Comparison of the relation between the length frequency distribution and sample size for several commercial fish species in the Wentai fishing ground

One of the basic requirements for conducting fishery stock assessment and management research is to obtain effective fishery and biological information for target species. Length frequency distribution (LFD) can provide insight into fishery population dynamics and is easier to collect compared with other data type such as age structure and abundance index. However, it is hard to determine the effective sample size required to describe the length structure for specific fish stock. The quantitative relationship between the precision of LFD and the sample size of fish length data could be regarded as an indicator to optimize the sample size of fishery survey design. In order to quantify and compare the relationship between the precision of LFD and sample size of fish length data for three commercial fish species (Larimichthys polyactis, Pennahia argentata and Harpadon nehereus) on the Wentai fishing ground, a simulation study was conducted to calculate the mean square difference (MSD) of LFD based on actual sample size and various simulated sample sizes based on observed length data collected during spring and summer on the Wentai fishing ground. The simulation study assumed that the observed length data are able to reflect the true length structure for three species. In addition, t-test was used to examine the significant difference of mean length for three species in spring and summer. The results showed that, ① Factors such as biology difference and seasonal variation would affect the relationship between the precision of LFD and sample size. All three species showed significant difference of mean length in spring and summer. The dominate length ranges for three species were different between two seasons. ② The accuracy for estimating the precision of LFD would be impacted by the setting numbers of size bin. The MSD values increased with the number of size bin when sample size was small and this relationship did not have seasonal variation. We recommend setting 10 size bins when the effective sample size was set at 50. ③ The precision of LFD could be considered as an effective indicator to optimize sample size. Given the assumption that 50 fish samples per each station can reflect the true length structure for three species in this study, we recommend setting the sample size at the range of 30-40 individuals for L. polyactis, 42-48 individuals for P. argentata and 34-40 individuals for H. nehereus in spring.