Project number: 1993-237
Project Status:
Budget expenditure: $251,081.00
Principal Investigator: Tony Koslow
Organisation: CSIRO Oceans and Atmosphere Hobart
Project start/end date: 29 Jun 1994 - 29 Apr 1999


1. To develop multifrequency-based methods of analysis for an expanding and diverse range of applications in Australian fisheries and associated ecosystems

Final report

ISBN: 0 643 06192 4
Author: Rudy Kloser
Final Report • 2017-09-29 • 11.07 MB


The development of the ECHO software has enabled the collection and analysis of large multifrequency acoustic data sets. The data can now be processed in a timely manner (via overlays) to quality assure and interpret underlying acoustic characteristics in the signals. From these analyses we can perform seabed and biomass processing on the individual frequencies. Alternatively the three frequencies can be mixed in a visual display that highlight distinct species groupings.

In the deep water orange roughy fishery the multi-frequency analysis clearly identified the dominant species groupings that occur. These combined echograms show that the community composition around the spawning aggregation is complex and not truly represented in commercial trawling operations. This will have a significant impact on the interpretation of acoustic biomass estimates. The ability to remotely sense the acoustical dominant species composition using multi-frequencies is a major advance in fisheries acoustic research.

The ECHO software has enabled us to develop methods to remotely sense fisheries habitat as well as ensuring high data quality. The software developed in this study was used to process acoustic seabed data obtained from the FRDC South East Fishery research project, Bax and Williams (1998) using simple classification methods. This project and associated software development was used to provide data to the acoustic benthic habitat FRDC project 93/058, Pitcher et al (1998). This enabled the development of enhanced bottom classification methods. We have further developed the processing techniques in this report and demonstrated that the use of multiple frequencies can improve seabed misclassification rates from 27% at a single frequency to 8% at multiple frequencies. This could lead to a major advance in our ability to correctly classify different fisheries habitats and to monitor the long term stability of these habitats.

Related research