Using accurate insonification volumes to estimate fish abundance from acoustic survey data

Abstract

Acoustic transect-surveys paired with biological sampling of detected fish targets are an effective method for surveying fish in large freshwater and marine basins. While echo data can be deconvoluted for detected fish targets and biomass through well-established biophysical models by echo-counting, trace-counting and echo-integration methods, commonly adopted geometric models for insonification volumes of transecting sound-beams are either inaccurate or erroneous, which can significantly bias estimated fish densities and abundances for surveyed basins. In this paper, we present accurate models for insonification volumes of acoustic beams to correct these errors. The practical value of the models is demonstrated with results from the acoustic survey data acquired in two major nursery lakes for sockeye salmon (Oncorhynchus nerka) on Vancouver Island in British Columbia, Canada. By using the correct ping-to-ping insonification volumes and taking considerations of unsampled volumes in the upper water column, estimates using accurate insonification volumes were within a 5 % difference between the echo-counting and trace-counting biophysical models. The presented results and methodologies can be readily replicated and are broadly applicable to acoustic surveys of fish abundance in freshwater and marine environments.