of local stakeholders in conservation activities (Drew, Reference Drew2005). We therefore used a combination of surveys and interviews to assess the abundance, distribution and conservation status of B. muricatum across the Andaman and Nicobar Islands. Using generalized linear models, we test whether abundance is influenced by a combination of benthic variables and the presence of marine protected areas, and discuss our findings in the light of developing conservation strategies for B. muricatum around these islands.
Study area
The Andaman and Nicobar archipelago of India is part of the Indo-Myanmar and Sundaland biodiversity hotspots in the south-eastern Bay of Bengal (Davidar et al., Reference Davidar, Devy, Yoganand, Ganesh, Boyle and Boontawee1995; Roberts et al., Reference Roberts, McClean, Veron, Hawkins, Allen and McAllister2002). The archipelago comprises 350 islands and a land area of 8,249 km2, with a total coastline of 1,962 km. It includes the Andaman group (> 325 islands, 24 inhabited, 6,408 km2) and the Nicobar group (21 islands, 13 inhabited, 1,841 km2), separated from each other by the Ten-degree Channel. The Andaman Islands have 105 protected areas (9 national parks and 96 wild sanctuaries), all of which encompass coast and surrounding waters, and the Nicobar Islands have seven community-protected marine areas, with restrictions on fishing and other resource use activities (Patankar et al., Reference Patankar, D'Souza, Alcoverro and Arthur2015). The islands are influenced by both the south-west and north-east monsoons (May–December).
Methods
We surveyed sites across the archipelago, to account for variability in coastal morphology, oceanographic conditions, geography and reef resources. In total we surveyed six sites at Camorta Island, four at Great Nicobar Island and Nancowry Island, two at Aves, Katchal, Little Nicobar, Interview, Twins, Cinque, Sister, Rutland and Eastern Reef islands, and one site at all other locations. The number of sites around each island were selected based on the size of the island, aspect, and location and accessibility of the reef (Fig. 1).
Density and abundance
At each site we estimated the abundance of B. muricatum, carrying out an underwater visual census at a total of 75 reef sites across 51 islands, using SCUBA diving, during November 2013–April 2015. At each site we counted B. muricatum along five random 50 × 10 m transects (500 m2) delineated with a 50 m fiberglass tape. Identification of B. muricatum is relatively straightforward because of its large size and the presence of a unique bulbous bump on the head. On sighting, we counted the number of individuals and visually estimated their size. A prior calibration was conducted to refine the accuracy and precision of our underwater size estimates until they were within 10% of actual lengths.
Benthic cover
We quantified the benthic cover (live coral cover, turf algae, macroalgae, sand, rubble, soft coral and others) at each transect by photographing five 1 × 1 m quadrats, one at each 10 m along the 50 m transect, using an underwater camera without flash. Per cent benthic cover categories were measured using Adobe Photoshop 7.0 (Adobe Systems, San Jose, USA).
Documentation of local knowledge
To investigate the awareness and perception of B. muricatum by fishers we conducted a total of 99 semi-structured interviews (Huntington, Reference Huntington2010) during January 2014–May 2015, in three villages in South Andaman (Wandoor, 38 interviews; Junglighat, 15; Mohanpura, 7), two villages in Mayabunder in Middle Andaman (17) and five villages in the Nicobars (Tapong, 4; Hitui, 4; Ramjav, 7; Masala Tapu, 4; Kakana, 3). These are the major fishing villages in the archipelago. All respondents were male, as fishing activities are predominantly carried out by men in this region. The number of interviews in each region was based on the intensity of fishing, the ethnic heterogeneity of the fishing community and the diversity of fishing gear used. Thus, there were 60 interviews in South Andaman, which has the highest number of fishing vessels, several ethnic groups (Telugu, Bengali, Malayali and Tamil) and a high diversity of fishing gear (gillnets, trawls, long lines, hook and lines, cast nets, and hand-held wooden spears and harpoons), and 17 in Middle Andaman and 22 in Central Nicobar, where spearfishing and hook-and-line are the common fishing methods used, by Karen and Nicobari communities. We used the snowball sampling method to select the interviewees in each village (Kvale, Reference Kvale1996). Interviews were conducted in Hindi in South Andaman, and in in Karen and Nicobari in the Middle Andaman and Nicobar Islands, respectively.
To minimize any potential bias, all interviews were carried out by VP, with help of a local interpreter in the Middle Andaman and Nicobar Islands. Most interviews lasted 10–20 minutes and were conducted at fishers’ houses or at fish landing sites. The questions included personal data about the respondents (island and village of residence, age, gender, level of education), preferred fishing gear, target and non-target species captured, and awareness of B. muricatum. In particular, the fishers were asked if they were aware of B. muricatum in their waters, and were shown photographs, to avoid erroneous results arising from the use of different local names for the same fish. We then asked fishers how many times they had seen the fish in the previous year, and how many times they had seen aggregations (i.e. > 25 individuals). When people provided information about their experience of catching B. muricatum, we requested descriptive details of the method of fishing, time of fishing (day/night) and where they fished. Interviews were recorded in writing and later transcribed for quantitative analysis. Before an interview, respondents were provided with information about the nature and expected goals of the study, complete confidentiality was assured, and interviews proceeded only with consent of the interviewee.
