中西太平洋金枪鱼延绳钓钓钩深度分布及其影响因素  

Hook depth distribution and influencing factors of tuna longline fishing in Western and Central Pacific Ocean

作　　者：刘志强[1] 郭绍健 王禹程 周成[1,2,3] 吴峰 万荣[1,2,3] LIU Zhiqiang;GUO Shaojian;WANG Yucheng;ZHOU Cheng;WU Feng;WAN Rong(College of Marine Living Resource Sciences and Management,Shanghai Ocean University,Shanghai 201306,China;National Engineering Research Center for Oceanic Fisheries,Shanghai Ocean University,Shanghai 201306,China;Key Laboratory of Sustainable Exploitation of Oceanic Fisheries Resources,Ministry of Education,Shanghai 201306,China)

机构地区：[1]上海海洋大学海洋生物资源与管理学院,上海201306 [2]国家远洋渔业工程技术研究中心,上海201306 [3]大洋渔业资源可持续开发教育部重点实验室,上海201306

出　　处：《上海海洋大学学报》2024年第4期1020-1030,共11页Journal of Shanghai Ocean University

基　　金：国家重点研发计划(2023YFD2401301,2023YFD2401305);农业农村部全球渔业资源调查监测评估(公海渔业资源综合科学调查)专项(D-8025-23-1003)。

摘　　要：为掌握中西太平洋金枪鱼延绳钓作业中钓钩深度分布及其影响因素,根据2021年8—10月随“淞航”号调查船在中西太平洋海域调查时采集的作业参数、钓钩深度和环境数据等信息,利用支持向量回归模型(Support vector machine,SVR)和广义加性模型(General additive model,GAM)分析不同作业参数和环境因素下钓钩深度的分布情况,探讨各因素对钓钩深度的影响。结果表明:(1)钓钩深度的计算值中1号钓钩深度最小,为(130.03±3.32) m;8号钓钩深度最大,为(363.25±36.52) m。实测值中1号钓钩深度最小,为(130.81±17.94) m;7号钓钩深度最大,为(329.91±54.37) m。(2)1号钓钩的SVR模型整体拟合度最小(R^(2)=0.38);6号钓钩的整体拟合度最大(R^(2)=0.77)。SVR模型训练值的均方差中6号钓钩最小,为0.138;8号钓钩最大,为0.309。测试值的均方差中4号钓钩最小,为0.086;5号钓钩最大,为0.282。(3)GAM分析显示0~64 m、128~192 m水层平均流速和投绳速度与各钓钩深度相关性最高,其次分别为64~128 m、192~256 m、256~320 m水层平均流速和船速。320~384 m水层平均流速和风速与各钓钩深度均无显著相关性。(4)SVR和GAM的钓钩深度预测值与实测值平均差异分别为1.75%和6.38%。SVR预测值与实测值随钓钩位置变化趋势一致,吻合性较好。研究结果有助于了解作业中金枪鱼延绳钓钓钩分布规律,进而为优化钓具性能提供基础参考依据。To comprehensively examine the hook depth distribution and influencing factors of tuna longline fishing in Western and Central Pacific Ocean,data collected aboard the vessel "Songhang" from August to October 2021 were analyzed.A combination of Support Vector Regression(SVR) and Generalized Additive Model(GAM) was employed to investigat the distribution of hook depths across different operational parameters and environmental variables.The objective was to elucidate the impact of each factor on hook depth.The results showed that:(1) Among the calculated values of hook depth,Hook No.1 had the smallest value at(130.03±3.32) m,while Hook No.8 had the largest value at(363.25±36.52) m.Regarding the measured values,Hook No.1 had the smallest depth at(130.81±17.94) m,whereas Hook No.7 exhibited the largest depth at(329.91±54.37) m.(2) The SVR model exhibited varying goodness of fit,with Hook No.1 displaying the lowest(R~2=0.38) and Hook No.6 demonstrating the highest(R~2=0.77) overall fit.The mean squared error(MSE) for training data was smallest for Hook No.6(0.138) and largest for Hook No.8(0.309),while for testing data,it was smallest for Hook No.4(0.086) and largest for Hook No.5(0.282).(3) GAM analysis reveals that water layer average velocities at 0-64 m and 128-192 m,as well as the casting speed,exhibit the highest correlations with hook depth across all hooks.Following these,water layer average velocities at 64-128 m,192-256 m,and 256-320 m,along with vessel speed,show significant correlations.However,water layer average velocity at 320-384 m and wind speed demonstrate no significant correlations with hook depth.(4) The average differences between predicted hook depths by SVR and GAM models and the actual measured values are 1.75% and 6.38%,respectively.The predictions from the SVR model align well with the measured values,showing consistent trends with the variation in hook positions,thus demonstrating good agreement.The results of this study will contribute to understanding the distribution pattern of tuna l

关 键 词：中西太平洋 金枪鱼延绳钓 钓钩深度 支持向量回归机模型 广义加性模型 

分 类 号：S973.3[农业科学—捕捞与储运]