Coral Reefs (2008) 27:85–95 DOI 10.1007/s00338-007-0293-z

REPORT

A quantitative comparison of recreational spearWshing and lineWshing on the Great Barrier Reef: implications for management of multi-sector coral reef Wsheries

A. J. Frisch · R. Baker · J-P. A. Hobbs · L. Nankervis

Received: 3 May 2007 / Accepted: 6 August 2007 / Published online: 25 August 2007 © Springer-Verlag 2007

Abstract This study compared the catch composition, equitably across both Wshing sectors. A management strategy catch per unit eVort, and incidental impacts of spearWshers of this type will simplify enforcement of Wsheries regulations and lineWshers engaged in a structured Wshing program and avoid discrimination of particular Wshers in local whereby Wshing eVort was standardized across time, space communities where both Wshing methods are socially or and skill level. It was found that (1) the catch composition culturally important. of both groups of Wshers overlapped considerably, (2) the numbers of target Wsh caught by spearWshers (156) and Keywords SpearWshing · LineWshing · Catch per unit lineWshers (168) were not signiWcantly diVerent, (3) the eVort · Selectivity · Coral trout · Bycatch mean size of target Wsh caught by spearWshers (1.95 § 0.1 kg, §SE) was signiWcantly larger than the mean size of target Wsh caught by lineWshers (1.27 § 0.06 kg), and (4) spear- Introduction Wshers retained 43% more biomass of target species than did lineWshers (304 versus 213 kg, respectively). However, OverWshing is deemed to be one of the greatest threats to lineWshers used »1 kg of bait for every 3 kg of target Wsh the future of coral reefs (Jackson et al. 2001; Bellwood that were captured. LineWshers also caught far more under- et al. 2004; Newton et al. 2007). Understanding and manag- sized, undesirable, or protected Wshes (i.e., bycatch) and ing the eVects and yields of reef Wsheries is therefore cru- caused far more pollution (i.e., lost gear) than did spearWsh- cial for conserving coral reefs, and for bringing wealth and ers. It is concluded that the overall impacts of recreational stability to the tens of millions of people who use or depend spearWshing and lineWshing on Wshery resources of the on coral reefs as a source of food or income (Pauly et al. Great Barrier Reef are broadly equivalent (per unit of Wshing 2002; Newton et al. 2007). An important step in this direc- eVort), and that management regulations should be applied tion is to understand how various Wshing methods impact upon both target and non-target organisms. Two of the most common reef-Wshing methods used in Communicated by Environment Editor R. van Woesik. the Indo-PaciWc region are lineWshing, otherwise known as W A. J. Frisch (&) · R. Baker · J-P. A. Hobbs · L. Nankervis angling, and spear shing (Hundloe 1985; Wright and Rich- School of Marine and Tropical Biology, ards 1985; Dalzell et al. 1996; Pet-Soede et al. 2001; Cin- James Cook University, Townsville, QLD 4811, Australia ner and McClanahan 2006). LineWshing typically involves e-mail: [email protected] the use of a baited steel hook attached to a weighted nylon line. SpearWshing typically involves the use of a rubber- A. J. Frisch · J-P. A. Hobbs ARC Centre of Excellence for Coral Reef Studies, propelled spear, as well as associated dive gear such as James Cook University, Townsville, QLD 4811, Australia mask, snorkel and Wns. Although SCUBA is used in a few places, the vast majority of spearWshing is performed by Present Address: breath-hold divers (Dalzell et al. 1996; Gillett and Moy R. Baker NOAA Fisheries, SEFSC Galveston Laboratory, 2006). In most places, both lineWshing and spearWshing are Galveston, TX 77551, USA conducted from small (4 to 6 m), outboard-powered boats 123 86 Coral Reefs (2008) 27:85–95 with 1–5 Wshers per boat (Hundloe 1985; Wright and Rich- prevent meaningful, quantitative comparisons of spearWsh- ards 1985; Myers 1993; Dalzell et al. 1996). ing and lineWshing using available data. Even in developing nations, most reef Wshers have the The aim of this study, therefore, was to compare the option of either spearWshing or lineWshing, since both meth- catch composition, CPUE and incidental impacts of recrea- ods are relatively simple and require little capital invest- tional spearWshing and lineWshing in the context of a coral ment (Dalzell et al. 1996). In places where management reef Wshery. To do this, small teams of spearWshers and regulations exist (e.g., Australia), spearWshing and lineWsh- lineWshers were engaged in a program of structured Wshing ing tend to be managed uniformly (i.e., the same input and on the Great Barrier Reef (GBR), Queensland, Australia. output controls), mainly because of the lack of speciWc Importantly, both groups of Wshers participated concur- information pertaining to each Wshing method. However, it rently at the same reef site, thereby avoiding any spatial or is crucial for reef Wshery managers to acquire intimate temporal biases. Also, catch and eVort data were recorded knowledge of the selectivity, catch per unit eVort (CPUE) directly (i.e., “on site”) to maximize accuracy. and incidental impacts of both spearWshing and lineWshing. This is to enable formulation and implementation of Wshing regulations that (1) ensure target species are exploited on a Materials and methods sustainable basis, (2) encourage Wshing methods that mini- mize bycatch, habitat degradation, and Wshing-associated Field surveys pollution, and (3) allow shared Wshery resources to be allo- cated equitably among diVerent user groups (King 1995; Two teams of Wshers, one consisting of two lineWshers (one Walters and Martell 2004). Given the current concern over hook and line per person) and the other consisting of two the sustainability of reef Wsheries (Pauly et al. 2002; Bell- spearWshers (one spear or speargun per person; no SCUBA) wood et al. 2004; Newton et al. 2007), there are remarkably conducted normal Wshing activities aboard two small (4– few studies on the relative impacts of spearWshing and 6 m) boats (one team per boat). To standardize Wshing eVort lineWshing with respect to coral reef species (McClanahan among teams and over time, lineWshers and spearWshers and Mangi 2004). operated concurrently during synchronized “sessions” Because spearWshers choose the Wsh they shoot, this form which were 1.5 h in duration. A total of 45 sessions were of Wshing is regarded as highly selective, both in terms of conducted during twelve one-day “trips” (3–4 sessions per species and size (Dalzell 1996; Mann et al. 1997; Harper trip) that were spread across 14 months (May 2005 to July et al. 2000). One advantage of this selectivity is that 2006). Consecutive sessions during the same trip were sep- spearWshing has minimal impact on non-target species arated by a short recess (0.25–1 h) and all sessions were (Eckersley 1997). Another noteworthy characteristic of completed between 0800 and 1700 hours. Given that some breath-hold spearWshing is that it is limited to shallow water. locations may favor one or other Wshing method (Long Hence, the proportion of target Wshes available to spearWsh- 1957; Connell and Kingsford 1998), each session was con- ers is typically less than the proportion available to lineWsh- ducted at a diVerent site (i.e., an arbitrary area of coral reef ers (Mann et al. 1997). Despite this limitation, spearWshing that was approximately 1 km2). Sites were chosen haphaz- is often perceived to be more eYcient (in terms of CPUE) ardly and both teams Wshed the same site at the same time. than lineWshing (Long 1957; Eckersley 1997; Gillett and Sites were spread across seven coral reefs in the Townsville Moy 2006), although this is yet to be substantiated. The region of the GBR (Fig. 1) where local Wsh assemblages are range of previous estimates of CPUE for spearWshing and considered to be relatively intact (PandolW et al. 2003; Wil- lineWshing overlap considerably (Hundloe 1985; Wright and kinson 2004). For a complete description of the habitat, as Richards 1985; Dalzell 1996; Dalzell et al. 1996; Harper well as estimates of the distribution and abundance of target et al. 2000; Pet-Soede et al. 2001; McClanahan and Mangi Wsh species, see Done (1982) and Newman et al. (1997). To 2004). Furthermore, these estimates vary (within method) standardize access to reef habitats at each location, Wshing by at least an order of magnitude. This variability is proba- was restricted to depths of ·15 m (approximately the maxi- bly attributable to the fact that data were collected at diVer- mum working depth of most spearWshers). Since the focus ent times and places, using diVerent types of Wshers of this study was on reef Wshing, all participants were dis- (commercial, artisanal, subsistence or recreational). Addi- couraged from targeting pelagic species such as mackerel tionally, most of the data were collected indirectly, either by (F. Scombridae) and trevally (F. Carangidae). boat-ramp surveys, which tend not to accurately quantify Wshing eVort (Wright and Richards 1985), or by question- Sampling considerations naires and personal logbooks, techniques that are often biased and unreliable (Tarrant and Manfredo 1993; King Most popular food Wshes are non-uniformly distributed 1995; Connolly and Brown 1995). Together, these problems across small spatial scales (Newman et al. 1997; Connell 123 Coral Reefs (2008) 27:85–95 87

