Activity B - Validated Data

Assemble training datasets from validated, historical echosounder and co-variate data (FY24-FY26)

Developing advanced analytical methods for echo classification requires high-quality, validated (accurately attributed to lowest taxonomic level) acoustic data and associated co-variate data. Most Centers have accumulated these data for stock assessment, but not in formats that are useful for input to advanced analytical models. In addition, co-variate data (e.g., oceanographic data, eDNA, optics, net catches, life history) and acoustical characteristics of species of interest are critical to improving the effectiveness of echo classification.

Validated acoustic data can be obtained from empirical observations and theoretical scattering models. Empirical observations consist of acoustic data synchronized and co-located with net catches, optical images, or other samples of the origins of acoustic backscatter. Theoretical scattering models provide species-specific predicted backscatter and target strength that are used to confirm empirical observations over a wide range of environmental conditions, anatomical features, and acoustic frequency spectra. We will assemble example training datasets from pre-existing data sources that includes ancillary data used for acoustic validation, which will form the basis for developing advanced analytical methods for echo classification.