论文部分内容阅读
长鳍金枪鱼作为高度洄游的大洋性鱼类,因其经济价值高、资源量丰富而成为世界海洋渔业的主要捕捞对象之一。本研究根据中西太平洋金枪鱼渔业委员会(WCPFC)提供的南太平洋延绳钓长鳍金枪鱼渔业2000~2012年13年的渔业生产统计资料,结合卫星遥感获得的海洋环境资料,利用广义线性模型(GLM)和广义可加模型(GAM)对其资源密度进行CPUE标准化。GLM模型结果表明,月、纬度、表温、表温梯度、盐度、海面高度、涡动能、100米溶解氧以及交互项年与纬度、月与纬度、月与经度对CPUE影响较大。根据AIC准则,选择最优的GAM模型,其对CPUE的最大解释偏差为58.9%,模型自变量对因变量的最高相关系数达58.4%。GLM模型和GAM模型标准化后的CPUE变化趋势相同,除2011年外均明显低于或接近名义CPUE。2000~2012年资源密度年间和月间变化较大,资源密度最高的年份为2001年,资源密度最高的月份为6~8月,GLM模型和GAM模型标准化后的平均CPUE分别为11.48尾/千钩、12.41尾/千钩和13.54尾/千钩、18.63尾/千钩。本研究中,GAM模型较GLM模型更合适。
As a highly migratory marine fish, albacore tuna has become one of the main fishing targets of the world’s marine fisheries due to its high economic value and abundant resources. Based on the statistical data of fishery production of longline Tuna fishery from 2000 to 2012 in South Pacific Ocean provided by the Western and Central Pacific Ocean Tuna Fisheries Commission (WCPFC) and the marine environmental data obtained by satellite remote sensing, the generalized linear model (GLM) And Generalized Additive Model (GAM) CPUE standardization of its resource density. The results of GLM model show that monthly, latitude, surface temperature, surface temperature gradient, salinity, sea level, eddy energy, dissolved oxygen of 100 meters and interaction term year and latitude, month and latitude, month and longitude have a great influence on CPUE. According to the AIC criterion, the optimal GAM model is selected, the maximum interpretation deviation of CPUE is 58.9%, and the highest correlation coefficient of model independent variable to dependent variable is 58.4%. The trend of the CPUE after the standardization of the GLM model and the GAM model is the same, with the exception of 2011 all significantly lower or close to the nominal CPUE. In 2000 ~ 2012, the resource density varied greatly between years and months, the year with the highest resource density was 2001 and the month with the highest resource density was from June to August. The average CPUEs after GLM and GAM standardized were 11.48 / Hooks, 12.41 Tails / Thousand Hooks and 13.54 Tails / Thousand Hooks, 18.63 Tails / Thousand Hooks. In this study, the GAM model is more suitable than the GLM model.