Global block kriging to estimate biomass from acoustic surveys for zooplankton in the western Aleutian Islands

Abstract

Because of patchy distributions and the expense of collecting and processing net samples, zooplankton population data are often characterized by broad confidence intervals, with little detailed information on vertical and horizontal distributions. Although acoustic techniques and optical plankton counters combined with nets can supply much more detailed information on plankton distributions, application of classical statistical procedures to such data may be effected by pseudoreplication from autocorrelation in closely spaced samples. We apply two-dimensional kriging for population mapping and global estimation. The technique utilizes a computational estimator of the global block kriging mean, which can be applied to large data sets common to hydroacoustic surveys. The techniques are outlined using examples from acoustic data taken in the western Aleutian Islands. This technique has two fundamental advantages: (i) it minimizes pseudoreplication by accounting for autocorrelation, permitting rigorous statistical inferences, and (ii) it provides an effective technique for visualizing the results. These advantages may make this a useful technique for identifying changes in the size and distribution of populations.