Assigning ground truth to hydroacoustic density estimates
Improving accuracy and precision in fish community assessments.                Workshop 24-28 October 2011

Information updated 2011 08 24

In this workshop, European researchers will learn from and work together with North American researchers. The main objective is to test assignment methods and approaches using data from various large lakes in Europe, and to discuss the suitability of different methods for different types of lakes (e.g. depending on size and productivity). Recent methods to apportion catches and assign ground truth data to hydroacoustic density estimates will be applied.

Methods to be applied, compared and discussed are 1) Keep it simple stupid (KISS), 2) Hierarchical averaging (HAM), 3) Nearest-trawl (Near), 4) Nearest-trawl method with depth stratification (NearD), and 5) Classification tree (Tree). We will also be introduced to a fish community simulator developed by Jean Adams, statistician at the Great Lakes Science Center, Michigan U.S., using the R programming language (R Development Core Team 2011). The highly flexible simulator allows a creation of a pelagic fish community based on knowledge of species composition, size structures of each population, and habitat preferences of each species. The lake simulator may thus demonstrate the power of knowing the true numbers of fish in a given study area, to aid comparison of apportionment methods.

We believe that improved methods will have a broad potential for application in assessments of fish populations and species assemblages, especially in large lakes. More precise estimates of fish communities in large lakes may be used e.g. for protection of sensitive species and sustainable use of fish resources.