Study site and fishery

Permission to conduct this research within the Maldives was granted by the Marine Research Centre, Republic of Maldives; no specific permits were required. The Republic of Maldives is an archipelagic atoll chain straddling the equator in the central Indian Ocean, running north-south along 73°E from about 7°N to 1°S (Fig 1). There are some 1,200 islands, grouped across 26 natural atolls. Some 185 of these islands are inhabited.

Download:

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

Fig 1. Map of Maldives.

https://doi.org/10.1371/journal.pone.0177391.g001

Pole-and-line fishing is carried out using traditional fishing boats known as mas dhonis. There are currently about 1,000 such vessels, which are typically about 18–30 m long. While the fishery remains fundamentally the same as it has been for centuries, mechanization and technological improvements and mechanization have dramatically increased efficiency in recent decades.

The livebait pole-and-line fishery consists of two distinct components: catching baitfish and catching tuna. In the Maldives these are invariably conducted from the same vessel, typically within the same fishing trip. Baitfish are usually caught inside the atolls using simple lift-nets deployed from the boat and subsequently kept alive in the vessel. A variety of small fishes are used as livebait, the most important being the silver sprat, Spratelloides gracilis (see S3 Table). The Maldivian baitfishery has been relatively well-studied [32–36] and this component of the livebait pole-and-line fishery is not considered further here.

Tuna fishing occurs offshore, in the open ocean, typically within about 12–70 nautical miles (22–130 km) of the atolls [37]. Fishermen locate schools of tuna most frequently by the presence of seabirds or another fishing boat, or by association with a floating object. Floating objects include natural logs, drifting Fish Aggregating Devices (dFADs, originating from the western Indian Ocean purse seine fishery) and Maldivian anchored Fish Aggregating Devices (aFADs). The Maldivian Ministry of Fisheries and Agriculture maintains a network of some 50 aFADs deployed throughout the country, mostly within about 12–15 nautical miles (22–28 km) of the atolls [38]. Approximately 50% of the Maldivian pole-and-line tuna catch is currently taken from these aFADs [39], up from an estimated 44% in the early 1990s [40].

When a school of tuna is located, water is sprayed on the surface of the sea and livebait are thrown into the water, to encourage a feeding frenzy. Several (5–15) fishermen stand at the stern of the vessel and use non-baited barbless hooks with lures to catch the tuna and flick them on board. Fish are stored on ice, and boats typically return to shore to sell their catch the same day.

Bycatch monitoring

In this study we considered all tuna landings to be ‘catch’, with everything else being ‘bycatch’, following the definition of the Indian Ocean Tuna Commission [41]. ‘Bycatch’ is subdivided here into ‘byproduct’ (non-tuna species that are retained for sale or consumption), ‘discards’ (any catch that is discarded dead, including damaged tunas) and ‘release’ (any catch that is released alive). However, we summarize our results in sufficient detail that they can be re-worked under different definitions.

One hundred and six fishing days were observed between August 2014 and November 2015 (S1 Table) by experienced fishery observers. In most cases (n = 89) there were two observers, on 17 days three observers and a single observer on just one trip. The trips were planned to cover both monsoon seasons, and all regions of the Maldives, but with most effort in the south where pole-and-line fishing is now concentrated.

On site, vessels were chosen opportunistically; vessels ranged from 16–35 m in length, with 8–23 crew. Most fishing trips started in the late evening, included night baitfishing, a full day searching and fishing for tuna offshore, and a return to shore in the early evening. On one occasion the fishing boat stayed out at sea overnight, tuna fishing on two successive days (thus there were 105 fishing trips, and 106 observed days). There were 86 trips in which tuna fishing was conducted, on 87 days. Most trips in which fishing was not conducted consisted of either trips with insufficient bait catch or days in which schools of tuna could not be located. Mean trip duration was 22h 08min (+/- 6h 45min). During the 106 days, baitfishing was carried out on 86 days, with 99 baitfishing events observed (on the other 20 occasions livebait was still available from the previous trip). A troll line was sometimes towed behind the fishing vessel, while searching for tuna. Three skipjack and one brown lesser noddy (Anous tenuirostris) were caught by troll, but are included with the other pole-and-line catch and bycatch in the results tables.