Data collection 145o E N Britomart Rf The identity, size, number and fate of all captured Wsh were Walker Rf Trunk Rf recorded shortly after capture (n.b. a Wsh was deWned as any Bramble Rf vagile Wshery resource, including WnWsh, crustaceans and cephalopods). Where possible, Wsh were identiWed to spe- Curacoa Is. Palm Is. cies level according to Randall et al. (1990) or Jones and Morgan (1994). Some of the less abundant species were Barber Is. identiWed to the genus or family level only. Body size Cairns 20 km (§0.5 cm) was recorded as total length (TL), carapace length (CL) or mantle length (ML) for WnWsh, crustaceans Queensland and cephalopods, respectively, and the fate of each Wsh was Great Barrier Reef categorized into one of the following groups: A = kept Townsville to eat; B = kept for bait; C = released in good condition (i.e., no major injury and able to swim without diYculty); D = released or escaped in poor condition (includes any 22o S animal that sustained spear-induced injuries); E = discarded Australia dead. Fish that were temporarily hooked on a line but escaped before landing were disregarded. Incidental impacts of Wshing were also recorded. This included the amount of pollution (lost gear), the quantity of bait used, and the number of times that Wshers deployed their vessel’s anchor Fig. 1 Map of the study area. Fishing eVort was spread across seven reefs (see inset, upper right) in the Townsville region of the Great (i.e., the number of “hangs”). Barrier Reef, Australia Data analysis and Kingsford 1998), a source of variability that may con- Fish catches were compared in terms of the taxonomic found potential diVerences between Wshing methods. Thus, composition, number, length and weight of Wsh, the latter it would have been desirable to engage multiple boats per of which was estimated using length–weight conversion Wshing method. This was not possible, however, because formulae (Kulbicki et al. 2005; Froese and Pauly 2007). If only two boats were ever available at any one time. None- this information was not available (or the species could not theless, the use of only one boat per method was considered be identiWed), the conversion formula of a morphometri- unlikely to alter the project’s outcomes because (1) both cally-similar, congeneric species was used. Importantly, teams of Wshers were highly mobile and thus capable of formula substitutions were considered unlikely to inXuence moving to a new location (within each site) if local Wsh overall results because relevant Wsh were rare and consti- abundances were considered unsatisfactory, and (2) any tuted only a small proportion of the total catch. advantage or disadvantage to one or other method was The numbers of Wsh retained by each Wsher group were likely to be limited to a single session and each session con- compared using a
2 goodness-of-Wt test, while length–fre- stituted only a small proportion (2.2%) of total eVort. It quency distributions of target species were analyzed using a should also be noted that the number of Wshers per boat was Kolmogorov–Smirnov test (Zar 1999). A Student’s t test limited to two, thereby allowing estimation of intra-method was used to compare mean Wsh weights, although data were variability while minimizing the potential for bias associ- initially transformed (log10 [x + 1]) to overcome heterosce- ated with competition between Wshers in the same boat. dasticity (Zar 1999). A Student’s t test was also used to Given that catch rates are greatly inXuenced by a Wsher’s compare mean CPUEs of legal (¸minimum legal size; skill (Hundloe 1985; Lincoln-Smith et al. 1989; Mann et al. MLS) Wsh, both in terms of the number and biomass of cap- 1997), Wshing ability was standardized by selecting only tured Wsh. To calculate CPUE, the number or biomass of those people who (1) regarded themselves as competent Wsh that were captured during each ‘trip’ was divided by Wshers, (2) had several years of experience in the relevant the amount of time spent Wshing, thus giving units of either Wshing method, and (3) had been reef Wshing within the last Wsh h¡1 or kg h¡1 (per Wsher). 6 months. Also, as many diVerent Wshers as could be To determine which factor(s) had the greatest inXuence recruited were engaged in the study. Hence, there were nine on Wsh catches, least-squares classiWcation and regression diVerent Wshers per method (N = 18), and the median num- tree (CART) analysis was used to examine the CPUE of ber of trips per Wsher was two. legal Wsh (excluding bycatch) in relation to trip, Wsher and 123 88 Coral Reefs (2008) 27:85–95 method (De’ath and Fabricius 2000). This type of analysis retained) notable quantities of snappers (14%) and emperors successively “splits” the data into increasingly homogenous (F. Lethrinidae; 11%), but they did not catch any spiny lob- clusters by minimizing the residual sums of squares for sters, parrotWsh or legal-size rockcods (Fig. 2). each split, analogous to least squares regression (De’ath and Fabricius 2000). In separate analyses, the CPUE of all Size distribution legal Wsh and of legal coral trouts (Plectropomus spp.) only, were used as dependant variables, while combinations of The mean weight of Wsh caught by spearWshers trip, Wsher and (in all analyses) method were used as (1.95 § 0.1 kg, §SE) was signiWcantly greater than the explanatory variables. In each case, the “best” tree models mean weight of Wsh caught by lineWshers (1.27 § 0.06 kg, were chosen by bootstrapped cross-validation using both excluding bycatch) (t322 = 7.0, P < 0.001) (Table 2). Simi- the minimum and minimum + one standard error (1SE) larly, the total biomass of Wsh retained by spearWshers rules (Breiman et al. 1984). Coral trouts were analyzed sep- (304 kg) was greater than the total biomass of Wsh retained arately as a single group because all species of Plectropo- by lineWshers (213 kg, excluding bait). With regard to the mus are equally sought-after and many Wshers cannot dominant target species (Plectropomus spp.), the mean distinguish between them (Frisch and van Herwerden length of individuals caught by spearWshers (48.9 § 0.7 cm 2006). TL) was signiWcantly greater than the mean length of indi- W CART analyses were performed using S-PLUS 2000 viduals caught by line shers (42.9 § 0.5 cm TL) (t258 =6.9, computer software with the TREEs Plus supplement (Math- P < 0.001) (Fig. 3). Approximately 23% of Plectropomus soft, Seattle, USA). All other statistical analyses were caught by lineWshers were below the MLS, while only 3% of performed using SPSS computer software (SPSS, Chicago, Plectropomus caught by spearWshers were below the MLS. USA). For each test, the relevant assumptions were The length–frequency distributions of Plectropomus cap- checked a priori and a signiWcant diVerence was considered tured by spearWshers and lineWshers were signiWcantly to exist if P < 0.05 (Zar 1999). All data listed in the text and diVerent (Kolmogorov–Smirnov Z =3.14, P < 0.001) (Fig. 3). Wgures are the (untransformed) arithmetic mean § 1SE. This result did not change even when undersize (

