STANDARDIZED CPUE OF SOUTH ATLANTIC ALBACORE (THUNNUS ALALUNGA) BASED ON TAIWANESE LONGLINE CATCH AND EFFORT STATISTICS DATING FROM 1967 TO 2012

SCRS/13/7 Collect. Vol. Sci. Pap. ICCAT, 7(3): 1234-1246 (14) STANDARDIZED CPUE OF SOUTH ATLANTIC ALBACORE (THUNNUS ALALUNGA) BASED ON TAIWANESE LONGLINE CATCH AND EFFORT STATISTICS DATING FROM 1967 TO 12 Feng-Chen Chang 1 and Shean-Ya Yeh 2 SUMMARY The standardized abundance index of South Atlantic albacore, dating from 1967 to 12, based on Taiwanese longline catch and effort statistics using the Generalized Liner Model (GLM) procedure, was carried out in the present study. Factors such as year, quarter, subarea, bycatch effects of bigeye tuna, yellowfin tuna, and swordfish were used to obtain the yearly standardized CPUE trend from 1967 to 12. The quarterly standardized CPUE series from the third quarter of 1967 to the fourth quarter of 12 were also obtained by using quarter-series, subarea, bycatch effects of bigeye tuna, yellowfin tuna, and swordfish as factors of concern. CPUE trends, both yearly and quarterly, thus obtained indicated that the abundance in number of South Atlantic albacore declined from the late 19s to 199, then increased until the mid-199s, and leveled off since the early s up to 12. The quarterly trend, as compared to its respective yearly trend, often showed a significant peak per year implying a consistent recruitment pattern of this resource. RÉSUMÉ La présente étude a estimé un indice d abondance standardisé du germon de l Atlantique Sud, correspondant à la période 1967-12, en se fondant sur les statistiques de prise et d effort des palangriers du Taipei chinois, à l aide d une procédure de modèle linéaire généralisé (GLM). Les facteurs, tels que année, trimestre, sous-zone, effets de la prise accessoire du thon obèse, de l albacore et de l espadon ont été utilisés en vue d obtenir la tendance de la CPUE standardisée annuelle de 1967 à 12. On a également obtenu la série de CPUE standardisée trimestrielle à partir du troisième trimestre de 1967 jusqu au quatrième trimestre de 12, en utilisant des séries trimestrielles, sous-zone et effets de la prise accessoire du thon obèse, de l albacore et de l espadon comme facteurs de préoccupation. Les tendances de la CPUE, à la fois annuelles et trimestrielles, ainsi obtenues ont indiqué que l abondance en nombre de germons de l Atlantique Sud avait chuté à partir de la fin des années jusqu en 199, puis avait augmenté jusqu au milieu des années 9, pour se stabiliser à partir du début des années jusqu en 12. La tendance trimestrielle, comparée à sa tendance annuelle respective, a souvent fait apparaître un pic considérable par an, ce qui implique un schéma de recrutement constant de cette ressource. RESUMEN En este estudio se estimó el índice de abundancia estandarizado de atún blanco del sur, para el periodo 1967-12, basándose en las estadísticas de captura y esfuerzo del palangre de Taipei Chino, utilizando un modelo lineal generalizado (GLM). Factores como el año, trimestre, subárea, efectos de captura fortuita de patudo, rabil y pez espada fueron utilizados para obtener la tendencia de la CPUE estandarizada anual desde 1967 hasta 12. Se obtuvieron también la series de CPUE estandarizadas trimestralmente desde el tercer trimestre de 1967 hasta el cuarto trimestre de 12 utilizando series trimestrales, subáreas y efectos de captura fortuita de patudo, rabil y pez espada como factores de inquietud. Las tendencias de CPUE estandarizadas, tanto anual como trimestralmente, obtenidas así indicaban que la abundancia en número de atún blanco del sur descendió desde finales de los sesenta hasta 199, se incrementó hasta mediados de los noventa, y se estabilizó desde principios de los años hasta 12. La tendencia trimestral, cuando se compara con su tendencia respectiva anual, mostraba a menudo un pico importante por año, lo que implica un patrón de reclutamiento coherente de este recurso. KEYWORDS South Atlantic albacore, GLM, fisheries management 1 Overseas Fisheries Development Council, Taipei 6, Taiwan; E-mail: fengchen@ofdc.org.tw ; d93248@ntu.edu.tw 2 Institute of Oceanography, National Taiwan University, Taipei 6, Taiwan; E-mail: sheanya@ntu.edu.tw 1234

