The use of sonar systems to look at zooplankton is certainly not new and various researchers have concentrated their efforts on using these tools both for broad survey and size characterization. Currently, two commercial systems, one manufactured by Biosonics (Seattle, WA) and the other by Simrad (Norway) are routinely used in both echocounting and echointegration studies. Although much has certainly been learned from data processed from both of these systems, neither can count or track multiple organisms at high frame rates when target densities exceed even moderate amounts. Our efforts in developing new sonar systems have been dedicated to creating a system which has the capability of resolving, in both space and time, single organism tracks. From a technical point of view, the solution to this problem is clear: multibeam sonar imaging systems with high repetition, or frame, rates.
The system that we have designed to achieve the goal of both finer spatial resolution and more frequent sampling in time utilizes two arrays of eight sonar transducers each, to create a three-dimensional image of "targets" in the field of view of the sonar. One aspect of sonar imaging systems, similar to the type that we designed, is that the two cross range dimensions have fixed resolution in angle rather than distance. Our system, FTV (for Fish TV) has beams which are approximately 2 degrees by 2 degrees. In the third direction, range, the resolution is determined essentially by the "effective" pulse length. We are currently using a pulse which provides a range resolution of approximately 3 cm. The system operates at a frequency of 450 kHz.
At a distance of 4 meters, the system has resolution cells which are about 15 cm x 15 cm x 3 cm. The repetition rate of the system is variable, but can be as high as five frames s-1. Since the system has eight two-degree by two-degree beams, it images a "wedge" of space which is 16 degrees by 16 degrees. The depth of the wedge can also be varied to a range as large as 20 meters. Presently, we have tested the system to a range of four meters; at this distance, a single 0.5 cm sized animal can be localized and tracked in three dimensions.
During the past year, the system has undergone numerous field tests. In one instance, we operated the system mounted on a Phantom IV ROV from the SIO RV Robert Gordon Sproul. The system was utilized to depths of 80 meters. Two deployment modes were used. In one mode, the ROV was raised and lowered in the water column to measure both the amount of scatterers and also their 3-dimensional spatial distribution. In the other mode, the system was kept at a fixed depth so that the trajectories of various targets could be measured. Figure 1 shows the result of processing 42 frames of data and the resultant trajectories of 6 euphausiid-like targets that were tracked simultaneously within the field of view of the system.
We have also been experimenting with some new methods which utilize structured lighting in order to characterize three-dimensional distributions of chlorophyll-a. Although structured lighting has been used before in underwater surveys, our idea to use it to map 3-dimensional distributions of phytoplankton appears to be novel. As illustrated in Figure 2, the technique consists of scanning a stripe of blue light, parallel to the image plane of a sensitive digital CCD camera. The resulting fluorescence induced from the incident illumination is then recorded. The stripe is then moved to a slightly different position, adjacent to the first beam, where another picture is taken, and so on. The resulting set of 2-dimensional images are then processed with an algorithm which computes the chl-a concentration in the 3-dimensional volume by taking into account the attenuation and the emission of light due to both the water and the chl-a.
Computer simulations and laboratory experiments performed with a two component system consisting of water and chl-a, indicate that the technique can resolve chl-a concentrations as low as 0.1 mg m-3 in volumes as large as a cubic meter with a spatial resolution of 1 cm. That is to say, a 1 cubic meter volume can be resolved into volumetric elements of 100 x 100 x 100 one centimeter cells. The primary limitation of the technique is the "low" level of light emission due to the fluorescence when compared with the relatively high attenuation of water in the emmited, or red wavelength. Other setups, which use only scattered light, and not the fluoresced emission may theoretically be used to image larger volumes.
Figure 3 shows the results of a lab experiment designed to characterize the system's capability in measuring chl-a concentration in a 3 component system. As illustrated, unfiltered seawater drawn from the SIO pier (chl-a concentration of 1.06 mg m-3) was injected with a volume of dinoflagellate culture measured at 2.18 mg chl-a m-3. Then, a volume of warm distilled water was floated on top of the seawater. A 12 cm x 12 cm x 12 cm volume in the center of a larger test volume was selected for imaging. Serial section images were collected using the technique described above and then reconstructed. The accompanying 3-dimensional rendered volume (using Sunvision software) shows a projected view of the 3-dimensionally reconstructed volume. As illustrated in the figure and the accompanying table, good quantitative agreement was obtained between samples obtained from the volume and the calibrated samples.
In summary, we have presented two "new" types of imaging systems. We believe that the availability of this technology presents many opportunities in that they provide a view of the ocean that has never before been available. Both systems are aimed at obtaining the three-dimensional information (e.g., relationships of predator to prey and herbivore to food) which is important to the goals of the U.S. GLOBEC program. Perhaps one of the most exciting and new opportunities concerns our plans to use both systems together so that both animal trajectories as well as distributions of phytoplankton can be measured to learn about foraging strategies. (The authors are members of the Underwater Imaging Group (headed by J. Jaffe) of the Marine Physical Laboratory at SIO).
Acknowledgments. This work is a result of research sponsored in part by the National Science Foundation under Grant OCE 89-14300, in part by ONR grants N00014-93-I-0121, N00014-89-J-1419 and in part by NOAA, National Sea Grant College Program, Department of Commerce, under Grant No. NA89AA-D-SG138, project number R/OE-24, through the California Sea Grant College.
Palowitch, A. W. and J. S. Jaffe. 1992. Experimental determination and modeling of volumetric optical properties. Proc. Ocean Optics X1, 1992 Int. Symp. on Optical Applied Science and Engineering, 1750, 104-113.
Palowitch, A. W. and J. S. Jaffe, In press. Three dimensional ocean chlorophyll distributions from underwater serial sectioned fluorescence images. Applied Optics.