(Cephalopholis cyanostigma) as well as two each of barra- (t46 = 2.16, P = 0.036). However, CART analyses revealed mundi cod (Cromileptes altivelis) and red bass (Lutjanus that CPUE varied more among trips, or among Wshers bohar), both of which are protected by law (Anon. 2003). (within trips), than among Wshing methods, regardless of Most of the bycatch was subsequently released, although which catch component (“all legal Wsh” or “legal coral trout some species were often used as bait (predominantly fusil- only”), dependent variable (Wsh h¡1 or kg h¡1), or combina- ier, Caesio cuning). Excluding bycatch and bait, the num- tion of explanatory variables (trip, Wsher, method) were bers of Wsh retained by spearWshers (156) and lineWshers used in the analyses. In all cases, two-leaf trees were the (168) did not diVer signiWcantly from a 1:1 ratio (Table 2). best models, as indicated by the minimum and minimum + For both groups of Wshers, the retained portion of the 1SE rules (Table 3). catch was dominated by coral trouts: Plectropomus leopar- dus, Plectropomus maculatus and Plectropomus laevis (Fig. 2). Incidental impacts Together, these three species comprised 62% of spearWshers’ catch and 73% of lineWshers’ catch (excluding bycatch). SpearWshers landed and retained 76% of the Wsh they Aside from coral trouts, spearWshers also caught signiWcant attempted to catch. The remaining Wsh either escaped with quantities of spiny lobsters (F. Palinuridae; 17%), rockcods spear-induced injuries (21%) or were discarded dead (3%) (F. Serranidae; 6%), parrotWsh (F. Scaridae; 6%) and snap- (Fig. 4). The latter occurred when a speared Wsh was found pers (F. Lutjanidae; 4%). In contrast, lineWshers caught (and to be

Table 1 Summary of lineWshers’ and spearWshers’ catch after 135 h of structured Wshing on the Great Barrier Reef. All Wsh sizes are listed as cm Total Length (TL), unless otherwise noted Taxon Target Minimum SpearWshing LineWshing species?a legal sizeb Number Size range Number Size range caught (mean) cm caught (mean) cm

