考慮年齡未確定性於兩階段成長模型：太平洋黑皮旗魚（Makaira nigricans）年齡與成長研究

Abstract

本研究於2016年9月至2018年1月在台灣三個魚市場採集616組背鰭切片（115 - 275公分eye fork length, EFL）合併51組耳石（17 - 172公分EFL）完成太平洋黑皮旗魚（Makaira nigricans）年齡判讀。耳石微輪紋計數日齡為213 - 484天，推測一歲平均體長為145公分EFL，並以該結果驗證硬棘切片第一年輪形成位置。雌雄魚硬棘切片讀齡平均百分誤差分別為10.77和9.23%，平均變異係數分別為15.23和13.06%。雌魚硬棘切片估計年齡範圍為1.5歲至11歲，雄魚為0.5歲至10.5歲。硬棘邊緣成長率分析得知每年8 - 9月間形成一成長輪紋。整合本研究及過去發表之年齡體長資料，以兩種成長模型 (traditional von Bertalanffy growth model，VBGM；two-stanza growth model， TSGM) 及兩種估計方法（標準非線性方法及考慮讀齡誤差方法）在模型選取準則下進行模型評估比較。結果兩種估計方法均支持使用TSGM。TSGM有較小的akaike’s information criterion 或 deviance information criterion 數值，且VBGM傾向低估最大平均體長。本研究建議TSGM結合讀齡誤差估計方法適合作為太平洋黑皮旗魚成長模型，且較具彈性以考慮判齡不確定性。 本研究建議區域漁業管理組織（Regional Fisheries Management Organization）未來在進行太平洋黑皮旗魚資源評估時應採用該成長曲線。

Age determination of the Pacific blue marlin (Makaira nigricans) was performed by combining methods of otolith of 51 young (17 – 172 cm eye fork length, EFL cm) and sectioned dorsal fin spines of 616 individuals (115 – 275, EFL cm) collected from three fish markets of Taiwan from September 2016 to January 2018. Otolith micro-increments were counted with ages of 213 – 484 days old. The estimated mean length (145 cm) at year 1 and the growth rate were used to validate the position of the first annual growth band in sectioned fin spines. The average percent error is 10.77 and 9.23% and the coefficient of variation is 15.23 and 13.06% for male and female annulus counts of fin spines, respectively. The ages of the female and males specimens were estimated to be 1.5 – 11 years and 0.5 – 10.5 years, respecitively. The marginal increment ratio suggested that the growth bands of blue marlin formed once a year (between August and September). By integrating the length-at-age data of the present study with the published data of juveniles and large individuals, two growth models (traditional von Bertalanffy growth model, VBGM; and two-stanza growth model, TSGM) by two estimation approaches (standard nonlinear approach and ageing-error approach) were evaluated based on the model selection criteria. The results support the use of the TSGM over the VBGM for both estimation approaches because the TSGM has a lower akaike’s information criterion or deviance information criterion value and the VBGM tends to underestimate the maximum averaged length. This study suggests that the TSGM coupled with the ageing-error approach appeared suitable for modeling the growth of the Pacific blue marlin with greater flexibility for incluing ageing uncertainty. This study recommend that the developed growth curves to be considered in the future stock assessment by the Regional Fisheries Management Organization.