Despite the importance of understanding the role of animal behavior in submarine ecosystems, ecologists who study these systems have very few tools for observation. The FTV systems are 3-dimensional acoustic “imaging” systems that permit observation of reflective targets in their native environment at frame rates up to 4 frames per second. Through several cruises (British Columbia, Canada and the Gulf of Aquaba, Eilat, Israel) we have recorded the trajectories of, literally, hundreds of thousands of zooplankton. Analysis of the data has given us a “glimpse” of their world. The FTV systems that operate at frequencies of 435 kHz and 1.6 MHz have been under development in our lab for over a decade. Results demonstrate how these systems can be used to infer animal behavior and its consequences for the ecosystems.
The FTV 435 and 1.6 high frequency sonar imaging and tracking systems are multibeam acoustic sensing systems designed to infer the location and hence trajectories of marine animals; primarily zooplankton. The 435 system can record the reflections from animals as small as 7 mm at ranges as great as 7 m. The 1.6 MHz system is designed to track even smaller animals: 1 – 2 mm copepods, however can only achieve this amazing performance at ranges less than 3m. Given that these animals only reflect somewhere between 1 millionth and one billionth of the sound incident upon them makes this a difficult task.
The OASIS system was designed to record optical images of animals from which acoustic reflections were being recorded. The OASIS system therefore consisted of FTV, a sensitive CCD camera and a strobe light. The FTV system monitored the location of a specific 3-dimensional spot called the “magic voxel”. When an animal entered this location the sonar triggered the camera to take a picture of the animal. Thus, exact correspondence between animal type, size, and orientation were correlated with the acoustic reflection.
FTV 435 and 1.6 use a set of 16 rectangular transducers arranged in two groups of eight in order to resolve animal location (See photo (38 kb) of the system). Eight of these transducers are used as transmitters and the other eight transducers are used as receivers. The transducers measure 9.07 cm. x 0.907 cm. x 0.5cm thick. The field of view of the transducers in the 435 system are 2 degrees by 20 degrees. In the 1.6 systems they are .7 degrees x 6 degrees. Considering all of the transducers gives a field of view of 16 degrees by 20 degrees in the 435 kHz system and 6 degrees by 8 degrees in the 1.6 MHz system. The systems are 8 x 8 x 512 (range bin) imaging systems. Animal location is inferred by looking at the acoustic reflectivity of targets within each beam as a function of range. If the reflected sound exceeds some predetermined level, a target is judged to be present.
In order to understand the overall performance of the system the relationships between the individual transducers must be specified. In FTV, each of the set of 8 transducers is arranged in a spiral like pattern, where each transducer is stacked upon the other and also pointed into a slightly different direction. In order to achieve a contiguous field of view, with no spatial holes, the transducers are clocked at a two-degree angle relative to each other. Thus, if all of the 8 transducers were either transmitting or receiving together, the field of view (FOV) would be approximately 16 degrees by 20 degrees. In order to achieve specified system performance, the two arrays of 8 transducers each are placed next to each other, but rotated 90 degrees with respect to each other.
While all of the receivers are operated simultaneously, that is, they are all digitizing at the same time, each of the transmitters is operated one at a time, sequentially. Viewing each resolution element as one square of an eight x eight graph, FTV forms a 3-dimensional sonar image by sequentially transmitting on each of the rows, and concurrently receiving on the entire set of columns. In this way, the entire FOV is “built up” for each image from the set of ensonified rows and received columns. Presently, the delay time between each transmission can be as short as the round-trip travel time of the sound. This allows an entire 3-dimensional image to be collected in .25 seconds, or at a rate of 4 Hz. Since the fast data storage of the current set of digitizing hardware is up to 2048 samples in length, the system creates a 3-dimensional matrix of backscattered sound with dimensions 8 x 8 x (up to) 2048 samples. In fact, since each of the digitizing channels uses a quadrature demodulator with a real and imaginary component, and each value is 16 bits long, the system produces 512 Kbytes of image data with each operating cycle. Thus, images can be collected at this rate until hard disk storage is exhausted.
- Principal Investigators: Jules Jaffe
- Chief Engineer: Edward Reuss
- Mechanical Engineer: Fred Ulhman – Hydrophone fabrication, chassis
- Software Development: Ed Reuss – Data acquisition, instrument control, GUI
- Lab Assistants: Mathew Diebolt
- Graduate Students: Duncan McGehee, Alex De Robertis, Chad Schell