Serranidae (rockcods and coral trout) Cephalopholis boenak No 38 0 – 5 17–27 (22) Cephalopholis cyanostigma No 38 0 – 140 16–34 (25) Cromileptes altivelis No No take 0 – 2 48–50 (49) Epinephelus caeruleopunctatus Yes 38 4 39–53 (47) 0 – Epinephelus coioides Yes 35 1 51 0 – Epinephelus fasciatus No 38 0 – 5 26–29 (28) Epinephelus fuscoguttatus Yes 50 5 59–92 (70) 2 42–42 (42) Epinephelus merra No 38 0 – 5 25–37 (31) Epinephelus ongus No 38 0 – 9 23–33 (29) Epinephelus polyphekadion Yes 50 1 47 0 – Plectropomus laevis Yes 50 7 45–65 (58) 1 49 Plectropomus leopardus Yes 38 84 36–61 (48) 154 29–62 (43) Plectropomus maculatus Yes 38 9 46–58 (50) 5 33–66 (46) Lethrinidae (emperors) Lethrinus atkinsoni Yes 25 1 44 10 29–41 (35) Lethrinus laticaudis Yes 30 0 – 1 44 Lethrinus lentjan No 25 0 – 2 30–30 (30) Lethrinus miniatus Yes 38 1 53 10 34–53 (46) UnidentiWed Lethrinidae No 25 0 – 9 31–39 (35) Lutjanidae (snappers) Aprion virescens Yes 38 0 – 1 53 Lutjanus argentimaculatus Yes 35 3 52–61 (57) 0 – Lutjanus bohar No No take 0 – 2 31–48 (40) Lutjanus carponotatus Yes 25 3 31–37 (34) 19 21–37 (31) Lutjanus fulviXamma No 25 0 – 9 25–30 (28) Lutjanus quinquelineatus No 25 0 – 4 21–23 (22) Lutjanus russelli Yes 25 1 41 7 28–35 (33) Lutjanus sebae Yes 55 1 46 8 33–50 (40) UnidentiWed Lutjanidae No 25 0 – 2 23–28 (26) Palinuridae (spiny lobsters) Panulirus longipes Yes – 4 9–10 (9)c 0– Panulirus penicillatus Yes – 2 12–13 (13)c 0– Panulirus versicolor Yes – 22 8–14 (12)c 0– Miscellaneous Caesio cuning No – 0 – 49 24–33 (27) Carangoides spp. No – 0 – 6 43–77 (59) Choerodon venustus Yes 30 2 38–52 (45) 0 – Echeneis naucrates No – 0 – 1 50 Grammatorcynus bicarinatus Yes 50 0 – 2 80–81 (81) Naso unicornis Yes 25 1 49 0 – Opistognathus sp. No – 0 – 1 36 Plectorhinchus chaetodontoides Yes 25 1 60 0 – Scarus microrhinus Yes 25 9 40–51 (45) 0 – Synodus sp. No – 0 – 1 24

123 90 Coral Reefs (2008) 27:85–95

Table 1 continued Taxon Target Minimum SpearWshing LineWshing species?a legal sizeb Number Size range Number Size range caught (mean) cm caught (mean) cm

UnidentiWed Labridae No 25 0 – 10 21–25 (24) Nebrius ferrugineus (shark) No – 0 – 1 85 Triaenodon obesus (shark) No – 0 – 1 90 Octopus sp. (octopus) No – 0 – 1 – Sepia sp. (squid) Yes – 1 15d 0– Live rock or coral No – 0 – 2 – a Target species were those that Wshers usually retained to eat b Many non-target species have a minimum legal size by default under Queensland State law (Anon. 2003) c Carapace length (cm) d Mantle length (cm)

Table 2 Catch statistics for spearWshers and lineWshers after 135 h of structured Wshing on the Great Barrier Reef SpearWshing LineWshing

Diversity of catch (no. of species) 21 ¸32a Total catch (no. of Wsh) 163b 485 Legal catchc (no. of Wsh) 156 168 Biomass of legal catchc (kg) 304 213 Mean Wsh weightc (kg) § SE 1.95 § 0.10** 1.27 § 0.06** CPUEc (Wsh h¡1) § SE 1.08 § 0.12 1.17 § 0.15 CPUEc (kg h¡1) § SE 2.22 § 0.23* 1.57 § 0.20* Bait consumption (kg pilchards) – 52.8 Bait consumption (kg reef Wsh) – 19.4 Pollution (lost gear) 1 knife, 2 spear tips, 96 hooks (plus associated lead 1 gun rubber weights and nylon line) Number of hangsd 70 155 Levels of signiWcance are denoted by asterisks (* P < 0.05; ** P <0.001) a The exact number of species was unknown because some Wsh were grouped into families or genera b Excludes organisms that escaped injured (i.e., category D) c Excludes organisms that were smaller than the minimum legal size (see Anon. 2003) or regarded as bycatch (i.e., not usually retained after cap- ture, unless used for bait) d Refers to the total number of times that Wshers deployed their vessel’s anchor bear eggs, in which case it was protected by law (Anon. During the study, lineWshers used 72.2 kg of baitWsh. 1995). LineWshers retained 35% of their catch for the pur- This consisted of 19.4 kg of reef Wsh (predominantly C. pose of eating. A further 12% were retained for bait. The cuning) and 52.8 kg of pre-purchased pilchards (Sardinops remaining Wsh were either released in good condition spp.) (Table 2). Assuming a mean pilchard weight of 45 g (47%), released in poor condition (5%), or discarded dead (Froese and Pauly 2007), the total number of Wsh used for (1%) (Fig. 4). Fish in the latter two groups generally bait (including reef Wsh) was »1,233. This equates to 2.5 suVered from hook-induced injuries and (or) barotrauma. baitWsh for each Wsh that was captured, or 7.3 baitWsh for Whilst the exact amount of collateral mortality could not be each Wsh that was kept to eat. determined (due to the prospect of delayed eVects), it was With regard to Wshing-associated pollution, spearWsh- found that at least 87 reef Wsh were incidentally killed or ers lost four items of Wshing gear, while lineWshers lost 96 substantially injured as a result of lineWshing (i.e., sum of hooks plus associated lead weights and nylon line categories B, D and E, Fig. 4). In contrast, 51 reef Wsh were (Table 2). Seventy-seven of these hooks were lost when incidentally killed or injured as a result of spearWshing (i.e., they became “snagged” on the substrate. The remaining sum of categories D and E, Fig. 4). 19 hooks were lost when “hooked” Wsh subsequently 123 Coral Reefs (2008) 27:85–95 91