Introduction In the Atlantic Ocean, two stocks of albacore, separated by 5 N latitude, were assumed for the fishery management. Taiwanese longline fishery, followed Japanese footstep, has become one of the major fishing fleets utilizing this resource since 19s. According to the ICCAT report, annual catch of South Atlantic albacore ranged from 25, mt to 35, mt in the last decade. Taiwanese catch of South Atlantic albacore comprised of 7% or more of the total. As one of the fishing nations that utilized this resource, it is equally our responsibility to acquire the catch and effort statistics for the purpose of monitoring its status. Taiwanese longliners in the Atlantic composed mainly of two types of fishing gears, i.e., regular longliner and deep longliner. The regular longliner, which commenced since 19s and is also called traditional longliner, is mainly targeting on albacore. Since mid-198s, another type of lonliner or so called deep longliner, which equipped with 7 degree centigrade or more freezing capability, emerged and mainly targeting on bigeye and yellowfin tunas. Unfortunately, it was not until mid-199s when the logbook reporting system was able to distinguish their major identity by the addition of the number of hooks per basket used in new reporting logbook. Nevertheless, historic task2 data series compiled by Taiwanese fisheries managerial sectors and reported to the ICCAT since late 19s thus became one of the important data sources to investigate the long-term abundance fluctuation of this resource. The main purposes of this study were thus to standardize the South Atlantic albacore abundance indices, based on Taiwanese 1967-12 task2 data series, by using Generalized Linear Models with identifiable factors as year, quarter, fishing locations, bycatch information for the purpose of minimizing the aforementioned incompatibility may have aroused in the data set, which were collected over a rather vast area-time-fishery spectra. Materials and Methods The task2 data, aggregated by month and by 5 statistical block from 1967 to 12, were compiled and provided by Overseas Fisheries Development Council of Taiwan. Nominal CPUE was defined as catch in number per 1, hooks. Although the Atlantic water mass is generally considered having the North Atlantic mid-ocean gyre and South Atlantic mid-ocean gyre, the delineation of North Atlantic albacore from South Atlantic albacore was set at 5 N latitude. Furthermore, the habitat of South Atlantic albacore is currently designated and separated from the Indian Ocean by the E longitude. As of the entire habitat for South Atlantic albacore, it is thus designated currently as from 5 N southward and set E as its eastward boundary condition. GLM with normal error structure (Robson, 1966; Gavaris, 198; Kimura, 1981) was used in present study to standardize yearly and quarterly CPUE series of the South Atlantic albacore. Factors used in the yearly standardization are year, quarter, subarea, effects of bycatch, which includes bigeye tuna, yellowfin tuna and swordfish. Factors used in the quarterly standardization, however, are quarter-series, subarea, effects of bycatch, which includes bigeye tuna, yellowfin tuna and swordfish. The nominal CPUE values of those bycatch species were calculated and coded by quantile. GLM models constructed in present study for yearly and quarterly standardizations are as follows: Yearly generalized linear model with normal error structure: LOG(CPUE ijklmn +const)=μ+year i +QUARTER j +SUBAREA k +CODEBET l +CODEYFT m +CODESWO n +ξ i jklmn where LOG: natural logarithm; CPUE ijklmn : nominal albacore CPUE (catch in number per hooks) in year i, quarter j, subarea k, and bycatch of BET l, YFT m, SWO n, μ: intercept, or overall mean for correction; const: constant (% of the overall mean albacore nominal CPUE); YEAR i : main effect of year i; QUARTER j : effect of quarter j; SUBAREA k : effect of subarea k; CODEBET l : effect of bycatch (bigeye tuna); 1235

