| 基于神经网络的印度洋长鳍金枪鱼(Thunnus alalunga)时空分布与海洋环境关系研究 |
| |
| 引用本文： | 程懿麒,张俊波,汪金涛,雷林.基于神经网络的印度洋长鳍金枪鱼(Thunnus alalunga)时空分布与海洋环境关系研究[J].海洋与湖沼,2021,52(4):960-970. |
| |
| 作者姓名： | 程懿麒 张俊波 汪金涛 雷林 |
| |
| 作者单位： | 上海海洋大学海洋科学学院 上海 201306;上海海洋大学海洋科学学院 上海 201306;农业农村部大洋渔业开发重点实验室 上海 201306;国家远洋渔业工程技术研究中心 上海 201306;大洋渔业资源可持续开发教育部重点实验室 上海 201306;农业农村部大洋渔业资源环境科学观测实验站 上海 201306 |
| |
| 基金项目： | 国家自然科学基金项目，NSFC41876141号；国家重点研发计划，2019YFD0901404号；上海市科技创新行动计划，19DZ1207502号；自然资源卫星遥感技术体系建设与应用示范项目，202101004号。 |
| |
| 摘 要： | 长鳍金枪鱼(Thunnusalalunga)是主要的经济性金枪鱼鱼种之一,其空间分布与环境因子存在着密切联系。利用2012—2019年印度洋长鳍金枪鱼生产数据和海洋环境数据,包括海表面温度(sea surface temperature, SST)、叶绿素浓度(chlorophyll a, chl a)和海表面盐度(sea surface salinity, SSS)构建印度洋长鳍金枪鱼时空分布神经网络模型。以空间(经度,纬度)、环境因子(SST, chl a, SSS)为解释变量,局部渔获量为因变量,变化隐含层节点数,构建了18个BP空间分布模型,并采用10×10交叉验证模型稳定性,以均方误差(meansquareerror,MSE)、平均相对方差(averagerelativevariance,ARV)以及拟合优度(R~2)作为不同模型精度与稳定性的评判标准,最终选取5-18-1(隐含层节点18)模型为最佳模型,其平均MSE值为0.02232,平均ARV值为0.511。利用最优模型预测结果与同期实际捕捞产量进行叠加对比发现两者具有一致性。环境因子敏感性分析表明海表温度显著影响印度洋长鳍金枪鱼渔场分布,其贡献率达到0.2。印度洋长鳍金枪鱼高精度BP神经网络时空分布模型为其资源的可持续开发与动态管理提供了一种新思路。
|
| 关 键 词： | 长鳍金枪鱼 印度洋 渔场预报 BP神经网络 时空分布 |
| 收稿时间： | 2021/1/5 0:00:00 |
| 修稿时间： | 2021/3/15 0:00:00 |
STUDY ON THE RELATIONSHIP BETWEEN TEMPORAL-SPATIAL DISTRIBUTION OF INDIAN OCEAN ALBACORE(THUNNUS ALALUNGA) AND MARINE ENVIRONMENT BASED ON NEURAL NETWORK |
| |
| CHENG Yi-Qi,ZHANG Jun-Bo,WANG Jin-Tao,LEI Lin.STUDY ON THE RELATIONSHIP BETWEEN TEMPORAL-SPATIAL DISTRIBUTION OF INDIAN OCEAN ALBACORE(THUNNUS ALALUNGA) AND MARINE ENVIRONMENT BASED ON NEURAL NETWORK[J].Oceanologia Et Limnologia Sinica,2021,52(4):960-970. |
| |
| Authors: | CHENG Yi-Qi ZHANG Jun-Bo WANG Jin-Tao LEI Lin |
| |
| Institution: | College of Marine Sciences of Shanghai Ocean University Shanghai 201306, China;College of Marine Sciences of Shanghai Ocean University Shanghai 201306, China;Key Laboratory of Oceanic Fisheries Exploration, Ministry of Agriculture and Rural Affairs, 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 Ocean University, Shanghai 201306, China;Scientific Observing and Experimental Station of Oceanic Fishery Resources, Ministry of Agriculture and Rural Affairs, Shanghai 201306, China |
| |
| Abstract: | Albacore (Thunnus alalunga) is one of the main economic tuna species, and its spatial distribution is closely related to environmental factors. In this research, BP artificial neural network method was applied, with which a neural network model was established based on the production data and environmental data of T. alalunga from 2012 to 2019. The output factor was the normalized fishing production data, and the input factors were spatial factors (including longitude and latitude), marine environmental factors, including sea surface temperature (SST), chlorophyll-a concentration, (chl a), and sea surface salinity (SSS), from which 18 BP neural network models were run using 10×10 cross-validation model stability, mean square error (MSE), average relative variance (ARV), and goodness of fit (R2) as the judgement of accuracy and stability of different models. Finally, Model 5-18-1 (hidden layer node 18) was chosen as the best model; its average MSE value is 0.022 32, and average ARV value is 0.511. The optimal model prediction results and the actual catch data during the same period were superimposed and compared, showing consistent results. Meanwhile, sensitivity analysis of environmental factors showed that SST significantly affected the distribution of T. alalunga fisheries in the Indian Ocean, with a contribution rate of 0.2. The high-precision BP neural network temporal-spatial distribution model of T. alalunga provides a new idea for the sustainable development and dynamic management of the resources. |
| |
| Keywords: | Albacore tuna Indian Ocean fishing ground forecast BP neural network temporal-spatial distribution |
| 本文献已被 CNKI 等数据库收录！ |
| 点击此处可从《海洋与湖沼》浏览原始摘要信息 |
| 点击此处可从《海洋与湖沼》下载免费的PDF全文 |
|