125 Table 3 Summary of classiWcation and regression tree (CART) anal- Spearfishing yses of spearWshers’ and lineWshers’ catch per unit eVort (CPUE) 100 Linefishing Catch Dependant Explanatory Tree Splitb a 75 component variable variables size W c W ¡1 W W 50 All legal sh sh h Method, sher, trip 2 sher Wsh h¡1 Method, trip 2 trip

Number of fish 25 kg h¡1 Method, Wsher, trip 2 Wsher kg h¡1 Method, trip 2 trip 0 Legal coral Wsh h¡1 Method, Wsher, trip 2 Wsher us e d idae ida ous trout only Wsh h¡1 Method, trip 2 trip janidae opomus ephel t in thrinidae Scar ¡1 Lu Le W W Ep Palinur kg h Method, sher, trip 2 sher Plectr Miscellane kg h¡1 Method, trip 2 trip Taxon a The number of homogeneous clusters (“leaves”) in the best tree W W Fig. 2 The composition of spear shers’ and line shers’ catch after model 135 h of structured Wshing on the Great Barrier Reef. Data exclude b The explanatory variable that accounted for the greatest amount of organisms that were smaller than the minimum legal size (see Anon. variability among clusters 2003) or regarded as bycatch (i.e., not usually retained after capture, c unless used for bait) Excludes organisms that were smaller than the minimum legal size (see Anon. 2003) or regarded as bycatch (i.e., not usually retained after capture, unless used for bait) 35 Spearfishing d Includes all Plectropomus spp. that were larger than the minimum 30 Linefishing legal size (see Anon. 2003) 25

