Building a toolbox for studying marine ecology using large ocean sonar datasets
Abstract
Echosounders are high-frequency sonar systems specialized for monitoring fish and zooplankton. They are the workhorse in observing life in the ocean and are widely used in fisheries and marine ecological studies. The recent explosion of the availability of echosounder data from diverse ocean observing platforms has created unprecedented opportunities to study marine ecosystems at broad spatial and temporal scales. However, this acoustic data deluge has also brought a multitude of challenges, from data access and analysis to interpretation.In this talk I will discuss our work in developing data-driven methodologies and software tools to tackle these challenges. I will show how matrix decomposition techniques can extract biologically meaningful patterns from complex echosounder observations, producing a compact representation that is further conducive to systematic analysis with other ocean variables. This work is accompanied by our development of echopype, an open-source software package to enable interoperable and scalable echosounder data processing. These advances form the foundation for our continuing effort to gain ecological insights from large volumes of echosounder data.