CODEYFT m : effect of bycatch (yellowfin tuna); CODESWO n : effect of bycatch (swordfish); ξ ijklmn : error with distribution character of N(,σ 2 ). Quarterly generalized linear model with normal error structure: LOG(CPUE iklmn +const)=μ+quarter-series i +SUBAREA k +CODEBET l +CODEYFT m +CODESWO n +ξ iklmn where LOG: natural logarithm; CPUE iklmn : nominal albacore CPUE (catch in number per hooks) in quarter-series i, subarea k, and bycatch of BET l, YFT m, SWO n, μ: intercept, or overall mean for correction; const: constant (% of the overall mean albacore nominal CPUE); QUARTER-SERIES i : main effect of quarter-series i; SUBAREA k : effect of subarea k; CODEBET l : effect of bycatch (bigeye tuna); CODEYFT m : effect of bycatch (yellowfin tuna); CODESWO n : effect of bycatch (swordfish); ξiklmn: error with distribution character of N(,σ 2 ). SAS Ver. 9.3. statistical package was used in both cases to obtain solutions. Results and Discussion A constant 1.628251, which was obtained by averaging all Taiwanese longliners nominal albacore CPUE reported from 1967 to 12 in the South Atlantic and divided by, was determined and added to each nominal albacore CPUE before using SAS solver for the purpose of avoiding zero albacore catch rate problem (ICCAT, 1996). The bycatch of bigeye tuna, yellowfin tuna and swordfish was included and coded by quantile. The four quantile intervals of nominal CPUE were (1) ~.1833,.1833~1.5425, 1.5425~ 3.625526, and greater than 3.625526 for bigeye tuna; (2), greater than to.147184,.147184~.656347, and greater than.656347 for yellowfin tuna; and (3), greater than to.728725,.728725~.28491, and greater than.28491 for swordfish. For elucidating geographical distribution characters of South Atlantic albacore resource, an aggregated (from 1967 to 12) geographic distribution map of nominal albacore CPUE in number was shown in Figure 1. As shown in Figure 1, significant area aggregation with different level of catch rate was observed. In particular, an aggregation with higher catch rate appeared between S and 45 S of the South Atlantic. The same pattern was also observed in Figure 2, which was obtained exactly the same procedure used to obtain Figure 1 but using nominal albacore CPUE in weight instead of number. To divide appropriately the South Atlantic albacore s entire habitat into subareas was one of the attempts used in present study for providing corrections stemmed from area contrast. Three subareas, as shown in Figure 3, were thus used in present study followed the resultant obtained by Wu and Yeh (2). The ANOVA tables, as shown in Table 1 and 2, which were provided by SAS solver, indicated that (1) factors assigned both in yearly model and in quarter-series model were statistically significant; (2) factor subarea played an important role in explanation of its orthogonal variation to the total; (3) comparatively, factor bycatch of yellowfin tuna played a less significant role as its mean square was relatively low, although still significant; (4) the determination coefficient R-square approached 8% in both cases indicated the explanatory resultant by the two models were quite significant. The yearly nominal CPUE trend and its respective yearly standardized CPUE series thus obtained were tabulated in Table 3, and plotted in Figure 4. The yearly standardized CPUE series showed a continuous decline from the beginning of the Taiwanese longline fishery to 199, then increased till mid-199s, and leveled off since early s up to 12. The normalized residual pattern from this model was shown in Figure 5. As shown in Figure 5, main distribution of residuals ranged from 1.65 to +1.65 and obviously centered at zero as mode. The Q-Q plot of those residuals was also shown in Figure 6 indicating the fitting was not far from normal distribution. 1236