20 250 15 Spearfishing Linefishing 10 200 Number of fish fish of Number 5

0 150 30 33 36 39 42 45 48 51 54 57 60 63 66 100 Total length (cm) Number of fish Fig. 3 Size distribution of coral trout (Plectropomus spp.) that were 50 captured during 135 h of structured spearWshing and lineWshing on the Great Barrier Reef. The x-axis labels represent size-class midpoints. Arrows illustrate the minimum legal sizes of Plectropomus leopardus 0 (38 cm), Plectropomus maculatus (38 cm) and Plectropomus laevis ABCDE W (50 cm). Species-speci c sample sizes for P. leopardus, P. maculatus Fate category and P. laevis were 84, 9 and 7 (respectively) for spearWshing and 154, W 5 and 1 (respectively) for line shing Fig. 4 The fate of Wsh that were captured during 135 h of structured spearWshing and lineWshing on the Great Barrier Reef (A kept to eat; W B kept for bait; C released in good condition; D released or escaped broke the line. The ultimate fates of these sh were in poor condition; E discarded dead) unknown. LineWshers moved their boat more frequently during each Wshing session than did spearWshers. The mean dura- Gillett and Moy 2006). However, the eYciencies of these tion of each hang was »58 min for spearWshers and two Wshing methods have never been compared directly, »26 min for lineWshers. As such, lineWshers deployed their at least not without the potential for bias associated with anchor more than double the number of times than did boat-ramp surveys, questionnaires or logbooks (Wright and spearWshers (Table 2). Richards 1985; Tarrant and Manfredo 1993; King 1995; Connolly and Brown 1995). The results of this study thus provide the empirical data that are necessary for reef man- Discussion agers to make informed decisions regarding equitable allo- cation of shared Wshery resources among diVerent Wshing It is widely perceived that spearWshing is more eYcient sectors. than lineWshing with respect to exploiting shared Wshery The Wndings of this study do not support the perception resources (Long 1957; Eckersley 1997; Mann et al. 1997; that spearWshing is more eYcient at capturing target Wsh 123 92 Coral Reefs (2008) 27:85–95 than is lineWshing. Both groups of Wshers were found to spearing an under-size Wsh is considered to be wasteful harvest similar numbers of target Wsh, as well as exploit a (Personal communication, reef Wshers), spearWshers tend to similar biomass of Wsheries resources (i.e., when bait con- avoid Wsh that appear close the MLS (Fig. 3). LineWshers, sumption was included; see below). Additionally, because on the other hand, have the potential to accurately measure of the considerably larger bycatch, as well as the loss of a captured Wsh’s size without substantially harming it. substantial amounts of Wshing gear, it seems that lineWshing Hence, lineWshers are more likely than spearWshers to retain has the potential to inXict greater incidental impacts on smaller Wsh that approach the MLS. GBR ecosystems per unit of eVort than does spearWshing. In ecological terms, it may be beneWcial to target larger Wsh, since it increases the proportion of the population that Catch composition reproduce before recruitment to the Wshery (King 1995). However, it is also undesirable to target very large Wsh, Although the diversity of food Wshes on the GBR is very because they make a disproportionately large contribution large (Randall et al. 1990; Grant 1997), it is evident that to the reproductive output of the population (Sadovy 1996; both spearWshers and lineWshers caught and retained only a Birkeland and Dayton 2005). It is therefore pertinent to small number (20) of the available species, and that there consider the relative sizes of Wsh caught by each Wshing was a high degree of overlap with respect to the catch com- method. The mean sizes of coral trouts (Plectropomus spp.) position of both Wshing sectors (Table 1). These results sug- caught by spearWshing and lineWshing were 48.9 and gest, Wrstly, that spearWshers and lineWshers have strong 42.9 cm TL, respectively. These Wshes attain sexual matu- preferences for a select few species and secondly, that rity at »35 cm TL and normally grow to ¸70 cm TL spearWshers and lineWshers compete for many of the same (Heemstra and Randall 1993; Ferreira 1995). The mean size Wsh. In reef Wsheries elsewhere, there tends to be much less diVerence of Wsh caught by spearWshing and lineWshing overlap in species composition, and species preferences are (i.e., 6 cm TL) was therefore small relative to the potential often obscure (Dalzell et al. 1996; Harper et al. 2000; size of the Wsh. As such, the diVerence between methods McClanahan and Mangi 2004). This is probably because with regard to Wsh size is unlikely to be ecologically signiW- most popular species (e.g., groupers and other apex preda- cant. tors) are no longer abundant on many coral reefs (Roberts 1995; Jennings and Lock 1996; Sadovy et al. 2003), which Catch per unit eVort forces local Wshers to retain most (or all) of the other Wsh they catch (Dalzell et al. 1996; McClanahan and Mangi Excluding bycatch, the CPUEs of spearWshers and lineWsh- 2004; Author’s personal observation). ers were similar with respect to the number of Wsh captured, Coral trouts (Plectropomus spp.) were the principal tar- but diVerent with respect to the biomass of Wsh captured, get species of both Wsher groups, a result that reXects the since spearWshers generally caught larger Wsh than did line- fact that these Wshes are highly esteemed, relatively com- Wshers (Table 2). However, CART analyses revealed that mon, and vulnerable to both spearWshing and lineWshing the variability in CPUE among Wshing methods was small (Heemstra and Randall 1993; Sadovy et al. 2003). Species in comparison to the variability in CPUE among trips, or other than coral trouts were not generally targeted, but were among Wshers (Table 3). This suggests that the temporal captured opportunistically according to their respective and spatial aspects of Wshing eVort were more important vulnerabilities to each Wshing method. For example, spiny determinants of CPUE, measured in terms of either the lobsters were not caught by lineWshing because these number or biomass of Wsh, than which Wshing method was organisms do not take commonly used baits (Frisch 2007). employed. Whilst it is possible that some of the variability Similarly, few snappers and emperors were caught by in CPUE was a result of diVerences in skill among Wshers spearWshing because these organisms are relatively timid (Hundloe 1985; Lincoln-Smith et al. 1989; Mann et al. and diYcult to approach underwater (Personal communica- 1997) or the spatial heterogeneity of target Wsh populations tion, reef Wshers). (Newman et al. 1997; Connell and Kingsford 1998), it is unlikely that these inXuences were signiWcant given the Size distribution criteria for selecting Wshers and the structured nature of the sampling design (respectively) (see “Materials and It is evident that spearWshers generally caught larger Wsh methods”). than did lineWshers (Table 2). This pattern indicates that the An interesting (but anecdotal) observation was that the two Wshing methods have diVerent size selectivities (King CPUE (Wsh h¡1) of spearWshers was often lower than that of 1995; Dalzell 1996), although the diVerence was probably lineWshers during sessions when Wshing was generally exaggerated by minimum size limits. Because estimating regarded as “good”, but that the pattern was reversed during Wsh size can be diYcult underwater and inadvertently sessions when Wshing was generally regarded as “bad”. One 123 Coral Reefs (2008) 27:85–95 93 explanation is that spearWshers experienced “gear satura- Estimation of these parameters is therefore considered an tion” during “good” sessions. In other words, the time important topic for future research. required to shoot, retrieve and store a Wsh may have been Bait consumption represents a signiWcant collateral limiting for spearWshers when target Wsh were abundant impact of lineWshing, because small Wshes such as fusiliers (n.b., some spearWshers individually returned each captured often provide important ecosystem functions (Hobson Wsh to the boat in order to avoid attracting sharks). Another 1991; Graham et al. 2003). In general, lineWshers used 1 kg explanation, which is not mutually exclusive, is that of bait for every 3 kg of Wsh that were kept to eat. Although spearWshing was more eYcient than lineWshing when target it was mentioned earlier that spearWshers exploited a sig- Wsh were scarce, or were not concentrated in predictable niWcantly greater biomass of Wsh compared to lineWshers areas, perhaps because spearWshers actively searched for (304 versus 213 kg, respectively), the diVerence was small Wsh by swimming over a broad area, while lineWshers gen- when the biomass of bait used by lineWshers (72.2 kg) was erally selected discrete “spots” in which to deploy their added to their tally. Additionally, baitWsh are often supplied baited hooks. from geographically distant Wsheries (Western Australia, Studies of other shallow reef Wsheries across the Indo- California or India in the case of pilchards; Authors’ per- PaciWc indicate that spearWshers typically catch 0.4– sonal observation), thus extending the impacts of lineWsh- 8.5 kg h¡1, while lineWshers typically catch 0.5–5.1 kg h¡1 ing to areas that are well beyond local Wshing grounds. (Wright and Richards 1985; Dalzell et al. 1996; Pet-Soede Both spearWshers and lineWshers were responsible for et al. 2001; McClanahan and Mangi 2004). Although those some degree of pollution. In the case of spearWshers, this studies engaged artisanal Wshers and are thus not directly consisted of four pieces of lost spearWshing gear (two spear comparable to this study, it is noteworthy that the CPUEs tips, one knife, one gun rubber). However, gear of this observed here (2.22 and 1.57 kg h¡1 for spearWshing and nature is unlikely to harm reef organisms when it is lost at lineWshing, respectively) were within the range of previ- sea. In contrast, the gear lost by lineWshers (96 hooks