袁兴伟,姜亚洲,程家骅.利用Δ-分布模型法评估调查数据带有极端值的渔业生物的平均资源密度.渔业科学进展,2011,32(1):1-7 |
利用Δ-分布模型法评估调查数据带有极端值的渔业生物的平均资源密度 |
Estimating the average stock density with dominating large catches based on Δ-distribution model |
投稿时间:2010-03-18 修订日期:2010-05-13 |
DOI: |
中文关键词: 底拖网调查 平均资源密度 调查设计法 -分布模型法 评估效能 |
英文关键词: bottom trawl survey the mean of stock density design-based methods Δ-distribution model estimator efficiency |
基金项目:中央级公益性科研院所基本科研业务费专项基金(中国水产科学研究院东海水产研究所)资助项目(2008T03);科技部公益性项目-东海区渔业资源动态监测网 |
|
摘要点击次数: 2465 |
全文下载次数: 2522 |
中文摘要: |
海洋生物的空间分布具有不均匀性,从而造成底拖网调查数据频数分布显著偏斜。偶尔出现的极端值,会直接影响资源密度均值和方差的估计值。本文根据东海区渔业资源监测调查数据库中资料,利用Δ-分布模型法和调查设计法分别估算了2005年夏季的太平洋褶柔鱼Todarodes pacificus、2007年春季的鳀鱼Engraulis japonicus以及2008年秋季的竹荚鱼Trachurus japonicus的平均资源密度。结果表明,利用-分布模型法分布模型法对三种类资源密度均值的估算值均低于调查设计法,且-分布模型法对资源密度均值进行估算时的标准误也相对较小。因此,针对调查数据中出现极端值的情况,如果资源密度对数转换后服从正态分布,那么利用Δ-分布模型法对资源密度均值进行估算是一种稳健的评估方法,值得在渔业评估上进行推广。 |
英文摘要: |
Due to the patchy spatial distribution of marine biota, data collected from bottom trawlers exhibit a skewed distribution with many zero catches and some very large catches. A common problem in the analyses and interpretation of skewed survey data is that, a single immense catch may account for 50% or more of the total catch during the survey. These extreme values not only greatly affect the estimate of the means but also of the variance. Based on the data collected from the bottom trawl survey conducted in the East China Sea, we estimated the average stock density of Todarodes pacificus in summer 2005, Engraulis japonicus in spring 2007, and Trachurus japonicus in autumn 2008 by employing two methods, namely the Δ-distribution method and the design-based method. The results revealed that the estimators of the means and the standard errors of stock density for the three populations based on the Δ-distribution were obviously lower than those based on the design based method. It was concluded that the mean and the standard error would be overestimated using the design-based method. The Δ-distribution was more reasonable when the survey data follow lognormal distribution. Therefore,Δ-distribution is a robust estimating method and is suggested to be widely usable in further studies. |
附件 |
查看全文
查看/发表评论 下载PDF阅读器 |
关闭 |