During tuna fishing, detailed records of fishing activities and catch were maintained. Effort data included vessel details, number of fishermen/hooks, amount and type of bait used. Position, school association, sighting method and duration were recorded for each fishing event. Fishing events were defined as periods of active fishing that were separated by more than 10 minutes. School association was recorded as one of six categories: anchored FAD (aFAD), drifting FAD (dFAD), natural log, floating man-made object, seamount, or free school. Schools were defined as associated if the start of fishing was within 1 nautical mile (1.85 km) of any FAD or other floating object, or within 5 nautical miles (9 km) of a seamount following [9].

Because of safety concerns, sampling of the catch and bycatch could not be carried out during active fishing. Sampling was carried out during fishing events if there were brief stoppages in active fishing while the vessel steamed to reposition on the tuna school. Otherwise, sampling was carried out at the end of each fishing event.

For tuna catch, it was not possible to measure or weigh all skipjack and yellowfin caught, as fishermen were keen to put the catch on ice as quickly as possible. One hundred tunas were identified to species and measured (fork length) per fishing event. Also, for each fishing event, the total weight of tuna caught was estimated, and the proportion of skipjack and yellowfin was calculated based on number and average weight of 100 fish. For bigeye tuna (which were not separated from yellowfin tuna at sale), numbers were estimated by close inspection of the external appearance of a subset of up to 50 Thunnus per fishing event. Numbers of minor tuna species (frigate tuna [Auxis thazard] and kawakawa [Euthynnus affinis]) were counted on board immediately after capture; subsamples were measured and weighed. The total weights of major species (skipjack and yellowfin) were recorded when catch was weighed at sale.

For bycatch species, every individual caught was identified to the lowest taxon possible, counted, and its fate recorded. A subsample of each species was measured and weighed. The estimated size and condition of every individual discarded or released was recorded.

Qualitative notes describing activities were recorded daily, and a written report produced at the end of each multi-day sampling trip. A detailed description of the observer sampling protocol is available in [42].

Data analysis

As it was not possible to measure or weigh all catch and bycatch, some weights were estimated. For the main tuna species (skipjack and yellowfin), almost all the catch was sold at the end of each day’s fishing and was weighed by the buyer at the time of sale. In some cases, skipjack and yellowfin were weighed separately. In other cases, when skipjack and yellowfin were weighed together, the total weight was apportioned between the species according to the recorded proportions in the catch and their average weights. Average weights were estimated from the lengths measured, using established length-weight conversion factors (S2 Table). On days when there was more than one fishing event, total catch was then further apportioned between fishing events based on catch estimates made on board. Bigeye catch was estimated as a proportion of the Thunnus catch (i.e. what had been sold as yellowfin tuna), based on the sampling carried out that day.

Occasionally, bycatch individuals could not be weighed (e.g. when released alive by fishermen); in these cases it was sometimes possible to measure lengths, but more frequently lengths were estimated. Species specific length–weight relationships (S2 Table) were then applied to convert into weights. Two-tailed Mann-Whitney U tests were performed on bycatch rates for each fishing event.

Variation in bycatch rates

The results of this study demonstrate that there is considerable variability in bycatch rates. The most important factor influencing bycatch rate was tuna school association. Free schools were typically composed almost entirely of large skipjack, and yielded little bycatch. In contrast, associated schools contained mostly smaller-sized skipjack together with juvenile yellowfin and bigeye tunas, and yielded much more bycatch. Similar differences in tuna composition and bycatch rates between free schools and FAD schools are known from the purse seine tuna fisheries (e.g. [9, 54]), which also show variation by ocean (region) and retention or discard of small tunas. Indeed, purse seine bycatch ranged from a low of 0.9% for seamount-associated schools in the Indian Ocean [9] to 15.0% for FAD-associated schools in the Atlantic [55]. For the Maldivian pole-and-line fishery there may be differences between different types of associated schools, but the quantities of bycatch taken from the different school types during this study are too small to make meaningful comparisons.

