Estimating effective sample size for fish length distribution from fishery samples on Georges Bank: A comparative study

Because of the aggregative behavior of fish, the effective sample size is more informative than the actual sample size in estimating length distribution of a fish stock. In this study, using the data collected from the University of Massachusetts, Dartmouth SMAST Study Fleet, we estimated the effective sample for length distributions of six groundfish species on Georges Bank including Atlantic cod, winter flounder, witch flounder, yellowtail flounder, American plaice and haddock. Both statistical methods and simulation methods were applied. Analyses of skewness and kurtosis were used as a proxy for accurately representing distributional shapes and the Kolmogorov-Smirnov test was used to compare length distributions in the total sample and in the simulated resamples. This study suggests that the estimated effective sample size is much less than the number of fish measured. However, the sample size for each station should represent the number of fish caught at that station; otherwise it would cause loss of information. This study helps develop an optimal design for the surveys of groundfish and other species which balances the quality of the data collected and costs of the sampling effort.