plus ously reported values. It is also interesting that among those associated lead weights and nylon line) may potentially studies, the degree of variation within each Wshing method cause collateral mortality of other Wsh, sharks, turtles and was much greater than the degree of variation between Wsh- seabirds, either by ingestion of steel hooks or by entangle- ing methods, which is congruent with the Wndings of the ment in nylon line (Matsuoka et al. 2005). Incidentally, if present study. the rate of hooks lost during this project (i.e., 1 hook per 1.71 kg of “legal” coral trouts) is representative of the Incidental impacts entire GBR lineWshery, it is estimated that approximately one million hooks are lost annually on the GBR, since In general, spearWshing was found to be much more selec- the total catch of coral trouts is estimated to be 1805 t tive than lineWshing, both in terms of species and size. As (Williams 2002). a result, the total number of undersized, undesirable or It is well recognized that coral reefs can be adversely protected Wshes captured by spearWshers (7) was far less aVected by anchor damage, and that the amount of damage than the number captured by lineWshers (257). However, is proportional to (among other things) the number of times spearWshers’ bycatch was always released dead, while an anchor is deployed and retrieved (Davis 1977; Dinsdale lineWshers’ bycatch was generally released in good condi- and Harriott 2004). In the present study, lineWshers tion (Fig. 4). Furthermore, spearWshers injured an addi- recorded more than double the number of hangs than spear- tional 44 Wsh that escaped before capture (category D, Wshers did, suggesting that lineWshers potentially cause a Fig. 4). Whilst the fate of these Wsh is not known, it seems greater amount of anchor damage per unit of Wshing eVort. likely that a signiWcant proportion of them subsequently This diVerence is related to the fact that lineWshers must died as a result of their injuries. Similarly, it is likely that retrieve their anchor each time they wish to move to a new a signiWcant proportion of line-caught Wsh that were area, while spearWshers generally select a single anchorage released in seemingly good condition (category C, Fig. 4) and thereafter swim between areas. also died, perhaps as a result of post-release predation, In summary, spearWshers and lineWshers were found to stress, delayed barotrauma, or apparently minor mouth catch similar numbers of target Wsh. They were also found injuries (Cooke and Schramm 2007; Rudershausen et al. to exploit a similar biomass of Wshery resources (i.e., when 2007). The proportion of Wsh that suVer signiWcant eVects bait consumption was included). Together, these results or die as a result of Wshing-induced injuries contributes to suggest that spearWshers and lineWshers have equivalent the incidental impacts of spearWshing and lineWshing and access to shared Wshery resources, and that the direct thus should be integrated into any contemporary manage- impacts of each Wshing method are not vastly diVerent. ment regime. Unfortunately, post-release mortality rates It thus seems appropriate that management regulations are not available for any of the species encountered here. be implemented equitably across both sectors of the GBR 123 94 Coral Reefs (2008) 27:85–95 recreational Wshery. This will simplify enforcement of Wsh- De’ath G, Fabricius KE (2000) ClassiWcation and regression trees: a eries regulations and avoid discrimination of particular powerful yet simple technique for ecological analysis. Ecology W W 81:3178–3192 shers in local communities where both shing methods Dinsdale EA, Harriott VJ (2004) Assessing anchor damage on coral are socially or culturally important. However, this type of reefs: a case study in selection of environmental indicators. Envi- management strategy may need to be reviewed once fur- ron Manage 33:126–139 ther information is obtained about the incidental impacts of Done TJ (1982) Patterns in the distribution of coral communities W V across the central Great Barrier Reef. Coral Reefs 1:95–107 shing (e.g., post-release mortality rates and the e ects of Eckersley Y (1997) War beneath the waves: spearWshing debate sur- lost gear). Lastly, two cautions are issued with regard to faces. Geo Australasia 19:18–26 extrapolating the results of this study to other reef Wsheries. Ferreira BP (1995) Reproduction of the common coral trout Plectrop- Due to diVerences in skill, behavior and motivation of rec- omus leopardus (Serranidae: Epinephelinae) from the central and W northern Great Barrier Reef. Bull Mar Sci 56:653–669 reational versus artisanal or subsistence shers (Ruddle Frisch AJ (2007) Growth and reproduction of the painted spiny lobster 1996), the conclusions presented here may not be applica- (Panulirus versicolor) on the Great Barrier Reef (Australia). Fish ble to reef Wsheries outside of developed nations such as Res 85:61–67 Australia. Secondly, this study applies only to reef Wsheries Frisch AJ, van Herwerden L (2006) Field and experimental studies of hybridization between coral trouts, Plectropomus leopardus and where SCUBA is prohibited, because without the limita- Plectropomus maculatus (Serranidae), on the Great Barrier Reef, tions imposed by breath-hold diving, the CPUE of Australia. J Fish Biol 68:1013–1025 spearWshing is likely to be substantially greater than that Froese R, Pauly D (2007) FishBase. International centre for living W reported here. aquatic resource management, Manila (http://www. shbase.org) Gillett R, Moy W (2006) SpearWshing in the PaciWc islands: current W status and management issues. Food and Agriculture Organisa- Acknowledgments The authors are grateful for the numerous shers tion of the United Nations, Rome who participated in this study. Thanks are also due to the Queensland V W Graham NAJ, Evans RD, Russ GR (2003) The e ects of marine Department of Primary Industries and Fisheries for providing nancial reserve protection on the trophic relationships of reef Wshes on the support. James Cook University provided logistical assistance and Great Barrier Reef. Environ Conserv 30:200–208 John Frisch, Adam Smith and David Welch provided valuable com- Grant E (1997) Guide to Wshes. E.M. Grant Publishers, Scarborough ments on an earlier draft of the manuscript. Katherine Munkres assisted W W Harper DE, Bohnsack JK, Lockwood BR (2000) Recreational sheries in with preparation of the gures and Mark McCormick provided statis- Biscayne National Park, Florida, 1976–1991. Mar Fish Rev 62:8–26 tical advice. This research was conducted with permission from the Heemstra PC, Randall JE (1993) Groupers of the world. FAO Species Great Barrier Reef Marine Park Authority (Permit no. G05/15590.1). Catalogue 16. Food and Agriculture Organization of the United Nations, Rome Hobson ES (1991) Trophic relationships of Wshes specialized to feed References on zooplankters above coral reefs. In: Sale PF (ed) The ecology of Wshes on coral reefs. Academic Press, San Diego, pp 69–95 Anon Fisheries regulation (1995) Queensland Government, Brisbane, Hundloe T (1985) Fisheries of the Great Barrier Reef. Prestige Litho Australia (http://www.legislation.qld.gov.au) Printing, Brisbane Anon Fisheries management plan (2003) Queensland Government, Jackson JBC, Kirby MX, Berger WH, Bjorndal KA, Botsford LW, Brisbane, Australia (http://www.legislation.qld.gov.au) Bourque BJ, Bradbury RH, Cooke R, Erlandson J, Estes JA, Bellwood DR, Hughes TP, Folke C, Nystrom M (2004) Confronting Hughes TP, Kidwell S, Lange CB, Lenihan HS, PandolW JM, Pet- the coral reef crisis. Nature 249:827–833 erson CH, Steneck RS, Tegner MJ, Warner RR (2001) Historical Birkeland A, Dayton PK (2005) The importance in Wshery manage- overWshing and the recent collapse of coastal ecosystems. Science ment of leaving the big ones. Trends Ecol Evol 20:356–358 293:629–637 Breiman L, Friedman JH, Olshen RA, Stone CG (1984) ClassiWcation Jennings S, Lock J (1996) Population and ecosystem eVects of Wshing. and regression trees. Wadsworth International Group, Belmont In: Polunin NVC, Roberts C (eds) Reef Wsheries. Chapman and Cinner JE, McClanahan TR (2006) Socioeconomic factors that lead Hall, London, pp 193–218 to overWshing in small-scale coral reef Wsheries of Papua New Jones DS, Morgan GJ (1994) A Weld guide to crustaceans of Australian Guinea. Environ Conserv 33:73–80 waters. Reed, Chatswood Connell SD, Kingsford MJ (1998) Spatial, temporal and habitat related King M (1995) Fisheries biology, assessment and management. Black- variation in the abundance of large predatory Wsh at One Tree well Science, Oxford Reef, Australia. Coral Reefs 17:49–57 Kulbicki M, Guillemot N, Amand M (2005) A general approach to Connolly NA, Brown TL (1995) Use of angler diaries to examine length–weight relationships for New Caledonian lagoon Wshes. biases associated with 12-months recall on mail questionnaires. Cybium 29:235–252 Trans Am Fish Soc 124:413–442 Lincoln-Smith MP, Bell JD, Pollard DA, Russell BC (1989) Catch and Cooke SJ, Schramm HL (2007) Catch-and-release science and its eVort of competition spearWshermen in southeastern Australia. application to conservation and management of recreational Wsh- Fish Res 8:45–61 eries. Fish Manag Ecol 14:73–79 Long EJ (1957) Anglers versus spearmen. Sea Front 3:130–141 Dalzell P (1996) Catch rates, selectivity, and yields of reef Wshing. In: Mann BQ, Scott GM, Mann-Lang JB, Brouwer SL, Lamberth SJ, Polunin NVC, Roberts C (eds) Reef Wsheries. Chapman and Hall, Sauer WHH, Erasmus C (1997) An evaluation of participation in London, pp 161–192 and management of the South African spearWshery. S Afr J Mar Dalzell P, Adams TJH, Polunin NVC (1996) Coastal Wsheries in the Sci 18:179–193 PaciWc Islands. Oceanogr Mar Biol Annu Rev 34:395–531 Matsuoka T, Nakashima T, Nagasawa N (2005) A review of ghost Wsh- Davis GE (1977) Anchor damage to a coral reef on the coast of Florida. ing: scientiWc approaches to evaluation and solutions. Fish Sci Biol Conserv 11:29–34 71:691–702 123 Coral Reefs (2008) 27:85–95 95

