Australian Vessel Performance in the East Coast Tuna Longline Fishery

Item Type article

Authors Campbell , H. F.; McIlgorm, A.

Download date 26/09/2021 04:40:09

Link to Item http://hdl.handle.net/1834/26465 Australian Vessel Performance in the East Coast Tuna Longline Fishery

H. F. CAMPBELL and A. MciLGORM

Introduction Japanese vessels which follow the to fish surface schools within 50 n.mi. from shore (Fig. 2). For this reason, the The Australian east coast tuna long­ stocks in both the northern and south­ line fishery is currently exploited by do­ ern zones of the fishery. The Japanese Australian fishery, unlike the Japanese mestic vessels and by Japanese vessels vessels target the full range of tuna and fishery, can be modeled as a single-spe­ under an access agreement. The Japa­ billfish species available in the Austra­ cies fishery, with catches of other spe­ nese fleet access to eastern Australian lian EEZ, the total catch being 6,068 cies regarded as by-catch. An analysis of waters is shown in Figure 1. About 100 metric tons (t) on average in the years L'le performance ofthe multispecies Japa­ Japanese vessels were involved in 1989, 1987-89. The average catch composi­ nese vessels is contained in McIlgorm although numbers have since fallen to tion by weight was: 2,589 t yellowfin (1995). around half that level. Their fleet con­ tuna, Thunnus albacares (about 43% of This paper describes the results of an sists of three groups of vessels: large the total by weight), 557 t bigeye tuna, analysis of the catching performance of (200-500 Gross Registered Ton, GRT) Thunnus o/;Jesus (9%), 1,453 t albacore, the Australian vessels in the fishery. A longliners which fish in the southern Thunnus alalunga (24%), 699 t sword­ sample of 3,860 daily observations on part of the Japanese fishery (south of fish, Xiphias gladius (11.5 %), and 770 domestic vessels engaged in the fish­ lat. 25°S) when the southern bluefin t marlin (12.5%) (148 t black marlin, ery in the period 1987-90 was used to tuna, Thunnus thynnus, fishery is not Makaira indica, 281 t blue marlin, relate tuna catch to vessel characteris­ underway; smaller (150-200 GRT) Makaira mazara, and 341 t striped mar­ tics and operations. The effect on the longliners which fish in Australia's lin, Tetrapturus audax) (McIlgorm, catch offactors such as vessel type, fish­ northern tropical and subtropical waters 1995). ing conditions (moon phase, sea surface and in the adjacent Exclusive Economic The Australian vessels operate from temperature), fishing practices (soak Zones (EEZs); and a small group of various ports along the coast. They are time, patrolling the longline), location typically smaller «15 m) than the Japa­ in terms of distance from the coast, and nese vessels, operate with lighter seasonal and annual fluctuations in monofilament gear than the Kuralon I stocks is estimated. used by the Japanese, and target stocks H. F. Campbell is Professor of Applied Econom­ Thna Production Model ics, University of Queensland, SI. Lucia, which are closer to the surface. Of the Queensland, Australia 4072. A. McIlgorm is As­ sociate Director (Fisheries), Australian Maritime approximately 100 Australian boats The basis of the analysis is the Cobb­ College, P.O. Box 986, Launceston, Tasmania, engaged in the fishery, 50% reported Douglas production function which is Australia 7250. catching only yellowfin tuna. The catch a simple economic model of production composition by weight of the Austra­ which describes the harvesting of tuna ABSTRACT-A sample ofdaily observa­ lian fleet during 1987-89 averaged 295 by an individual vessel: tions on the activities ofAustralian vessels longlining for yellowfin tuna, Thunnus t yellowfin tuna (85%), 9 t bigeye tuna albacares, during 1987-90 was analyzed, (3%), 31 t albacore (9%), 5 t swordfish using a productionjunction approach, to de­ (1 %), and 7 t striped marlin (2%) termine the effects ofvessel characteristics (McIlgorm, 1995). The overwhelming where h is daily harvest in metric tons, and operational practices and conditions. concentration of yellowfin tuna in the E is the daily amount ofeffort measured Significant differences were found between the tunafisheries in the northern and south­ Australian vessels' catch, as compared in thousands of hooks fished, x is the ern regions of the inshore yellowfin tuna with the Japanese vessels, is explained stock of tuna susceptible to the vessel's fishery in the east Australian Exclusive Eco­ by the tendency of Australian vessels fishing gear during the day, A is the nomic Zone. The type of vessel used, and catchability coefficient, and a and f3 are fishing practices such as soaktime, patrol­ constants. A special form of this model, ling the longline, and choice of surface 1 Mention of trade names or commercial firms water temperature were found to have sig­ does not imply endorsement by the National in which a = f3 = 1, was adopted by nificant effects on yellowfin tuna catch rates. Marine Fisheries Service, NOAA. Schaefer (1967) in a time-series study

