NEAFWA 2017

AUTHORS: Aaron Bunch, Dan Goetz, Dustin Martin - Virginia Department of Game and Inland Fisheries

ABSTRACT: Monitoring studies aimed at assessing status and trends are imperative for addressing fisheries management goals and objectives because of the dynamic nature of fish populations. However, routine surveys often overlook sample size considerations needed for statistical validity in making fisheries management decisions. We evaluated three datasets using bootstrap techniques to determine how coefficient of variation (CV) of catch per unit effort (CPUE; fish/hr) estimates respond to sample size (e.g., total number of sites, or sample days depending on survey type). Datasets were obtained from a variety of monitoring studies across Virginia, which included electrofishing and creel surveys from lakes, rivers, and small streams. Typically, CV curves will show a negative relationship (i.e., exponential decay trend) as sample size increases. We found this to be true with the number of sites needed to accurately estimate Largemouth Bass Micropterus salmoides CPUE in a large impoundment. However, the opposite trend occurred with angler CPUE for Largemouth Bass and Blue Catfish Ictalurus furcatus from James River creel surveys. In this case, despite high and increasing trend in CV, there was still a point of diminishing returns for added sampling effort above eight creel days for estimating CPUE of those target species. When determining adequate sample size biologists should consider overall study objectives, time and budget constraints, and the distribution of the experimental unit in question. Lastly, periodic evaluation to determine sampling efficiencies is essential for maintaining data integrity and standardization, and should become standard practice.