Incorporating effective sample sizes into sampling designs for reef fish

Reef fisheries are usually sampled via two stage cluster sampling. Sampling efficiency for cluster sampling is low because of non-independence among fish samples collected from individual clusters/trips. This study establishes a new procedure for incorporating effective sample sizes into cluster sampling designs. The first step in this new procedure is to examine the degree of non-independence, the within and between trip variability, and the effective sample sizes of historical cluster samples to identify factors that influence sampling efficiency. The second step is to examine how changes in the effective sample size influence the precision of the length frequency distribution. The third step is to evaluate how different sampling designs change effective sample sizes. In general, the trip sample size should be small when the within trip variability is low. The number of sampling trips should be large when between trip variability is high. However, an optimum sampling design is also subject to limitations in cost and sampling opportunities. The procedures established here can also be applied to fishery independent surveys.