57(3-4) 35 of the eastern Pacific yellowfin tuna fishery, and it has been extensively used in bioeconomic studies to relate total harvest to fishing effort and fish stock. The model has also been used in cross­ sectional studies to estimate the rela­ tionship between the catch and effort of individual vessels. Examples are the Strand et al. (1981) study of the Atlan­ • tic surf clam fishery, and the Bjorndal reat Barrier Reef Marine Park (1989) study of the North sea herring fishery. Other cross-sectional studies, such as those of Morey (1986) and Squires (1987), estimate cost functions derived from general forms of produc­ tion functions which include the Cobb­ Douglas as a special case. 25° S-- Queensland 25° S The Cobb-Douglas production func­ tion cannot be directly estimated be­ cause of the absence of observations on the stock of fish encountered by each AFZ vessel. For this reason it will not be possible to estimate the catchability New South Wales coefficient, A, or the coefficient of stock, {3. However, indirect measures of the 34°S------=..::...:...... ;"". 34°S factors influencing catchability, and of stock levels, can be generated by means of a series of dummy variables. Using these measures, together with the ob­ servations on catch and effort, the con­ stant, a, on fishing effort in the produc­ tion function can be estimated. The fac­ • Permanent closure tors affecting the tuna stock encountered by each vessel, and the vessel's catch­ ability coefficient can be divided into III Longline only the categories of vessel characteristics, fishing practices, and stock levels. Vessel Characteristics l!m Handline only Figure I. - Japanese vessel access to the east Australia Fishing Zone (AFZ) in 1989. Four types of vessels are used in the The areas south oflat. 34'S, adjacent to the Great Barrier Reef Marine Park and Box domestic longline fishery. Planing "171" have been closed to fishing since 1984. longliners are high-speed, high-horse power, low-displacement hull vessels of up to 15 m length and are commonly sels which are specifically designed for to retrieve fish and rebait hooks, or by used in Australian rock lobster fisher­ the fishery will have an advantage in moon phase. Catchability is thought to ies. Multipurpose vessels are displace­ terms ofcatchability, and that the speed rise in the darker phases of the moon ment-hull vessels under 15 m that un­ of the planing longliner and the length (DPIE2), and tidal patterns are also as­ dertake alternative fisheries such as of purpose-built vessels will provide an sociated with moon phase. The choice trapping, droplining, or potting, as well additional advantage. of where to fish in terms of distance as tuna longlining. Many trawlers con­ from the coast (0-12, 12-50,50-100, Fishing Practices vert to longline gear when the tuna or >100 n.mi.), or in terms of sea sur­ stocks are available, whereas purpose­ Several choices which are made by face temperature may affect the stock built longliners are displacement-hull vessels may affect catchability and/or level in the locality. vessels that can be up to 18 m in length the level of stock open to exploitation. and have been designed specifically for Catchability may be affected by the 2 DPIE. 1990. New South Wales logbook co­ longlining and droplining of fish for length of soak time, by intermittent pa­ ordinators report. Aust. Fish. Serv., Dep. Primary fresh markets. It is possible that the ves­ trolling of the longline during the day Ind. Energy, Canberra, Aust. Unpubl. doc.