The quarterly nominal CPUE trend and its respective quarterly standardized CPUE series thus obtained were tabulated in Table 4, and plotted in Figure 7. The quarterly standardized CPUE series showed a continuous decline from late 19s to 199 with higher fluctuation, then increased till mid-199s, and leveled off since early s up to 12. The general trend appeared in quarterly CPUE series was very similar with those obtained in yearly CPUE trend. Although quarterly trend having more fluctuations, it was very interesting to point out that every four quarters always appeared a high peak strongly implied that a consistent recruitment may have coming in every year. The normalized residual pattern from this model was shown in Figure 8. As shown in Figure 8, main distribution of residuals also ranged from 1.65 to +1.65 and obviously centered at zero as mode. The Q-Q plot of those residuals was shown in Figure 9 indicating the fitting was not far from normal distribution. Fishing intention maybe well acknowledged through notification on number of hooks per basket. It is very unfortunate that the information on noting of using number of hooks per basket only available since 1995, when a new format of including number of hooks per basket was established and delivered for Taiwanese longliners. Logbooks recovered in the period of late 19s to mid-199s, in particular, perhaps be entangled with mixed fishing intentions yet not able to clarify its identity only through area-time factors thus may produce a biased CPUE trends. Efforts will be devoted to obtain suitable discriminant functions obtained from known fishing intention data set (1995 upward data set) and extrapolating into former entangled period. We hope, through such manipulations, will give a more persuasive resultant CPUE trend than current endeavours. Acknowledgments We are grateful for the Fisheries Agency and Overseas Fisheries Development Council of Taiwanese fisheries managerial sectors for their financial supports and tremendous efforts devoted to fisheries catch and effort statistical data collection and compilation. References Chang, F. C. and Yeh, S. Y. 11, Standardized CPUE of South Atlantic albacore (Thunnus alalunga) based on Taiwanese longline catch and effort statistics dating from 1967 to. Collect. Vol. Sci. Pap. ICCAT, 68(2): 58-592. Collette, B. B. and Nauen, C. E. 1983, An annotated and illustrated catalogue of tunas, mackerels, bonitos and related species known to date. In FAO Species catalogue, (FAO Fisheries Synopsised). Rome, Italy, 125(2). Gavaris, S. 198, Use of a multiplicative model to estimate catch rate and effort from commercial data. Can. J. Fish. Aquat. Sci., 37: 2272-2275. González-Garcés, A. and Fariña-Pérez, A. C. 1983, Determining age of young albacore, Thunnus alalunga, using dorsal spines. U.S. Dep. Comer., NOAA Tech. Rep. NMFS 8: 117-122. Hsu, C. C. 1996, Standardized catch per unit effort of Taiwanese longline fishery as abundance index of albacore stock in the Atlantic. Col. Vol. Sci. Pap. ICCAT, 46 (3): 123-1. ICCAT. 1996, Report of the bluefin tuna methodology session. In April 16 to 19, 1996. Madrid, Spain, 28pp. Kimura, D. K. 1981, Standardized measures of relative abundance based on modeling log(cpue), and their application to Pacific Ocean perch (Sebastes alutus), J. Cons. Int. Explore, Mer., 39:211-218. Koto, T. 1969, Studies on the albacore-xiv. Distribution and movement of the albacore in the Indian and Atlantic Oceans based on the catch statistics of Japanese tuna longline fishery in 1959-1967. Bull. Far Seas Fish. Res. Lab., 1: 115-129. Laevastu, T. and Horacio Rosa Jr. 1963, Distribution and relative abundance of tunas in relation to their environment. FAO, Fish. Rep., 6(3): 1835-1851. 1237