McClanahan TR, Mangi SC (2004) Gear-based management of a trop- Ruddle K (1996) Geography and human ecology of reef Wsheries. In: ical artisanal Wshery based on species selectivity and capture. Fish Polunin NVC, Roberts C (eds) Reef Wsheries. Chapman and Hall, Manag Ecol 11:51–60 London, pp 137–160 Myers RF (1993) Guam’s small-boat-based Wsheries. Mar Fish Rev Rudershausen PJ, Buckel JA, Williams EH (2007) Discard composition 55:117–128 and release fate in the snapper and grouper commercial hook-and- Newman SJ, Williams AM, Russ GR (1997) Patterns of zonation of line Wshery in North Carolina, USA. Fish Manag Ecol 14:103–113 assemblages of the Lutjanidae, Lethrinidae and Serranidae (Epin- Sadovy YJ (1996) Reproduction of reef Wshery species. In: Polunin ephelinae) within and among mid-shelf and outer-shelf reefs in NVC, Roberts C (eds) Reef Wsheries. Chapman and Hall, London, the central Great Barrier Reef. Mar Freshw Res 48:119–128 pp 15–59 Newton K, Cote IM, Pilling GM, Jennings S, Dulvy NK (2007) Cur- Sadovy Y, Donaldson TJ, Graham TR, McGilvray F, Muldoon GJ, Phil- rent and future sustainability of island coral reef Wsheries. Curr lips MJ, Rimmer MA, Smith A, Yeeting B (2003) While stocks last: Biol 17:655–658 the live reef food Wsh trade. Asian Development Bank, Manila PandolW JM, Bradbury RH, Sala E, Hughes TP, Bjorndal KA, Cooke Tarrant MA, Manfredo MJ (1993) Digit preference, recall bias, and RG, McArdle D, McClenachan L, Newman MJH, Peredes G, non-response bias in self reports of angling participation. Leisure Warner RR, Jackson JBC (2003) Global trajectories of the long- Sci 15:231–238 term decline of coral reef ecosystems. Science 301:955–958 Walters CJ, Martell SJD (2004) Fisheries ecology and management. Pauly D, Christensen V, Guenette S, Pitcher TJ, Sumaila UR, Walters Princeton University Press, Princeton CJ, Watson R, Zeller D (2002) Towards sustainability in world Wilkinson C (2004) Status of coral reefs of the world. Australian Insti- Wsheries. Nature 418:689–695 tute of Marine Science, Townsville Pet-Soede C, van Densen WLT, Pet JS, Machiels MAM (2001) Impact Williams LE (2002) Queensland’s Wsheries resources: current condi- of Indonesian coral reef Wsheries on Wsh community structure and tion and recent trends 1988–2000. Department of Primary Indus- the resultant catch composition. Fish Res 51:35–51 tries, Brisbane Randall JE, Allen GR, Steene RC (1990) Fishes of the Great Barrier Wright A, Richards AH (1985) A multi-species Wshery associated with Reef and Coral Sea. Crawford House Press, Bathurst coral reefs in the Tigak Islands, Papua New Guinea. Asian Mar Roberts CM (1995) EVects of Wshing on the ecosystem structure of Biol 2:69–84 coral reefs. Conserv Biol 9:988–995 Zar JH (1999) Biostatistical Analysis. Prentice-Hall, London

123