36 Marine Fisheries Review dummy variables representing the years 1988, 1989, and 1990, respectively ------25 0 S+--~~~-+-----,.l---+---- (with the default value of the dummy representing 1987), In(E) is the natural Queensland logarithm of the number of hooks fished in the day, T is a temperature-related ",--- ... variable which measures stock avail­ ._------~ ability, and St represents soak time and U is a random error term assumed to ------30 0 S+----#-.:?'~.f_-----+_--__'l:-- Lord Howe Island have zero mean and constant variance. The Australian vessels primarily har­ New South Wales . I vest yellowfin tuna. Researchers have determined upper and lower bounds for Australian Fishing Zone water temperatures where they are .... known to range, and fishermen search \ for a level of sea surface temperature '--... \ (SST) which they believe maximizes the "- - ... _-----, chance of locating fish. In this study, " the optimum water temperature was Victoria. assumed to be 21.5°C. Ifyellowfin tuna stocks tend to be most abundant in wa­ ters of 21.5°C, then it can be assumed that the further the actual SST is from this value, the lower the stock level. A crude indicator of the stock level avail­ Figure 2. - The area of the domestic fishing effort along the New South Wales and able to a vessel is given by l/exp(yT) Southern Queensland coasts. The hatched area represents the inshore fishing grounds favored by the domestic vessels. where T is the absolute value of the dif­ ference between the actual SST and the optimal temperature. If a vessel is fish­ Stock Levels where: In(h) is the natural logarithm of ing waters at 2lSC the stock indicator daily yellowfin tuna harvest, A is a con­ takes on a value of unity; this value de­ Harvest will be determined by the level stant, a, bl' b , b , cl' c ' c ' ml' illz, m , clines exponentially as the water tem­ of fish stocks susceptible to the vessel's 2 3 2 3 3 ql' q2' q3' YI' Y2' Y3' a, y, and 0 are coef­ perature diverges from the optimum. fishing gear during the day, the latter be­ ficients to be estimated, Pt is a dummy Since the larger the value of Tthe lower ing influenced by the general stock con­ variable taking the value I if the is the stock assumed to be in the vicin­ ditions in the EEZ. Tuna and billfish are longline was patrolled and zero other­ ity of the vessel, it would be expected migratory, and stocks are known to fluc­ wise, ZI' Z2' and Z3 are dummy vari­ that the estimated value ofy, the coeffi­ tuate on a seasonal and annual basis. ables representing the 12-50, 50-100, cient on T, would be negative. All these factors are accounted for in and >100 n.mi. zones from the coast The equation is first estimated by the production analysis. In addition the (with the default value of the dummy Ordinary Least Squares (OLS) using all analysis allows for the possibility of sig­ representing the 0-12 n.mi. zone), CI' the observations in the sample. Tests nificant differences between the north­ C , 2 and C3 are dummy variables repre­ indicated the presence of hetero­ ern and southern zones of the domestic senting the vessel classes multipurpose scedasticity in the sample, and standard fishery. The northern zone is defined as vessels, trawlers, and purpose built ves­ data transformations were performed in the area north of Sydney, from lat. 34° to sels, respectively (with the default value an unsuccessful attempt to eliminate 25°S; the southern zone is the area south of the dummy representing planing this problem. In the presence of hetero­ of Sydney, from lat. 34° to 38°S (Fig 2). , longliners), MI' M 2 and M 3 are dummy scedasticity in the sample, the OLS co­ variables representing the new moon, efficient estimators are unbiased but the Estimation first quarter, and full moon respectively estimated t statistics are biased. The It is convenient to estimate the pro­ (with the default value of the dummy sample is then divided into northern and duction model in logarithmic form. The representing the last quarter), QI' Q2' southern subsamples, and a log likeli­ equation estimated is: and Q3 are dummy variables represent­ hood radio test is used to determine ing the April-June, July-September, whether the production processes in the In(h) = A + aPt + blZI + b2~ + b3Z3 + and October-December quarters, re­ north and south are the same. This test CIC 1 + C2 C2+ C3 C3+ miMI + m2M 2 + spectively (with the default value of the indicated that the northern and south­ m3M3+ qlQI + q2Q2 + q3Q3 + Y1Y1+ dummy representing the January­ ern zones should be regarded as distinct Y2 Y2 + Y3 Y3 + aIn(E) + 1F + Oln(St)+U, March quarter), Y I, Y2, and Y3 are fisheries. The northern and southern