Lebeau, A. 1971, Etude de la biologie du germon de l Ocean Indien. Science et Peche, Bull. Inst. Peches Marit, 4:13pp. Morita, S. 1977, Estimated age composition of albacore harvests by Japanese and Taiwanese longline fisheries in the Atlantic Ocean. Col. Vol. Sci. Pap. ICCAT, 6 (2): 19-194. Nakano, H. 1996, Review of data collection system for the Japanese longline fishery and problems about standardization of CPUE. Col. Vol. Sci. Pap. ICCAT, 43 (3): 159-161. Nelder, J. A. and Wedderburn, R. W. M. 1972, Generalized Linear Models. J. Royal Stat. Soc. A, 135:37-384. Nishida, T., Ishizuka, Y. and Kishino, H. 1994, Estimate of catch rate of Japanese longline fisheries for southern bluefin tuna (Thunnus maccoyii). The trilateral meeting for southern bluefin tuna. SBFWS/1994/11. Robson, D. S. 1996, Estimation of the relative fishing power of individual ships. ICNAF Res. Bull., 3:5-15. Shannon, L. V. 1986, The tunas of the Benguela region off southern Africa- a synthesis. ICCAT Col. Vol. Sci. Pap., 25: 566-581. Stéquert, B. and Marsac, F. 1989, Tropical tuna- surface fisheries in the Indian Ocean. FAO Fish. Rep., 282: 13-17, 64-75. Uosaki, K. 1996, Updated standardized CPUE for albacore caught by Japanese longline fishery in the Atlantic Ocean. Col. Vol. Sci. Pap. ICCAT, 43: 311-317. Uosaki, K. 1999, Updated standardized CPUE for albacore caught by Japanese fishery in the Atlantic Ocean, 1975-1997. Col. Vol. Sci. Pap. ICCAT, 49 (3): 212-217. Uosaki, K., Updated standardized CPUE for albacore caught by Japanese fishery in the Atlantic Ocean, 1975-1999. Col. Vol. Sci. Pap. ICCAT, 52 (4): 1468-1474. Uozumi, Y. 1997, Standardization of biomass CPUE for swordfish caught by Japanese longline fishery in the south Atlantic. Col. Vol. Sci. Pap. ICCAT, 46 (3): 369-372. Wu, C. L. and Yeh, S. Y. 1999, CPUE standardization for South Atlantic albacore caught by Taiwanese longline fisheries, 1968-1996. Col. Vol. Sci. Pap. ICCAT, 49 (4): 191-199. Wu, C. L. and Yeh, S. Y., Demarcation of operating areas and fishing strategies for Taiwanese longline fisheries in South Atlantic Ocean. Col. Vol. Sci. Pap. ICCAT, 52 (5): 1933-1947. Wu, C. L. and Yeh, S. Y. 2, Geographic distribution and area demarcation on the fisheries resource of south Atlantic albacore. ACTA Oceanogra. Taiwan. (1):81-92. 1238

Table 1. Analysis of variance on standardizing South Atlantic albacore yearly CPUE using Taiwanese longline fishery data set from 1967 to 12 by GLM procedure. Dependent Variable: Logcpuen_alb Source DF Sum of Squares Mean Square F Value Pr > F Model 59 21287.3686 3.8286 69.62 <.1 Error 15535 52.24633.33732 Corrected Total 15594 26527.61492 R-Square Coeff Var Root MSE Logcpuen_alb Mean.82461 26.27573.58791 2.2372 Source DF Type III SS Mean Square F Value Pr > F year 45 1146.318771 25.47375 75.52 <.1 quarter 3 76.4887 25.493362 75.58 <.1 subarea 2 5163.494716 2581.747358 7653.73 <.1 codebet 3 484.365397 161.455132 478.64 <.1 codeyft 3.744756 3.581585.62 <.1 codeswo 3 132.38221 44.12674 1.82 <.1 Table 2. Analysis of variance of standardized South Atlantic albacore quarterly CPUE using Taiwanese longline fishery data set from 1967 to 12 by GLM procedure. Dependent Variable: Logcpuen_alb Source DF Sum of Squares Mean Square F Value Pr > F Model 192 21496.36995 111.926 342.74 <.1 Error 152 31.24497.32666 Corrected Total 15594 26527.61492 R-Square Coeff Var Root MSE Logcpuen_alb Mean.8339 25.85734.571543 2.2372 Source DF Type III SS Mean Square F Value Pr > F yq 181 1384.479768 7.64959 23.42 <.1 subarea 2 4829.754 2414.73527 7392.6 <.1 codebet 3 485.718334 161.96111 495.64 <.1 codeyft 3.981353 3.6451 11.21 <.1 codeswo 3 129.194181 43.64727 131.83 <.1 1239

Table 3. Yearly nominal and standardized CPUE trends of South Atlantic albacore based on Taiwanese longline fishery data set from 1967-12 using GLM procedure. Year Nominal CPUE Standardized CPUE 1967 37.764 18.16668 1968 36.466 18.66767 1969 31.924 19.88498 197 32. 14.9796 1971 34.789 15.12431 1972 29.418.2 1973 26.566 9.3719 1974 27.813.24663 1975 33.325 12.2838 1976 29.341 12.718 1977 34.645 13.8647 1978 35.617 12.2935 1979 31.725 11.973 198 32.527.463 1981 28.1 8.36113 1982 29.1 8.324 1983 28.581 8.16678 1984 29.374 9.18769 1985.125 8.6812 1986 32.13 8.54178 1987 24.444 7.62685 1988 19.746 5.4884 1989 16.974 4.88248 199 17.527 5.21856 1991 13.382 5.86999 1992 15.975 6.972 1993 15.473 5.97513 1994 19.568 7.59922 1995 14.852 7.57848 1996 14.149 8.5729 1997 14.283 7.62 1998 9.532 6.58646 1999 9.69 5.51952 8.964 5.97 1 11.746 6.17374 2 8.378 4.98416 3 8.9 4.667 4 7.896 5.67919 5 9.359 6.57294 6 18.194 5.1853 7 15.794 5.71563 8 16.641 5.936 9 11.573 5.7728 11.665 6.54867 11 11.2896 5.8758 12 9.3552 5.27636 12