Data analysis
We used descriptive statistics to analyse interview data. Bolbometopon muricatum spends most of its time foraging on structural reef carbonate (Bellwood et al., Reference Bellwood, Hoey and Choat2003; Donaldson & Dulvy, Reference Donaldson and Dulvy2004) and, as marine protected areas of sufficient size can protect such fish species (Sadovy Reference Sadovy2005; Green et al., Reference Green, Maypa, Almany, Rhodes, Weeks and Abesamis2015), we expected higher abundances in protected areas with a high per cent of live coral cover. The number of sites without B. muricatum was high (66 of 75) and therefore we used zero-inflated Poisson generalized linear models (GLM) to model abundance. An exploratory analysis revealed that these models performed better than a GLM with Poisson error (Vuong test statistic = 2.04, P < 0.05). The zero-inflated Poisson GLM is a two-step process in which abundances are modelled as count variables with a Poisson distribution conditional on the probability that the values are non-zero (a binomial variable). As B. muricatum was encountered at only a few sites we did not include multiple variables in a single model, to avoid overfitting. We compared the effects of live coral, soft coral and others, turf algae, and with and without a marine protected area. If the zero abundances in the data can be attributed entirely to the species’ biology, then the zero model should only contain a single intercept term. We used package pscl 1.4.9 in R 3.4 for regression analysis (R Development Core Team, 2016).
Results
Abundance and distribution
Around the 51 islands surveyed, we sighted a total of 59 individual B. muricatum at nine islands (seven in the Andamans and two in the Nicobars), with Sister Island in South Andaman having the highest abundance (29 individuals) and the largest shoal size (18 individuals), followed by Craggy Island in North Andaman (one shoal of 12 individuals), six individuals at Nancowry, four individuals each at Rutland and John Lawrence islands, and a single individual at Great Nicobar, Landfall, Cinque and Neil Islands. All the fish were > 50 cm except one subadult (< 50 cm) at Neil Island. We did not observe any juveniles. This could be because the habitat we sampled is not preferred by juvenile reef fish (Hamilton et al., Reference Hamilton, Almany, Brown, Pita, Peterson and Howard Choat2017). The average density of B. muricatum was 0.0262 individuals per ha at the nine sites where we sighted the fish, and the density across the entire archipelago was 0.0032 individuals per ha.
Influence of benthic cover and protected areas
Rutland Island in South Andaman had the highest live coral cover (mean 52.5 ± SE 0.1%, 2 sites); the lowest was recorded at Excelsior (n = 1), Kwantung (n = 1) and Landfall (n = 1) Islands (mean 21 ± SE 0.1%). Thirty-one islands surveyed had an intermediate cover of live coral (29–41%). Algal cover was low at most sampled sites, with the highest at Smith Island (19%, n = 1) in North Andaman and lowest (1%) at Katchal, West, Redskin, Kondul and Menchal Islands. Zero-inflated Poisson regression models for the effects of benthic variables and presence/absence of a marine protected area on the abundance of B. muricatum showed that live coral cover was an important predictor and any additional effects of benthic variables on the zero inflation part of the model could not be identified. The effects of live coral and presence of a protected area were positive in all count models. The best model contained effects of live coral, rubble and sand and management variables (Table 3).
Fishers characteristics
Of the 99 male respondents 53% were ≤ 40 years of age and 47% > 40, and level of education was low (most respondents had not completed secondary education). All interviewees were full-time fishers, except in South Andaman where 33% were part-time fishers, with farming as their alternative livelihood. All fishers used more than one type of fishing gear, to diversify catch composition. Hook-and-line, hand-held wooden spears and harpoons and gillnets were the most common fishing methods. Amongst these, hook-and-line was the most commonly used fishing gear across regions, followed by gills nets, spears and harpoons. The characteristics of respondents are summarized in Table 1.
Awareness/knowledge of B. muricatum
Most fishers were aware of the presence of B. muricatum in their waters (see Table 2 for a summary of respones to all questions). Amongst these, all fishermen in Central Nicobar and Middle Andaman had seen aggregations of the fish. Bolbometopon muricatum was most commonly referred to as tota macchi, which translates to parrotfish in Hindi, the most widely spoken language in the archipelago. All respondents in Middle Andaman and Nicobar had seen B. muricatum feeding, whereas in South Andaman only 22% had seen feeding. All the interviewed fishers who were aware of the presence of the fish in their waters were also aware that B. muricatum is not legally protected in the Andaman and Nicobar Islands (Anon., 1992). Most fishers in the Nicobars (91%), all fishers in Middle Andaman, and 8% of fishers in South Andaman had hunted the fish in their lifetime, using hand-held wooden spears or harpoons, mostly in daylight. All fishers who had caught the fish in their lifetime reported that the catch was opportunistic rather than targeted.
*Count model describes covariates for abundance data, and zero model describes covariates for sites with zero abundance.