57(3-4) 37 samples were then examined to deter­ no significant effect on yellowfin tuna dence and can be attributed to movements mine whether any groups of variables catch in the southern fishery, and these in the East Australian Current (DPIE2). could be excluded as having an insig­ variables were omitted from the final The results indicate that patrolling the nificant effect on yellowfin tuna catch. specification of the southern model. longline in the southern fishery in­ It was found that in the northern fish­ Patrolling the longline was found to creases yellowfin tuna catch by 16%, ery, yellowfin tuna catch did not vary have no significant effect on yellowfin holding all other factors constant. A 1% significantly with distance from shore, tuna catch in the northern fishery and increase in soak time was found to in­ and this set of variables was dropped this variable was also omitted. Distance crease yellowfin tuna catch by 0.2% in from the model. In the southern fish­ from shore was found to be significant the northern fishery, and 0.08% in the ery, it was found that yellowfin tuna in neither fishery, suggesting that per­ southern fishery. These results perhaps catch was not significantly affected by formance is influenced by local avail­ reflect the fact that longer soak times distance from shore, vessel class, or ability offish rather than choice offish­ (11.23 h avg.) backed up by patrolling moon phase, and these variables were ing ground. the longline are the norm in the south­ dropped from the model. The coefficients on the year dummies ern fishery, whereas the northern fish­ indicate that 1988 and 1990 were par­ ery employs shorter soak times (5.96 h Results ticularly good years for the northern avg.). When a hook encounters a tuna, Since the northern and southern fish­ fishery, whereas 1989 and 1990 were the bait is either removed or the fish eries were found to have different char­ good for the southern fishery. The fluc­ hooked. In either case, the hook ceases acteristics, results will be reported for tuations from year to year are quite sig­ to fish. Patrolling the line to remove each. Table 1 reports the coefficients nificant: for example, the 1989 yellow­ hooked fish and/or rebait will signifi­ estimated for each of the independent fin tuna vessel catch level in the south­ cantly increase catch if there is a sig­ variables in final versions of the esti­ ern fishery was more than double that nificant probability that the rebaited mating equations, together with t sta­ in the base year, holding all other fac­ hook will encounter another fish. Since tistics. Coefficient estimates which are tors constant. These annual fluctuations tuna travel in schools, the probability significantly different from zero at the are partly explained by movements of of rebaited hook encountering another 1%, 5%, and 10% levels of significance the East Australia Current. The coeffi­ fish depends on the probability of the are marked by asterisks. As mentioned cients on the seasonal dummies suggest longline encountering another school. above, these tests of significance may that there is little seasonal variation in It may be that short soaktimes used in not be reliable because of the problem yellowfin tuna catches for given levels the northern fishery make probability of of heteroscedasticity. ofeffort in the northern fishery, whereas the longline encountering a second in the southern fishery performance in school oftuna fairly low. The use of the Discussion the first quarter of the year is markedly shorter soaktimes in the north, despite the In the initial estimations it was found below that in the rest of the year. This apparent benefit of an increase in soak­ that moonphase and class of vessel had pattern is consistent with anecdotal evi- time, may be due to technical factors such as strong currents which rapidly move the line from its point of setting. Table 1.-Results of the regression for the period 1987-90.' The results suggest that in the north­ Variable North South Variable North South ern fishery purpose-built longliners and A (Constant) 0.874" 0.756' q2 (Season 2) 0.285'" 0.382' trawlers have a significant catching ad­ (2.28) (2.57) (1.72) (5.17) vantage over multipurpose vessels and a (Patrol) N/A2 0.151' q3 (Season 3) -0.235 0.365' planing longliners. For example, hold­ (3.50) (-1.36) (5.80) ing all other factors constant, a purpose­ m, (New moon) 0.107 Y, (1988) 0.464' 0.285' built longliner will have a 64% higher (1.42) N/A (5.04) (4.16) yellowfin tuna catch, and a trawler a m2 (First phase) -0.104 N/A Y2 (1989) -0.036 0.766' 33% higher yellowfin tuna catch than a (-1.36) (-0.47) (12.22) planing longliner. Against this advantage m (Full moon) -0.047 N/A 3 Y3 (1990) 0.263' 0.488' in yeUowfin tuna catch performance must (-0.62) (3.19) (709) be set any disadvantage in terms ofhigher c, (Multipurpose) 0.077 N/A Ii (InSt) 0.201' 0.08' costs of operating these vessels. However, (0.89) (4.12) (2.88) a survey by Campbell and McIlgorm c2 (Trawlers) 0.285' N/A ,,(In effort) 0.724' 0.535' (4.28) (11.65) (10.59) (1992) revealed that the cost per unit of effort for planing longliners was around c3 (Purpose built) 0.498" N/A Y (Temperature) -0.106' -0.068' (2.37) (-4.39) (-3.65) 1.8 times that of multipurpose vessels and trawlers. Cost data were not available for q, (Season 1) -0.141 0.407' Summary stalistic: R2 north = 0.267, south = 0.138. (-0.79) (7.83) Sample size (n) = 1,412 (north) and 2,448 (south). purpose-built vessels.

1 • = t significant at 1% level, ... at 5% level, ..... at 10% level (I ratios in parenthesis). SST recorded by fishermen consti­ 2 N/A = not applicable. tuted a significant variable in both fish-

