基于圆形统计和时间序列的月相对两种拖网渔业CPUE的影响  

Effects of lunar phases on CPUEs of trawl fisheries based on circular statistics and time series

作　　者：李成 石肖飞 张健[1,2] 肖佑鹏 黄斌 石建高[5] LI Cheng;SHI Xiaofei;ZHANG Jian;XIAO Youpeng;HUANG Bin;SHI Jiangao(College of Marine Sciences,Shanghai Ocean University,Shanghai 201306,China;National Engineering Research Centre for Oceanic Fisheries,Key Laboratory of Sustainable Exploitation of Oceanic Fisheries Resources,Ministry of Education,Key Laboratory of Ocean Fisheries Development,Ministry of Agriculture and Rural Affairs,Shanghai Ocean University,Shanghai 201306,China;Shanghai Kaichuang Deep Sea Fisheries Company Limited,Shanghai 200082,China;Representive Office in Madagascar,China National Fisheries Company,Mahajanga,Madagascar 401;East China Sea Fisheries Research Institute,Chinese Academy of Fishery Sciences,Shanghai 200090,China)

机构地区：[1]上海海洋大学海洋科学学院,上海201306 [2]上海海洋大学国家远洋渔业工程技术研究中心,大洋渔业资源可持续开发教育部重点实验室,农业农村部大洋渔业开发重点实验室,上海201306 [3]上海开创远洋渔业有限公司,上海200082 [4]中国水产有限公司马达加斯加代表处,Mahajanga Madagascar 401 [5]中国水产科学研究院东海水产研究所,上海200090

出　　处：《大连海洋大学学报》2023年第2期340-347,共8页Journal of Dalian Ocean University

基　　金：农业农村部全球重要鱼种资源动态监测评估项目;国家重点研发计划项目(2020YFD0900803)。

摘　　要：为了解和掌握月相对渔业产量或单位捕捞努力量渔获量(catch per unit effort,CPUE)的影响规律,利用2016—2019年(每年3月2日—11月29日)马达加斯加西海岸底拖网独角新对虾(Metapenaeus monoceros)和2017—2020年(每年6月15日—10月9日)西白令海中层拖网狭鳕(Theragra chalcogramma)的渔业生产数据,结合基于圆形统计的广义线性模型(GLM)和基于时间序列的广义加性模型(GAM)2种不同的月相量化和统计的方法,分析月相对拖网渔业CPUE的影响。结果表明:月相对独角新对虾的CPUE具有显著性影响(P<0.05),2种方法得出的影响趋势较为一致,较高CPUE出现在上弦月;基于圆形统计的GLM显示,月相对狭鳕CPUE具有显著性影响(P<0.05),较高CPUE出现在新月期,而基于时间序列的GAM显示,月相对狭鳕CPUE的影响不显著(P>0.05);交叉验证显示,基于圆形统计的GLM平均绝对误差(E_(MA))和均方根误差(E_(RMS))均小于基于时间序列的GAM,而GLM分析的决定系数R~2则大于GAM,表明前者的拟合具有更好的准确性、稳定性和拟合优度。研究表明,当周期性循环变量(月份、月相和小时等)具有较弱的显著性时,使用基于圆形统计的GLM更能反映月相对拖网渔业CPUE的影响。In order to understand and grasp the pattern of effect of lunar phase on fishery yield or CPUEs(catch per unit effort),the impact of lunar phases on CPUEs of trawl fisheries was analyzed by daily CPUEs of speckled shrimp(Metapenaeus monoceros)in the bottom trawl fisheries along the west coast of Madagascar from 2016 to 2019(from March 2 to November 29 each year)and walleye pollock(Theragra chalcogramma)in midwater trawl fisheries in the western Bering Sea from 2017 to 2020(from June 15 to October 9 each year),combined with two distinct methods of quantification of lunar phases,including generalized linear model(GLM)based on circular statistics method and generalized additive model(GAM)based on time series method.The results showed that the daily CPUEs of speckled shrimp was significantly affected by the lunar phases(P0.05),with the consistent trends in the two methods.The higher CPUEs were found in the first quarter moon period.The model fit of GLM based on circular statistics showed that the lunar phases had significant effect on the CPUEs of walleye pollock(P0.05),with the higher CPUEs in the new moon phase.However,there was no significant effect of lunar phases on CPUEs of walleye pollock in model fit of GAM based on time series(P0.05).Analysis of cross-validation indicated that both the mean absolute error(E_(MA))and root mean square error(E_(RMS))of GLM based on circular statistics were less than those of GAM based on time series,while the coefficient of determination(R~2)of GLM was larger than that of the GAM,indicating that the former model fit had better accuracy,stability,and fit superiority.The findings show that the circular statistical quantification method is more likely to reflect the influence of lunar phases on the CPUEs of trawl fisheries when periodic cyclic variables(monthly,lunar phase,hourly,etc.)are weakly significant.

关 键 词：月相 拖网 单位捕捞努力量渔获量(CPUE) 圆形统计 时间序列 交叉验证 

分 类 号：S931[农业科学—渔业资源]