Within the Maldives there are clear regional differences in bycatch composition. Bycatch rates in the north of Maldives (north of 2°N) were significantly higher (up to three times greater) than in the south (Table 5). There were also differences in species composition. Bycatch from the north included more rainbow runners and round scad, while dolphinfish dominated in the south. Among tunas, bigeye was relatively more abundant in the south of Maldives, while kawakawa and frigate tuna were relatively more abundant in the north. Differences in relative abundance of different pelagic fish species between these regions have been attributed to differences in oceanographic conditions [56].

It should be noted that Maldivian pole-and-line fishermen can clearly see what they are catching, and will not continue catching species which they cannot sell or consume. During this survey it was apparent that there was greater demand for bycatch species in the north of Maldives than in the south. This was partly because the main southern fish buyers would only buy skipjack, yellowfin and bigeye, while northern buyers would buy other species as well. Also, many fishermen in the north have access to Malé fish market (by far the largest in the country), where they could sell many species of fish. These factors may have encouraged northern fishermen to take more bycatch.

It is likely that there are also seasonal differences in bycatch. However, these are not apparent in our data, in part perhaps because of the relatively small sample sizes.

The issue of bait

In this study, as in most similar studies, the livebait component of the pole-and-line fishery was not considered bycatch due to different target species, different gears, different locations, and different management issues. For these reasons the catch of livebait is usually excluded from the definition of bycatch [13, 14, 15, 44]. One exception to this standard practice was provided by [25], wherein livebait was included in bycatch estimates for the Maldivian pole-and-line fishery and, based on published information from MRC, noted bycatch at 11.6% of the tuna catch. For comparison with [25], during this study it was estimated that total livebait catch was 13.0 t. That was equivalent to 8.9% of the tuna catch, or 11.3 kg of tuna per 1 kg of livebait caught. This is comparable to, but slightly more efficient than, catch rates estimated by previous Maldives studies (summarised in [36]), which reported catches within the range of 7.4–10.0 kg of tuna per 1 kg of livebait used, or 10.0–13.5% of tuna catch.

Tuna catches

While this report deals specifically with bycatch, some information on the tuna catch is relevant to interpreting the bycatch data.

Different school types showed markedly different tuna species composition and different sizes of skipjack tunas (Table 2 and Fig 2A). Skipjack made up nearly all (98%) the tuna catch from free schools but only half (51%) from associated schools. Furthermore, the skipjack caught from free schools were mostly uniformly large-sized, whereas smaller skipjack predominated in associated-school catches (Fig 2A and S5 Table). Across the entire study, skipjack lengths showed a bimodal distribution (Fig 2A); this is a well-known feature of Maldivian skipjack landings [57, 58]. Also, across the entire study, skipjack contributed 72.3% to the total tuna catch, which is close to the long-term estimated average for the national fishery of 75% [59].

The yellowfin tuna taken by pole-and-line were all juveniles, with an average fork length of roughly 40 cm in both free schools and associated schools (Fig 2B, S5 Table). This agrees with previous findings [60].

Juvenile bigeye tuna makes up a small proportion of the Maldivian pole-and-line catch (Table 2), but these fish are not always easy to distinguish from juvenile yellowfin tuna. Indeed, bigeye catches were generally not separated from yellowfin catches in official Maldivian catch statistics until 2014, when bigeye catches began being recorded separately in the logbooks. In this study, the proportion of bigeye in the Thunnus component of the pole-and-line catch was estimated at 6.1%, but with a higher proportion in the south (7.3%) than the north (2.7%) (S7 Table). This agrees with previous studies [61–63], which all found a higher proportion of bigeye in the south of Maldives than in the north and centre (S7 Table). In neighbouring Sri Lanka, inspection of ‘small yellowfin’ landings at Beruwela (6°29’N, the same latitude as Shaviyani Atoll in the north of Maldives) found 0.7% bigeye [64]. The cause of this latitudinal variation in bigeye tuna abundance is unknown, but may be related to a known deep oxygen minimum zone [65, 66] that could affect deep-swimming species such as bigeye.

