大西洋黄鳍金枪鱼资源丰度不同预报模型的比较分析  

Comparision of forecast models for Atlantic yellowfin tuna abundance

作　　者：丁鹏 邹晓荣[1,2,3] 丁淑仪 白思琦 张亚文 DING Peng;ZOU Xiaorong;DING Shuyi;BAI Siqi;ZHANG Yawen(College of Marine Living Resource Sciences and Management,Shanghai Ocean University,Shanghai 201306,China;Collaborative Innovation Center for National Distant-Water Fisheries,Shanghai 201306,China;Key Laboratory of Sustainable Exploitation of Oceanic Fisheries Resources,Ministry of Education,Shanghai 201306,China;School of Education,Shangdong Women s University,Jinan 250300,China;Qingdao Jimo District Bureau of Natural Resources,Qingdao 266200,China)

机构地区：[1]上海海洋大学海洋生物与资源管理学院,上海201306 [2]远洋渔业创新中心,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 [4]山东女子学院教育学院,山东济南250300 [5]青岛市即墨区自然资源局,山东青岛266200

出　　处：《大连海洋大学学报》2024年第4期666-674,共9页Journal of Dalian Ocean University

基　　金：渔业生产数据收集项目(D-8002-12-0127-2);农业农村部远洋渔业资源调查和探捕项目(D-8002-13-8004E2)。

摘　　要：为探讨海洋环境因子对黄鳍金枪鱼(Thunnus albacares)资源丰度的影响,利用中水集团13艘延绳钓渔船2016—2019年的渔业数据,结合海表面风速(WS)、叶绿素a浓度(Chl a)、涡动能(EKE)、0~500 m水层温度(T)和盐度(S)等19个环境因子,经过相关性分析筛选出叶绿素a浓度等11个海洋环境因子;采用卷积神经网络(CNN)、BP神经网络、长短期记忆网络模型(LSTM)、双向长短期记忆网络模型(BiLSTM)、卷积神经网络结合双向长短期记忆网络模型(CNN-BiLSTM)分别对筛选的海洋环境因子和黄鳍金枪鱼资源丰度关系进行研究并对比模型的预测性能。结果表明,在筛选后的11个海洋环境因子中,100 m水层温度、200 m水层盐度、300 m水层温度和纬度相对重要性指数之和约占55%,为影响黄鳍金枪鱼单位捕捞努力量渔获量(CPUE)的关键环境因子;使用CNN-BiLSTM模型预测的均方根误差和平均绝对误差比其他4个模型低。研究表明,CNN-BiLSTM模型对黄鳍金枪鱼资源丰度有较高的预测准确度,可用于预测黄鳍金枪鱼的资源丰度。In order to explore the impact of oceanic environmental factors on the abundance of yellowfin tuna(Thunnus albacares),the relationship between oceanic environmental factors and the abundance of yellowfin tuna was investigated by convolutional neural network(CNN),backpropagation neural network(BPNN),long short-term memory network(LSTM),bidirectional long short-term memory network(BiLSTM),and CNN combined with BiLSTM(CNN-BiLSTM)models based on the fishery data from 13 longline fishing vessels of the China National Fisheries Corporation from 2016 to 2019.The oceanic environmental factors included sea surface wind speed,chlorophyll-a concentration,eddy kinetic energy,water temperature and salinity in the 0-500 m water column and 19 environmental factors including salinity were subjected to correlation analysis,resulting in the selection of eleven marine environmental factors such as chlorophyll a concentration.The results showed that the key environmental factors affecting yellowfin tuna Catch Per Unit of Effort(CPUE)were shown to be chlorophyll-a concentration,100 m water temperature,100 m water salinity,sea surface wind speed,and eddy kinetic energy,the combined relative importance index of 100 m water temperature,200 m water salinity,300m water temperature,and latitude,which accounted for approximately 55%based on an importance analysis among the selected eleven marine environmental factors.The findings indicate that CNN-BiLSTM model has high prediction accuracy for yellowfin tuna abundance and can be used to predict yellowfin tuna abundance,and that provide a new approach to predict the abundance of yellowfin tuna.

关 键 词：黄鳍金枪鱼 海洋环境因子 CNN-BiLSTM模型 共线性分析 

分 类 号：S931.9[农业科学—渔业资源]