Effects of in situ target spatial distributions on acoustic density estimates

Abstract

One goal of acoustic-based abundance estimates is to accurately preserve spatial distributions of organism density and size within survey data. We simulated spatially-random and spatially-autocorrelated fish density and σbs distributions to quantify variance in density, abundance, and backscattering cross-sectional area estimates, and to examine the sensitivity of abundance estimates to organism spatial distributions and methods of estimating acoustic size. Our results show that it is difficult to simultaneously estimate fish density and maintain accurate σbs-frequency distributions. Among our acoustic backscatter estimation methods, a weighted-mean from a local search window provided optimal estimates of density, abundance and σbs. Other methods tended to bias either σbs or density estimates. This analysis identifies the relative importance of variance sources when estimating organism density using spatially-indexed acoustic data.