Comparison with other fisheries

In comparison to other fisheries, a bycatch rate of 0.5%-1% is very low. For example, there are estimates of global average bycatch from shrimp trawl of 62%, tuna longline of 28%, and tuna purse seine at 5% [15]. Worldwide, total pole-and-line fishing discards were estimated at 3,121 t, compared with 144,152 t for tuna purse seine in 2005 [15].

Within the Indian Ocean, in addition to pole-and-line, the other major tuna fishing gears are gillnet, purse seine and longline. In 2015, reported catches for tuna and tuna-like species were 534,000 t for gillnet (33% of the total), 415,000 t for purse seine (26%) and 260,000 t for longline (16%), compared to 102,000 t for pole-and-line (6%) (catch data from www.iotc.org, accessed January 2017).

Gillnet fisheries are not only the largest tuna fisheries in the Indian Ocean but are also believed to have the highest levels of bycatch. However, they have been poorly documented. Most of the bycatch sampling to date has been of landings, not catches, so discards are unknown. However, recent research in Pakistan, which incorporated observer data, reported bycatch amounting roughly 40% of tuna landings [67, 68]. If that were representative of all Indian Ocean tuna gillnet fisheries, then it would suggest bycatch in excess of 210,000 t in 2015. It should be noted, however, that a large part of the bycatch is of other edible finfish and of sharks, most of which are sold as byproduct.

The tuna purse seine fishery in the Indian Ocean is dominated by French and Spanish fleets. For these fleets, bycatch was sampled during 2003–2009 by onboard observers, and estimated at 11,590 t per year, which was 4.7% of tuna landings [69]. This likely underestimated total bycatch, since marine mammals and whale sharks were caught but excluded from the analysis, and rare catches of large numbers of some bycatch species were under-represented in the observer samples [69]. In addition, the proportion of dFAD-sets by tuna purse seiners has been increasing in recent years [70], which has likely resulted in the proportion of bycatch increasing. Nevertheless, if 4.7% were representative of all Indian Ocean tuna purse seine fisheries, then it would suggest bycatch of over 20,000 t in 2015.

Longline tuna fisheries cover all tropical and temperate waters of the Indian Ocean. Although the basic longline technique is fairly standard, there is much variation in gear materials and configuration, in setting depth and between ecological regions. Thus, there is considerable variation in tuna catch and bycatch, and bycatch rates can be particularly high in some tuna longline fisheries. For example, seabird bycatch has been unsustainably high in the southern Indian Ocean (e.g. [5, 71]), while shark bycatch can be especially high in tropical waters (e.g. [72, 73]). With such variability, it is difficult to make a robust estimate of total bycatch. However, a large study of bycatch on Taiwanese tuna longliners which covered much of the Indian Ocean found that bycatch and discards comprised 35.8% of total catch [13]. If this were representative of all Indian Ocean tuna longline fisheries, then it would suggest bycatch of the order of 145,000 t in 2015.

While there is much uncertainty in these bycatch estimates, it is clear that other tuna fisheries, particularly gillnet and longline fisheries, have much higher bycatches than the tuna pole-and-line fishery. And for all of these other fisheries, there is also the problem of ghost fishing gear: lost or discarded pieces of fishing gear which may continue to catch a wide variety of species for months or years [74, 75]. Within the Indian Ocean, ghost gillnets and purse seine FADs have been implicated in the mortality of particularly large numbers of silky sharks and olive ridley turtles, Lepidochelys olivacea [75, 76]. Such mortalities should be considered as part of the bycatch of the gillnet and purse seine fisheries. There is very little gear loss and effectively no ghost fishing from the pole-and-line fishery.

