黄河口鱼类底拖网调查采样断面数的优化

Sample size optimization for cluster design of bottom trawl fish surveys in the Yellow River estuary

优化调查采样设计方案,利用有限的调查成本获取准确可靠的渔业资源数据,对于开展独立于渔业的科学调查十分重要。根据2013年8、10月和2014年2、5月在黄河口及其邻近水域进行的渔业资源底拖网调查数据,选取短吻红舌鳎(Cynoglossus joyneri)和矛尾虾虎鱼(Chaeturichthys stigmatias)作为目标鱼种,以其平均个体体长、平均个体体重为调查采样优化目标,利用计算机模拟方法对黄河口水域的渔业资源底拖网调查生物学数据进行再抽样,以平均体长、平均体重估计值的相对估计误差(REE)、相对偏差(RB)和变异系数(CV)作为优化评价指标,对基于整群抽样方法的黄河口及邻近海域的调查采样断面数进行优化。结果表明,对于目标鱼种的平均体长、平均体重指标,模拟估计值的REE、RB和CV均随着断面数的减少不断增加,调查断面数少于3时,各指标的变化幅度较大。断面数由5减少至3,REE值平均增加2%,RB值平均增加0.13%,CV值平均增加1.95%,同时渔获量降低近40%。因此,断面数为3可视为黄河口及邻近海域可接受的最优调查断面数。

Fishery-independent surveys are essential for collecting high-quality data to support the stock assess- ments and management of regular fisheries. In general, such programs are more costly and time-consuming than commercial fishery-dependent programs. Thus, considerable interest exists in using computer simulations to opti- mize methods for obtaining high-quality data with limited sampling effort. Currently, high intensity fish- ery-independent bottom trawl surveys may negatively affect and disturb fish populations and the ecosystem of fragile estuarine habitats. These areas support many important fishery species; however, they are also among the most extensively affected and threatened aquatic environments due to fishing pressure and environmental stressors such as coastal development. In this study, we developed computer simulations to evaluate and optimize sampling of mean body length and weight of target fish species in a cluster sampling survey. For use in simulations, bottom trawl surveys were conducted in the Yellow River estuary and its adjacent waters during 2013 （August, October） and 2014 （February, May） to collect abundance and biological-trait data on red tongue sole （Cynoglossusjoyneri） and finespot goby （Chaeturichthys stigmatias）. The relative estimation error （REE）, relative bias （RB）, and coeffi- cient of variation （CV） were used to measure the performance （accuracy, precision, and efficiency） of sampling schemes. These indices increased for simulated data when the number of sampling sections decreased. In the cur- rent survey design, a reduction in sampling-section number from five to three would reduce sampling effort by 40%, while increasing REE by only -2% in about 40% of the catches. Thus, three sections are acceptable for sur- veys designed to obtain size-based indicators. This study also showed that sampling-effort optimization may vary between different survey objectives. Therefore, a post-survey＇analysis will improve fishery-independent survey designs based on specific survey goals, thereby yielding more effective survey data.