Dr. Dezhang Chu
Woods Hole Oceanographic Institution, Woods Hole, MA 02543
Program Manager: Dr. J. Michael Jech, NOAA/NMFS/NEFSC
Precise, accurate, and efficient estimates of the abundance/biomass of standing fish stocks will provide valuable information for better understanding the marine ecosystem processes. Splitbeam echo sounders have been used routinely as primary acoustic survey systems and have provided quantitative information on the abundance and biomass of fish stocks. Multi-beam sonar systems have been used primarily in seafloor bathymetry surveys and not until recently are used in fisheries acoustic surveys. Although the multi-beam sonar systems have a great potential to provide quantitative 3-D images of fish schools and a much more efficient way to estimate the abundance of the fish stocks accurately, there are no standardized quantitative data processing and visualization techniques available. The primary objectives of the proposed work are to streamline the procedures of quantitative multi-beam sonar data processing techniques and to develop a GUI-based software for processing and visualizing multi-beam sonar data.
It is my hope that the results from this project will lead to an improved acoustic assessment of the fish standing stocks and help us to better manage the marine ecosystem.
The proposed research consists of three major components: (1) multibeam sonar system calibration experiment; (2) development of the theory and techniques for quantitative multibeam sonar data processing; and (3) development of a GUI-base Matlab software package for implementing the quantitative multibeam sonar data processing.
1. Multibeam sonar calibration experiment:
For quantitative acoustic measurements, calibration with high quality is required. From March 5 to March 29, 2004, the calibration experiment was conducted in the sea-well on Iselin Dock at the Woods Hole Oceanographic Institution (WHOI) using the calibration facilities developed under the other projects (NSF OCE-0002664 and NOAA NA97OG0241). The sonar system was the Simrad SM2000/90 kHz provided by the Northwest Fisheries Science Center (NWFSC), NOAA/NMFS (Fig. 1).
The sonar system has 80 receiving channels with an operating option of using the internal or external transmitter. The sonar calibration data include (a) 2-D farfield (23 m) directivity (Fig. 2) and near field (12 m) measurements; (b) beampattern of individual beams (fig. 3); (c) Time-Varied-Gain (TVG) measurement from 0-50 dB with other settings kept the same as or similar to the settings commonly used in the acoustic surveys; (d) influence of pulse duration (150, 300, and 600 ms) on quantitative measurements such as on Target Strength (TS) measurement; (e) transmit power setting (high/low) on quantitative measurement; and (f) stability and variability measurement by recording the target echoes over a period of time. These data serve as a basis for achieving reliable and quantitative acoustic measurements.
2. Theory and techniques to quantify the acoustic sonar data:
To quantitatively estimate abundance and/or biomass of the marine animals such as fish schools from the raw multibeam acoustic data, a theory is developed to estimate the Target Strength (TS) for resolvable echoes and volume backscattering strength (Sv) for overlapping echoes. TS and Sv are the two important quantities in fisheries acoustic applications. The theory takes into account a variety of factors that can potentially affect the estimation accuracy of the interested biological quantities. One of these factors is the inconsistency between the estimated environmental parameters that are set as default parameters in the sonar system and the actual in situ parameters for both calibration and field measurements. This is the unique characteristic of the multibeam sonars since the accuracy of the beamformed data depends on the accuracy of the actual sound speed in water. The conversion factor from the raw acoustic data to target strength depends on other system settings including TVG, transmit power, and pulse duration. The equivalent beam angle Ψ or its logarithmic equivalence Ψ= 10log10 Ψ of each of the individual
beams (a total of 128 nominal beams for Simrad SM2000 series), another crucial parameter and has been estimated almost exclusively based on the theoretical prediction due to experimental difficulties, can be determined experimentally (Fig. 4) using the developed theory.
Sensitivity analysis of the system was also performed. These include the dependence of the beamwidth of the mainlobe ( !w), sidelobe level (SL) on uncertainties in positions of array elements (Fig. 5) and sound speed uncertainty, and beamform weighting function, etc. It is found from Fig. 5 that with 1-mm radial position uncertainty, the sidelobe level increases by about 1.5 dB. In general, the width of the mainlobe is quite stable, but the sidelobe level is much more sensitive to these uncertainties.
