Review of Statistical Aspects of Survey Sampling for Marine Fisheries

Abstract

Fishery surveys are an essential ingredient of modern fisheries stock assessment. To understand this, one must understand that the survey time series are the essential anchor that makes modern fishery stock assessment modeling possible. Without fishery surveys, fisheries stock assessment scientists would have great difficulty modeling absolute abundance, and therefore agencies would have difficulty setting levels of allowable catch. In this review, we explore the basic assumptions that need to be fulfilled for valid surveys to be accomplished. Although these assumptions are generally well understood, the complexity of survey sampling gear and the complexity of target animal behavior makes them difficult to fulfill in practice and can easily lead to problems when interpreting survey results from even the most carefully designed survey. In reviewing the literature surrounding fishery surveys, their sampling design, modeling, and methods of analysis, it became clear that fisheries scientists have long been preoccupied with coping with the intense variability found in fish catches. This variability is found within fishing hauls, between hauls, between area, time and depth strata. Coping with this variability, which is due to animal behavior, habitat variation, and the nature of fishing gear, will be a constant theme that ties together our review.