One of the more striking aspects of the pole-and-line bycatch recorded here is not what was included in the bycatch, but what was not present. While it is not surprising that whales, dolphins or turtles were not caught, the almost complete absence of discarded tuna is of note. In this study, only one skipjack and one yellowfin tuna were discarded, both depredated by sharks. Also during the course of this study, in interviews with 10 fishermen from 10 different boats, fishermen reported that they never discarded tunas because of size or damage; any tuna that could not be sold to processing plants would be retained for sale on local islands or for household consumption. In contrast, discards of small tunas have been a regular feature of western Indian Ocean purse seine fisheries. Thus, for the Soviet/Russian purse seine fishery during 1986–92, small tuna discards were estimated at about 8.3 t per 1000 t of tuna retained [10]. And for the European purse seine fishery, it was reported that tuna discards accounted for 54% of all bycatch [9]. However, IOTC [77] has recently banned discarding of skipjack, yellowfin and bigeye tunas (although not frigate tuna or other species) by purse seiners operating in its area of competence (IOTC Resolution 15/06), a move which should reduce tuna discards.

During this study, much of the non-retained bycatch was released alive, unlike many other fisheries in which most fish arrive on deck in poor condition or already dead (e.g. trawl, gillnet, some purse seine). In pole-and-line fishing, fish are brought on deck one at a time, and the flick-off de-hooking technique has very high survival rates [43, 78]. In the Maldivian pole-and-line fishery, the combination of barbless hooks, flick-off de-hooking and short time on deck is believed to promote similarly high survivorship. In comparison, longline tuna fishing can have high mortality rates of discarded fishes: 38–63% discarded dead for Taiwanese tuna longline fisheries [13], and 59% discarded dead for western-central Pacific longline fisheries [44]. In one study for Indian Ocean purse seine fishery, over 50% of the silky sharks brought on board in purse seine operations and released alive subsequently died, raising the total mortality rate of brailed sharks to 85% [79].

Seabird catches are typically minimal in pole-and-line fisheries, although 3 were caught during the course of this study. A review of Pacific pole-and-line bycatch found no records of seabird captures [43]. In a study of the Brazilian skipjack pole-and-line fishery, no seabird captures were reported, although fishermen did deliberately hit seabirds to scare them away [80]. In contrast, in the southern Indian Ocean, tuna longline fisheries have been a major source of mortality for several species of seabirds, particularly petrels and albatrosses (e.g. [5, 13]).

FADs & unobserved mortalities

As noted above, the Maldivian pole-and-line fishery makes full use of the network of approximately 50 aFADs deployed around the atolls [81, 82]. This network of aFADs was developed by the Government of Maldives to provide an opportunity for fishermen to locate and catch tuna with a relatively high likelihood of success, and it has worked, with roughly one third to one half of the total tuna catch now coming from aFADs ([39], present study).

Most FADs (both anchored and drifting) are deployed with swathes of netting underneath, which are believed to increase their aggregating power, but can lead to animal entanglement and death. In the western Indian Ocean purse seine fisheries, very high numbers of dFADs are employed: more than 7,600 dFADs were estimated to have been deployed in the Indian Ocean in 2009 [83], and 6,700 active dFADs were deployed by the French and Spanish purse seine fleets in the western Indian Ocean at the end of 2013 [84].

It has been estimated that roughly 480,000 to 960,000 silky sharks may be entangled and die annually in the netting of dFADs deployed by the purse seine fleet, a mortality 5 to 10 times higher than the reported catch [76]. There is also a problem for turtles, particularly olive ridley turtles (Lepidochelys olivacea), which frequently become entangled in FAD netting [75, 85]. Comparisons between non-entangling and entangling FADs suggest equal tuna attraction, while reducing tangling mortality of other marine life [86, 87]. IOTC has adopted a resolution to phase out entangling FADs, but progress appears to be slow.

The number of Maldivian aFADs is less than 1% of the number of dFADs deployed by the western Indian Ocean purse seine fisheries. The ecological impacts of these two types of FAD must therefore be very different, especially as aFADs tend to have smaller net panels than dFADs [88]. Nevertheless, there is scope for improvement. Maldivian aFADs are deployed with netting attached underneath, and Maldivian fishermen do sometime add extra lengths of netting to aFADs (pers. obs). While no entanglements were observed during this study, there are infrequent anecdotal reports of entangled turtles. Maldives is moving towards net-free aFADs, as recommended by IOTC (Resolution 15/08).