Table 4. Quarterly nominal and standardized CPUE trends of South Atlantic albacore based on Taiwanese longline fishery data set from 1967-12 by GLM procedure. Year*Quarter Nominal Standardized CPUE 19821 28.8961.1227384 19971.58 7.6687738 19673 67.412 29.9882 19822.832.754628 19972 22.3716 7.762918755 19674 11.266 13.23453 19823 25.594 8.1219638 19973 14.5625 8.1172848 19681 4.489 12.37339 19824 14.2398 4.965875443 19974 9.39 6.9693368 19682 7.526 27.5769 19831 24.44 6.963347353 19981 5.6854 6.681223252 19683 61.763 37.698 19832 37.9 9.85379 19982 16.6399 7.3593891 19684 8.249 11.48727 19833.9674 9.366291134 19983 14.2336 7.62787 19691 17.98 15.714 19834 18.5895 8.4156167 19984 3.585 5.428413151 19692 35.11 15.71792 19841 25.2644 9.3165365 19991 5.82 4.9112832 19693 47.495 22.45939 19842 49.8414 14.14152994 19992 14.939 5.315277467 19694 31.342 24.89749 19843 29.876 8.8752584 19993 11.2682 6.14563 1971.79 14.7249 19844 18.741 6.934788253 19994 5.3171 5.673869825 1972 48.468 14.49348 19851 28.89 8.145553729 1 5.8366 4.855991157 1973 46.473 19.86624 19852 39.6748 11.927122 2 13.1985 5.32226176 1974.592 11.81 19853 31.72 8.764467326 3 9.95 5.641765 19711 13.561 12.65159 19854 15.5784 6.9895536 4 7.2214 5.3714624 19712 46.982 17.742 19861 31.912 8.534839714 11 9.1858 6.8738 19713 61.3 23.26257 19862 41.8987.15763531 12 13.977 5.43354161 19714 17.875 11.83265 19863 29.8781 9.3633274 13 16.7965 6.664445934 19721 23.8 11.32343 19864 15.198 6.47478682 14 7.7787 5.568854347 19722 38.736 11.5893 19871 25.8926 7.52465322 21 7.1785 5.89624 19723 38.5 11.56858 19872.1874 9.1512216 22.1255 4.99526 19724 11.61 8.1 19873 21.7751 7.75513 23 9.2245 4.715964989 19731 24.393 6.91 19874 14.33 6.9968196 24 6.5289 5.1347181 19732 32.566 15.546 19881 17.2144 4.39927451 31 6.6653 5.251424396 19733 32.847 11.39362 19882 24.39 7.44982 32 13.9482 4.8198 19734 14.319 6.37635 19883 18.35 5.64914 33 7.9428 4.742396355 19741 17.342 7.63548 19884 16.5281 7.6526239 34 3.7578 3.786877768 19742 38.741 11.36554 19891 16.9126 4.751427537 41 5.11 4.5249581 19743 37.585 13.17621 19892 21.579 4.427984122 42 9.998 5.84843357 19744 15.594 9.49477 19893 19.4 6.74527 43 9.5855 5.934334961 19751 21.911 11.5977 19894.3576 4.194195359 44 6.85 6.673636512 19752 51.5 15.18443 1991 11.753 5.252552274 51 7.3785 6.697788725 19753 35.749 13.557 1992 23.9139 6.383878733 52 12.1578 6.41864853 19754 18.24 8.89282 1993 18.34 5.2756873 53 11.9479 6.717484418 19761 26.671.725 1994 11.57 3.483429678 54 6.3653 6.89321932 19762 39.89 16.42539 19911 13.89 5.846568636 61 15.12 5.39123472 19763 31.752 16.78869 19912 14.8715 4.771138979 62 26.3314 5.896926813 19764 16.61 8.93332 19913 12.672 6.61461 63 16.87 5.698951556 19771.4 12.53649 19914 12.776 6.7363818 64 8.2878 3.517583796 19772 44.278 17.6 19921 13.2255 6.353755572 71 11.2399 4.3573342 19773 35.682 15.22111 19922.9587 8.89843554 72 28.78 7.671763972 19774 21.555 11.46849 19923 16.7955 7.693412221 73 17.1 6.893888463 19781.446.76657 19924.2511 4.7827852 74 4.2 4.3475772 19782 43.472 13.69334 19931 14.6544 2.953448825 81 9.4515 5.67392 19783 36.225 12.54525 19932 23.881 9.1131233 82 22.8876 6.38595 19784 25.144 13.28335 19933 18.5135 6.85249794 83 27.172 7.193974567 19791 29.922.965 19934 11.1289 5.458829439 84 1.5723 4.865862168 19792 39.43 11.59621 19941 19.5579 8.142192 91 6.3532 5.559661171 19793.26 11.57919 19942 28.8247 8.224337278 92 16.9687 6.38685482 19794 22.781 11.76955 19943 16.3811 7.828441577 93 18.793 6.626331412 1981 33.655.338 19944 8.287 6.326617252 94 3.846 4.961659979 1982 44.678 13.55 19951.8642 7.7343683 1 7.757 5.2541553 1983 29.136 9.94978 19952 26.6187 9.34168 2 16.74 6.618194261 1984 22.174 9.8669 19953 14.6255 7.23947928 3 18.922 9.448851723 19811 28.466 7.921 19954.35 7.1229859 4 3.6727 5.5796649 19812 38.385.81199 19961 14.7431 7.943342772 111 6.779 5.156435247 19813 29.698 8.54639 19962.888 8.619887712 112 19.119 6.923636168 19814 15.693 6.93656 19963 14.3577 8.13657271 113 14.2133 6.5416 19964 7.3529 7.51937956 114 2.873 5.1647733 121.944 4.466494 122 15.2115 5.8278652 123 6.918 6.631 124.724 5.268298265 1241

