Impact of Sample Size Reduction on the Precision of Biomass Estimates in the Gulf of Alaska.

摘要

The sample sizes of the two most recent (2011 and 2013) Gulf of Alaska bottom trawl surveys were reduced by 20% and 33% (to 670 and 548 from 825 stations), respectively because of lack of funding and logistical reasons. In this study we examine the potential impact these sample size reductions has on the precision of the biomass estimates of 11 common species assessed by this survey. Surveys of sample sizes ranging from 300 to 825 stations were simulated by randomly sampling stations without replacement from the 2007 Gulf of Alaska survey. The coefficient of variance was plotted against sample size for each of the 11 species, and the rate of decline of the slope of the fitted curve was used to determine whether the reduced sample sizes of the two most recent surveys resulted in adequate precision. A threshold value of the slope (70% of its value at N = 300) was used to decide whether the reduced sample sizes of the 2011 and 2013 surveys were adequate for each species. The sample size of the 2013 (N = 548) survey resulted in none of the 11 species meeting the threshold, suggesting that the sample size that year was too low. In contrast, 3 species met the threshold at the sample size of the 2011 survey (N = 670) and 10 species met the threshold for a traditional full-scale survey (N = 825).