|
This description is from the Proceedings of the IEE Colloquium on "Recent Advances in Sonar Applied to Biological Oceanography," London, UK, June 1998.
FROM ACOUSTIC SCATTERING MODELS OF ZOOPLANKTON TO ACOUSTIC SURVEYS OF LARGE REGIONS
By: Timothy K. Stanton
Overview
The bioacoustics research group at the Woods Hole Oceanographic Institution (WHOI) has a broad program and is actively conducting field surveys, laboratory acoustic scattering experiments, and theoretical modeling of acoustic scattering. While the focus of the research has been on zooplankton, we have also investigated the scattering by internal waves and suspended sediment as they sometimes contribute significantly to the acoustic echoes. The field surveys are taking advantage of our newly developed towed instrument, BIOMAPER-II, that contains a suite of acoustical (43 kHz to 1 MHz), optical, and environmental sensors. The system has been used to map portions of Georges Bank near Cape Cod, MA, US. Our laboratory contains a wide range of acoustic transducers (24 kHz to 1.4 MHz, some of which are broadband) that are used to measure the acoustic backscattering as a function of acoustic frequency and animal orientation. The system has been used to measure the acoustic scattering properties of live tethered animals, either at sea on the deck of a ship with freshly caught animals or on land with live animals that have been maintained. The laboratory data have provided a basis for the development of physics-based acoustic scattering models of the zooplankton which have been used to interpret the field surveys. The models take into account the material properties of the animals (fluid-like, shelled, or gas-bearing), size, shape, and orientation, as well as acoustic frequency. Broadband echo classification methods have been developed based on the models and data. Finally, two new systems are under development: one is used to measure target strength of individuals in situ and the other is to be used to measure sound speed of the animals in situ.
Below, the collective efforts of the active sonar bioacoustics team at WHOI are briefly summarized with references given. Given the focus of this colloquium, only the research efforts on active acoustics are given. There are also substantial efforts, led by Peter Tyack, Bill Watkins, and Darlene Ketten, on passive bioacoustics (cetaceans).
Field Survey Instrumentation
We have recently completed construction of BIOMAPER-II (BIo-Optical Multi-frequency Acoustical and Physical Environmental Recorder) ( Fig. 1 ). This towed body contains a wide suite of acoustical, optical, and environmental sensors. 1) Acoustics. The system contains two identical sets of acoustic transducers, one set looking upward and the other looking downward. Two sets are used so that coverage of the water column both above and below the towbody is achievable when the system is towed well below the surface. The frequencies in each set are 43 kHz, 120 kHz, 200 kHz, 420 kHz, and 1 MHz. All systems are split beam except for the 1 MHz systems. 2) Optics. A video plankton recorder (VPR) is mounted on the towbody and can image zooplankton with 15 micrometer resolution. The VPR is used to help ground truth the acoustics. Also included in the optics package are an optical transmissometer and fluorometer, and various optical spectral devices (absorption, attenuation, down-welling irradiance, and up-welling irradiance). The wavelength bands of the optical spectral devices cover the visible spectrum and were selected to match those on the SeaWiFS (satellite) ocean color sensor. 3) Environmental sensors. Depth, temperature, and conductivity sensors are in this package.
The system can be towed either at constant depth or in an undulating fashion (i.e., "tow-yoed"). Inside the tow cable are optical fibers for high speed system control and data telemetry as well as conductors to transmit power to the underwater systems. On board the ship is a winch and portable van (6.1-m-long by 2.4-m-wide) that houses ship-board computers and displays as well as operators.
Acoustic Scattering Laboratory
The acoustic laboratory consists of an array of pairs of acoustic transducers that is mounted in a 3.7-m-long by 2.4-m-wide by 1.5-m-deep tank ( Fig. 2). The animals are suspended in the middle of the acoustic beams by a single thin monofilament line (as thin as 59 microns in diameter) that is acoustically transparent at all frequencies. The tank is filled with filtered seawater. Depending upon the application, the array is either mounted on the bottom looking in the upward direction or mounted at mid-depth and looking horizontally. In each pair, one of the two transducers is used as the transmitter and the other the receiver. By using two transducers, transmitter noise and ringing of the transmitter are less of a problem and closer scattering ranges can be achieved. Furthermore, the system is linear and is easy to calibrate. The frequency range of the system spans 24 kHz to about 1.4 MHz through a combination of narrowband and octave-band broadband transducers. The frequencies of the narrowband transducers are: 24 kHz, 50 kHz, 75 kHz, 125 kHz, 165 kHz, 200 kHz, 305 kHz, 345 kHz, 470 kHz, and 1 MHz. The broadband transducers have center frequencies at about 250 kHz, 500 kHz, and 1 MHz.
