应用分层抽样技术估计北部湾底拖网渔业产量

传统的按行政单元逐级上报的渔业产量调查方法在数据获取过程中受人为因素的干扰,而全面普查的方法受限于时间和经费,相比而言,抽样调查是一种科学合理的方法。本研究于2007年8月对北部湾底拖网产量进行了调查,以全体底拖网渔船为抽样总体,按功率段划分了抽样层次,按比例分配了抽样单元数,分别以生产渔船总数和总功率数推算总产量,并比较了这两种方法的方差。结果表明,用每kW平均产量估计总产量的变异系数为1.13%,用单船平均产量估计总产量的变异系数为14.65%。同时由于渔业抽样调查的难点在于总体船数的掌握,建议统计推断时,采用单位渔捞努力量渔获量（CPUE）推算总体渔获量。

Estimation of trawler catch in Beibu Gulf by stratified sampling

It is unavoidable to make errors by traditional survey and reporting system because of manmade interference. And comprehensive survey is not feasible in terms of time and money. Thus sampling survey is a scientific and reasonable method. A sampling survey was conducted to investigate trawler catch in Beibu Gulf in August 2007. We divided sampling stratifications by power and distributed sampling units by vessel and power ratio within trawler sampling population. Total catch and variations were calculated by total fishing vessels and total machine power. Coefficient of variation（ CV） was 1. 13%,calculated by average catch per kilowatt. CV was 14. 65%,calculated by average catch per vessel. It was difficult to understand the total fishing vessels in sampling survey. We suggested calculating total catch based on catch per unit effort（ CPUE） in fishery sampling survey.