Discussion
We carried out the first investigation of the distribution and abundance of the Vulnerable B. muricatum in the Andaman and Nicobar archipelago, using both quantitative and qualitative approaches. The species occurs broadly, from the northernmost island in Andaman (Landfall Island) to the southernmost island in the Nicobars (Great Nicobar Island) and, except for fishers from South Andaman, most respondents were aware of the presence of this fish, its aggregation behavior and had seen it feeding. Although we sighted 59 individuals from nine islands, the fish occurs patchily, with most sightings from only two islands, and with an apparently naturally low density. The mean density we recorded (0.0032 per ha) is comparable to that observed on reefs of Malaysia (0.0042 per ha), Myanmar (0.0016 per ha) and Thailand (0.0004 per ha; Kobayashi et al., Reference Kobayashi, Friedlander, Grimes, Nichols and Zgliczynski2011). These densities are, however, markedly lower than those in other areas (Great Barrier Reef: 30 per ha; Solomon Islands: 0.3–4.2 per ha; Hamilton & Choat, Reference Hamilton, Choat, Sadovy de Mitchson and Colin2012).
In areas where B. muricatum is a fishery resource, free diving spearfishers exclusively target nocturnal aggregations (Comeros-Raynal et al., Reference Comeros-Raynal, Choat, Polidoro, Clements, Abesamis and Craig2012). However, in the Andaman and Nicobar Islands all fishers reported that catch was opportunistic rather than targeted. The main fishing gear used in the islands is hook-and-line and, other than spear or harpoon fishers, most do not encounter B. muricatum on their regular fishing trips.
We found that the presence of a protected area, live coral and algal cover significantly influenced the distribution and abundance of B. muricatum. Although changes in the composition of these benthic characteristics as a result of disturbances could have negatively affected the abundance and distribution of this species in the Andaman and Nicobar Islands (Krishnan et al., Reference Krishnan, Dam Roy, George, Srivastava, Anand and Murugesan2011; Patankar et al., Reference Patankar, D'Souza, Kumaraguru and Arthur2012), no baseline data are available. Marine protected areas in which fishing and other activities are restricted aid the conservation of fish stocks (Donaldson & Dulvy, Reference Donaldson and Dulvy2004) and in the Andaman and Nicobar Islands the existing protected areas may be vital for the protection of B. muricatum and other fishes. Evidence of low abundance of B. muricatum on ocean reefs surrounded by deep waters, and traits such as limited dispersal and gregariousness, could also have influenced the distribution and abundance of this fish (Hamilton & Choat, Reference Hamilton, Choat, Sadovy de Mitchson and Colin2012; Munoz et al., Reference Munoz, Zgliczynski, Teer and Laughlin2014).
The current potential threats to B. muricatum in the Andaman and Nicobar Islands include incidental catch by fishers and degradation of coral reef habitats. Protected area designations and threatened species legislation are effective only if social conditions also encourage self-regulation (Sawchuk et al., Reference Sawchuk, Beaudreau, Tonnes and Fluharty2015). Although the present protected areas are supporting higher abundances of B. muricatum, fishing and benthic degradation could potentially result in future declines. We recommend extensive long-term population studies of B. muricatum in the archipelago, a regional Red List assessment for the species and inclusion in the Indian Wildlife (Protection) Act 1972. As reef fishery is growing, it will be necessary to implement regulations to avoid population declines, e.g. a ban on night fishing for B. muricatum. The latter would not compromise the livelihood of the fishing community as catch of B. muricatum is opportunistic. Considering the Vulnerable status of B. muricatum globally, the species could be a flagship for educational campaigns focusing on the importance of conserving similar functionally important fish groups and protecting the coral reefs of the islands.
Acknowledgements
The study was conducted with funding from the Research Fellowship Program, Wildlife Conservation Society, and Ravi Sankaran Inlaks Fellowship Program. We thank the Department of Environment and Forests, Port Blair, for granting us permits to carry out this work and for a boat for reef surveys; M. Gangal, R. Arthur, T. Alcoverro, N. Kelkar and E. D'Souza for support and advice; the Andaman and Nicobar Island's Environmental Team and Nature Conservation Foundation for help with SCUBA equipment and field logistics; and Saw John, Elrika D'Souza, Saw Berny, Saw Alexander, Saw Sawda, Tanvi Vaidyanathan and Sahir Advani for field assistance.
Author contributions
Study design, data collection and writing: VP; assistance with writing and preparation of map: TW; data analysis: AM.
Conflicts of interest
None.
Ethical standards
Nicobarese are a Scheduled Tribe of India, and entry into the islands is regulated by the Andaman & Nicobar Islands Protection of Aboriginal Tribes Regulation 1956. We obtained the requisite Tribal Area Entry Permits, to land and conduct research, from the Department of Tribal Welfare and District Commissioner's office at Port Blair. During each visit we also obtained permission to conduct research from the Chief Captain of the village and Chairperson of the Tribal Council of the area. We obtained all requisite clearances from the Nature Conservation Foundation's Institutional ethical committee and from the doctoral research committee of Marurai Kamaraj University prior to conducting fieldwork.