3. GraphicUserInterface(GUI) based Matlab software:
The software, Quantitative MultiBeam Sonar Processor (QMBSP), is developed specifically for Kongsberg Simrad SM2000 multibeam sonar system but can be easily extended to other multibeam sonar systems such as Reson Seabat 8000 series. The software has a Graphic User Interface (GUI) with multi-layer menu-driven operation options (Fig. 4). The major features of the software include: (a) User defined sonar system configuration with options of default sonar configurations of commonly used Simrad SM2000 multibeam sonar systems; (b) Capability of processing raw and beamformed data; (c) Bottom detection algorithm; (d) Target tracking
capability, positions and TS for resolved echoes; (e) Volume Backscattering Strength (Sv) display, which can be easily converted to abundance and/or biomass assessment; (f) Flexible visualization options. The biggest advantage of the QMBSP over the other existing multibeam sonar processing software is its capability of incorporating a full calibration data set.
Foote, K.G., D. Chu, T.R. Hammar, K.C. Baldwin, L.A. Mayer, L. C. Hufnagle, Jr., and M.J. Jech. “Protocols for calibrating multibeam sonar by standard-target method”, J. Acoust. Soc. Am., in press.
Hufnagle, L.C. Jr., D. Chu, K.G. Foote, T.R. Hammar, J. M. Jech. Calibrating a 90-kHz multibeam sonar: illustrating protocols. Proc. MTS/IEEE Oceans' 2004, Kobe, Japan, pp. 438-442.
Chu, D., L.C. Hufnagle, Jr., J.M. Jech, Quantitative acoustic measurements with multibeam sonars (abstract). To be presented at the ICES-Working Group on Fisheries Acoustics Science and Technology, 2005 meeting (WGFAST), April 19-22, Rome, Italy.
SUMMARY OF INTERACTION WITH NOAA
During the first year of the project, a multibeam sonar system, SM2000/90kHz, owned and operated by the Northwest Fisheries Science Center (NWFSC, NOAA/NMFS), was calibrated at the Woods Hole Oceanographic Institution (WHOI) using the calibration facilities developed previously with funding from the National Science Foundation through award number OCE-0002664. Larry Hufnagle, the operator of this multibeam sonar at NWFSC, participated in the entire calibration experiments. Team leader of the monitoring group at NWFSC, Guy Fleischer was also visited the WHOI for two days during the calibration trial. Since the calibration protocols of the multibeam sonar and conventional split-beam sonar systems are quite different, participation in the calibration could certainly provide a good opportunity for the operator to understand and experience the calibration procedures for multibeam sonars. As a result, it will help the acoustic team at the NWFSC to obtain more reliable acoustic assessment of fish stock based on the fully calibrated acoustic data.
I also provided the theoretical formulae to Dr. Mike Jech at the Northeast Fisheries Science Center (NEFSC, NOAA/NMFS), who also serves as the program manager of this project, allowing him to calculating the equivalent beam angle (! ) of his split-beam echo sounder based on the measured 2-D beampattern. I am planning to visit NWFSC in Seattle either next month or in May. The trip will serve multiple purposes: (1) give a talk on application of the acoustic technologies to fisheries applications, emphasizing the multibeam sonar application; (2) show them the new development of the software (QMBSP) that I have been working on; (3) process some of the field data collected during one of their survey cruises last year using the developed QMBSP; (4) refine the software to correct any possible errors in the program; and (5) exchange ideas on scientific research involving fisheries acoustics and explore the potential opportunities for future collaboration.
SUMMARY OF EDUCATION AND OUTREACH ACTIVITY
Outreach activities: In addition to the presentations in the international workshops and conferences, I have also collaborated with the SonarData Pty. Ltd, an Austrian software company, on improving the quality of their flagship software: Echoview. EchoView is a data analysis and visualization software package and has been used by many users around world, including different branches of the NOAA/NMFS, as a standard tool for processing split-beam acoustic survey data and fish stock assessment. The recent version of the Echoview (V3.10) is able to process and visualize the multibeam sonar data but only has limited capability of accurately including the calibration data. The discussions and communications between the key personnel from SonarData (Ian Higginbottom, manager and director, and Tim Pauly, director of the Technical Development) and myself last year led to an agreement that I would provide the theoretical formulae needed for more accurate multibeam sonar calibration so they would include the new enhanced calibration capability in the later version of Echoview, and they would allow me to evaluate their software by providing some special features and to ensure the correctness of the new calibration features added to their software. I believe that such collaboration will help the NOAA/NMFS to obtain better acoustic assessment of the fish standing stocks and in turn to better manage the marine ecosystem.