FIS004: Slippage mitigation acoustics (SMAC)

Start Date: Jan 2015

End Date: Sep 2015

Main Contractor(s): University of Aberdeen

Other Sponsor(s): European Fisheries Fund

Extended Title: Slippage Mitigation and Acoustic Characterisation (SMAC). Phase I: sonar adaptation and development of data processing tools.

Main Research Category: Discard Reduction - Technical

View Final Report

 

Project Objectives

Mackerel is the most important pelagic species in the Scottish fishing fleet, worth approximately £130 million p.a. at first landing.  In January 2015, the Scottish pelagic fleet, which is largely dependent on this species, will be the first to be subject to the landings obligation.  In the case of the mackerel fishery, the main challenge is to avoid slippage, the practice of releasing large amounts of (mainly small) mackerel caught in the net, back into the water.  The main objective of this research is to develop analytical techniques to allow the size of fish to be determined prior to capture and so avoid the problem of slippage.  This project provides the first, essential phase, to develop post-processing methods to analyse data from an adapted  sub-bottom profiler which will allow for broadband spectra to be created from fish schools.  These broadband spectra have been demonstrated theoretically to be capable of determining the size of mackerel in schools.  If that theoretical principle can be translated into the field, then it provides a means by which the size of mackerel can be determined prior to capture (and so avoid pelagic discarding).  This research answers FiS call 1, point 7: “Technical innovation in gear design and operation... to address the challenges of the landing obligation”.  The results of the project will be a suite of analytical tools that will allow a subbottom profiler, recently acquired by Marine Scotland Science, to be used as an innovative new broadband echosounder: this has further relevance to the call by representing a “significant advance over existing designs”.  The results will then be used in subsequent future projects to test the new device during the mackerel fishery (e.g. in autumn 2015).

 

Specific Objectives:

Objective 1.  Adaptation of the Edgetech 3200 sub-bottom profiler.  The raw data from this device will be interpreted to allow it to serve as a broadband echo sounder, processing whole water column echoes.   This will be achieved with the assistance of the WHOI.  WHOI have experience in using this device in this manner and are willing to share data processing routines (software) to transform the broadband echoes using pulse compression processing.  In the early part of the project, the research team will travel to WHOI and be trained to adapt the device.  The training will also cover calibration techniques (to be carried out locally in future phases, as described above) and analytical techniques (see objective 2 below). 

Objective 2.  Development of post processing software.  The WHOI software will be adapted and augmented to produce volume backscattering strengths (VBS) as a function of frequency (1.5 kHz to 60 kHz) – broadband spectra.  These will also be available at any chosen narrowband frequency, and then resolved vertically according to pulse duration and frequency, and horizontally according to pulse repetition rate and vessel speed. This will produce two dimensional echograms of objects such as fish schools which scatter sound in real time with vertical resolutions of the order of a few cm to a maximum of 20 cm at 1 KHz.  Crucially, the software will include facilities to produce broadband spectra of any area selected on the echogram.  

Objective 3.  Development of preliminary fish sizing algorithm.  Raw data from the sub-bottom profiler will be collected during a research cruise in October 2014 from schools of mackerel of known size.  These will provide the data for tasks 1 and 2 (above).  Once the above tasks are complete, then the broadband spectra from mackerel schools will be fit to theoretical scattering models to determine fish size. A distorted wave born approximation (DWBA) model will be applied using scattering model parameters fitted by a Bayesian approach to account for parameter uncertainty and to incorporate prior knowledge. This will be done on a standard PC using programming routines in the open source software language R which UNIABDN already has.

 

Summary of Project Outcomes

Mackerel is the most important pelagic species in the Scottish fishing fleet and the most valuable single species fishery in Scotland, accounting for 29% (£126 million) of the total value of Scottish landings in 2013.   The species consists of one stock, known as the North East Atlantic mackerel stock, and is managed under the European Union’s (EUs) Common Fisheries Policy (CFP), normally in agreement with other coastal states (Norway & the Faroe Islands).  In spite of recent disputes with Iceland over catch shares, the stock is exploited sustainably.  The total catch taken from this stock in 2013 was 932,000 t, which at a approximate value of over £1000.t-1 makes this a billion pound fishery, important to many other northern European nations such as Norway, Netherlands, Ireland, the Faroe Islands, and more recently, Iceland.

In January 2015, these northern European fleets will be the first to be subject to the new landings obligation, effectively banning the practice of discarding.  In the case of the mackerel fishery, the main challenge is to avoid slippage, the practice of releasing large amounts of (mainly small) mackerel caught in the net, back into the water.  The goal of this project was to develop analytical techniques to allow the size of fish to be determined prior to capture using a novel broadband sonar device and so avoid the problem of slippage.  As such this project was considered as the first, now completed phase, of an envisaged sequence of projects summarised in the flow chart below (Figure 1).

 

Figure 1.  Phased sequence of projects to deliver a system for the remote determination of mackerel size.  Green boxes represent the current completed project.  Red arrows and boxes indicate the critical phase: whether the broadband spectra (acoustic characteristics), collected from mackerel schools in the wild, conform to theory and allow for size to be estimated.  Blue boxes represent further phases required to achieve the ultimate goal of having a system for sizing mackerel for the fishing fleet.

 

This project (Phase 1) involved the adaptation of an EdgeTech 3200 sub-bottom profiler (shown on the front cover) recently acquired by Marine Scotland Science (MSS) to work as a low frequency echo sounder.  Concepts and computer code developed at the Woods Hole Oceanographic Institute (WHOI) were adapted to provide a set of tools to process whole water column echosounding data from the device. 

This document is a technical description of the routines used to adapt the sonar (mostly to do with pulse generation); to analyse data from the device so that broadband spectra can be measured; and to size fish with a preliminary sizing algorithm.  As such the document reads more like a manual.  This was necessary because the manual provided with the sub-bottom profiler describes how to use it for its intended purpose – geological surveying - not for whole water column broadband echosounding. 

In addition to describing methods to adapt the system for echosounding, there is an extensive suite of programmes adapted from programmes built by WHOI to process the data.  The steps are described in order to obtain (i) an echogram; and (ii) broadband spectra from the 3 channels effectively providing a broadband range from 1 to 60 kHz.  The complete code is not reproduced in this document: although  it forms the bulk of the work, it contains over 4000 lines of computer syntax which would have little meaning to anyone not familiar with coding and acoustics.  Calibration of the system is also described drawing on the experience of WHOI staff who were sub-contracted for the project. 

Finally a preliminary least squares algorithm to determine mackerel size based on simulations of mackerel broadband spectra is described.  Without any real data it is difficult to demonstrate its utility, hence the critical Phase 2 where real data should be collected at sea (Figure 1).  A Bayesian approach was also tested, but the run times (over 1 hour) make it impractical for use at sea.

The document will allow MSS staff and others to operate the EdgeTech sonar as a broadband echosounder and collect broadband spectra from mackerel in future projects (Phase 2).  This will allow for testing of the algorithm.  If this succeeds, then this can then be field tested further in the mackerel fishery (Phase 3).  Finally, commercial implementation of a stripped-down simple system could take place in Phase 4 in order to have a viable system available to the fishing industry.