9 8 7 W E 1967-12 ALB CPUE No./ Hooks >25 15-25 5-15 < 5 N S N S 9 8 7 W E Figure 1. Geographic distribution of South Atlantic albacore nominal CPUE (No./ Hooks) based on Taiwanese longline fishery data set from 1967 to 12. 9 8 7 W E 1967-12 ALB CPUE Wt./ Hooks > 37 27-37 7-27 < 7 N S N S 9 8 7 W E Figure 2. Geographic distribution of South Atlantic albacore nominal CPUE (Wt./ Hooks) based on Taiwanese longline fishery data set from 1967 to 12. 1242

Figure 3. Subarea delineation for South Atlantic albacore habitat. nominal CPUE (No./ hooks) 35 25 15 5 South Atlantic Albacore CPUE nominal standardized 21 18 15 12 9 6 standardized CPUE (No./ hooks) 3 1967 197 1973 1976 1979 1982 1985 1988 1991 1994 1997 3 6 9 12 Year Figure 4. Yearly nominal and standardized CPUEE (No./ Hooks) trends of South Atlantic albacoree based on Taiwanese longline fishery data set from 1967 to 12. 1243

Figure 5. Distribution of normalized residual obtained from yearly GLM model. Figure 6. The Q-Q plot for residualss obtained from yearly GLM model. 1244

8 South Atlantic Albacore CPUE 7 nominal CPUE (No./ hooks) nominal standardized standardized CPUE (No./ hooks) 1967.5 197 1972.5 1975 1977.5 198 1982.5 1985 1987.5 199 1992.5 1995 1997.5 2.5 5 7.5 12.5 Quarter-series Figure 7. Quarterly nominal and standardized CPUE (No./ Hooks) trends of South Atlantic albacore based on Taiwanese longline fishery data set s from 1967 to 12. Figure 8. Distribution of normalized residual obtained from quarterly GLM model. 1245

Figure 9. The Q-Q plot for residuals obtained from quarterly GLM model. 1246