Persistence: A Forgotten Need for Fixed-Station Surveys?

Standard sampling designs have associated design-based estimators that have well known statistical properties—they are designed to provide unbiased estimates of the mean and variance. Application of these estimators to fixed-station surveys is not technically valid but can result in unbiased differences between years if the data exhibit spatial persistence. A persistent system is one in which the station differences maintain themselves from year to year (i.e., the rank of catches among the stations persists from year to year). In a fully persistent system, the changes in abundance derived from fixed stations will be unbiased. Here, the persistence of a fixed-station survey intended to monitor the recruits of economically important species is evaluated. The results can be used to infer for which species changes in abundance estimated from fixed-station surveys will be more accurate than changes estimated from random surveys.