38 Marine Fisheries Review eries. In the north an additional 1° more tage over the south in terms of CPUE. longline in waters close to 21.5°C re­ or less than the optimal temperature re­ The coefficient on the log of effort is sults in significantly higher catch rates. duced yellowfin tuna catch by 11 %, an estimate of the percentage increase Overall, the northern region seemed whereas in the south the estimated re­ in yellowfin tuna catch per vessel in re­ to provide significantly higher yellow­ duction is 7%. This confirms the view sponse to a 1% increase in the level of fin tuna CPUE. Furthermore, the pre­ that local movements of yellowfin tuna effort. Again the comparison of the dicted response of catch to an increase stocks are influenced by water tempera­ northern and southern fisheries favors in effort per vessel was more favorable ture variations, and that information on the north: a 1% rise in vessel effort pro­ in the north than the south. This sug­ water temperature is important in the duces a 0.72% rise in yellowfin tuna gests that consideration might be given conduct of the fishery. catch in the northern fishery and a to developing the northern fishery further. The constant term in each production 0.54% rise in the southern fishery. function represents the log of catch for The comparison of the northern and Acknowledgments base values of the dummy variables, southern fisheries suggests that a shift This study is part of a larger study without taking into account the effects of vessel effort from the south to the north funded by the Fishing Industry Re­ of soak time, water temperature, and would increase total yellowfin tuna catch. search and Development Council local stock depletion caused by the level This result may reflect the fact that the (FIRDC) Project 90/98. The results of of individual vessel effort. It can be used southern fishery is operating close to the the larger study are available in a Ph.D. to calculate a base level of yellowfin limit of the distribution of yellowfin tuna thesis by A. McIlgorm, Department of tuna catch per unit of effort (CPUE) in in the South Pacific Ocean. Economics, University of Queensland. each fishery for assigned values of the The authors wish to thank the Austra­ soak time, temperature, and effort vari­ Conclusion lian Fisheries Management Authority ables. When effort and soak time are (AFMA) and the East Coast Tuna Man­ assigned their mean values, water tem­ In summary, the analysis of the 1987­ agement Committee (ECTUNAMAC) perature assumes its optimal value, and 90 sample provided some evidence of for providing catch and effort data. all dummy variables are set equal to seasonal and annual fluctuations in per­ zero, it can be estimated that 277 hooks formance: season 2 (July-September) Literature Cited set in the northern fishery yield 201 kg was clearly best in the north, whereas Bjorndal, T. 1989. Production in a schooling fish­ of yellowfin tuna and 345 hooks in the season 4 (January-March) was clearly ery: The case of the North Sea herring fish­ southern fishery yield 59 kg. The cor­ worst in the south; 1987 and 1989 were ery. Land Econ. 65(1):49-56. responding CPUE estimates are 0.73 kg the worst years in the north, and 1987 Campbell, H. E, and A. McIlgorm. 1992. A cost and income survey of the east coast tuna per hook in the north and 0.17 in the was the worst in the south. Moon phase longline fishery. Fish. Ind. Res. Develop. south. This comparison understates the was of no significance in the south and Counc. Proj. 90/98, Final Rep., Pt. I. Aust. relative performance of the southern of little significance in the north. Mar. ColI., July, 42 p. McIlgorm, A. 1995. An economic analysis of the fishery since, as can be seen from the Some fishing practices appeared to east Australian tuna longline fishery. Dep. positive signs of the coefficients of the provide a relative yellowfin tuna catch­ Econ., Univ. Queensland, Ph.D. thesis, season and year dummies, the base pe­ ing advantage. The use of trawlers and 336 p. Morey, E. R. 1986. A gen~ralized harvest func­ riod chosen for the comparison is the purpose-built vessels provided higher tion for fishing: Allocating effort among com­ worst season in the worst year of the catch rates in the north, holding other mon property cod stocks. J. Environ. Econ. Manage. 13:30-49. sample for the southern region. How­ factors constant. Patrolling the longline Schaefer, M. B. 1967. Fishery dynamics and the ever, the comparison can also be car­ resulted in a significantly higher catch present status of the yellowfin tuna popula­ ried out for the best season in the best in the south but not in the north. Soak tion of the eastern Pacific Ocean. Int. Am. Trop. Tuna Comm. Bull. 12(3):27-56. year for each region, and allowing ves­ time is important in both fisheries, but Squires, D. 1987. Fishing effort: its testing, speci­ sels in each region to adopt those fish­ more so in the north where there appear fication and internal structure in fisheries eco­ ing practices identified by the analysis nomic and management. J. Environ. Econ. to be significant gains to longer soak Manage. 14:268-282. as advantageous. When this is done the times; these apparent gains might be Strand, I. E., J. Kirkley, and K. McConnell. 1981. CPUE is 2.53 kg in the north, and 0.64 offset by operational difficulties result­ Economic analysis and the management of Atlantic surf clams. In L. G. Anderson (Edi­ kg in the south. Thus the northern re­ ing from drifting of the longline. In both tor), Economic analysis for fisheries manage­ gion appears to have a fourfold advan­ regions of the fishery placing the ment plans, p. 113-141. Ann Arbor, Mich.

57(3-4) 39