The pulse-echo electronics are controlled by a personal computer and include a programmable waveform generator and power amplifier at the transmission end of the system and a preamplifier, bandpass filter, and digital oscilloscope (for capture, display, and transfer of data to computer) at the receive end of the system.
The acoustic echoes are correlated with animal orientation in one of two ways: 1) Each animal is rotated by a computer-controlled stepper motor ( Fig. 2 ). The animal is rotated a small increment (typically 1 degree) after the echo from each ping is recorded. 2) A video camera is used to image a freely moving animal at the same time that it is being insonified (not shown). The data sets are coregistered by use of a common trigger. Enough pings are collected so that a full range of orientations are realized.
Acoustic Scattering Modeling
Given the diversity of animal types and applications, a wide range of acoustic scattering models have been developed. Because of the large number of animals, they have been grouped according to gross anatomical features: fluid-like (i.e., they do not support a shear wave), elastic shelled, and gas-bearing ( Fig. 3 ). The models vary in complexity and accuracy. Some are relatively simple and can predict average levels for the full usable range of frequencies while others are very complex and can predict, at least qualitatively, the structure of broadband echoes on a ping-by-ping basis ( Fig. 4 ).
The deformed cylinder method using the Distorted Wave Born Approximation (DWBA) has been very successful in predicting scattering by various elongated fluid-like zooplankton (euphausiids, copepods, and shrimp). A deformed sphere approach, using either rays or a modal-series solution has been used for elastic shelled bodies (gastropods and benthic snails). A hybrid solution that takes into account both the gas and tissue has been used for gas-bearing zooplankton (siphonophores). A description of our zooplankton scattering models can be found in Stanton (1989a,b; 1990), Stanton et al. (1993a,b; 1994a,b; 1996a, 1998a,b,c), Stanton et al. (in manuscript), Stanton and Chu (in manuscript), Chu et al. (1992, 1993), McGehee et al. (in press), Wiebe et al. (1990), and Greene et al. (1991).
Advanced Signal Processing And Echo Classification Of Broadband Echoes
Echoes due to a broadband incident signal are rich with information. With knowledge of the scattering properties of the animals, we have been able to size and classify the animals by use of signal processing. The methods rely on the condition that the animals are resolved acoustically and have currently only been tested under laboratory conditions: 1) By using either a model-based or empirical model of the scattering, the animals can be classified according to various statistical properties of the spectral signature (Martin et al., 1996; Martin-Traykovski et al., in press; Martin-Traykovski et al., submitted). 2) By temporally compressing the echo (similar to a matched filter), the resultant time series contains a series of peaks specific to the features on the animal. The statistics of the separation between the peaks have been used to both classify the animals according to anatomical group as well as to estimate their size (Chu and Stanton, in press). The pulse compression process also has the advantages of significantly improving the signal-to-noise ratio and range resolution.
Systems Under Development For In Situ Characterization Of Zooplankton
We have two systems currently under development: 1) One involves the use of a video camera that images the animal while at the same time it is insonified by an acoustic system ( Fig. 5 ). The system is designed for use in situ either on a remotely operated vehicle or mounted on a frame that is cast in the water. The system will provide in situ target strength information and will help to bridge the data collected in the laboratory with those collected by the field survey instrumentation. 2) The other system is currently in an early stage of development and is being designed to measure sound speed in animals in situ. Data from this instrument will help provide one of the key parameters for acoustic scattering models.
Results And Synthesis
As part of our scattering model development, we have made comparisons between theoretical predictions and both laboratory and field data. The laboratory data have involved single broadband echoes (for example, Fig. 6 ) as well as ensembles of echoes where ping-to-ping fluctuations and echo averages (from the same animal or same type of animal) were studied. The field comparisons involved aggregations of a diverse group of animals including copepods, euphausiids, and gastropods. The models developed in the laboratory have been relatively successful in describing the scattering from a complex assemblage of animals ( Fig. 7 ). We have also found biological conditions under which the acoustic scattering levels can be scaled according to ground truth data (VPR in this case) and used without scattering models to interpret the data (Wiebe et al., 1997; Benfield et al., in press).
Acknowledgements
This research was supported by the US Office of Naval Research, National Science Foundation, and the Woods Hole Oceanographic Institution.
|
