We conduct acoustic trawl surveys, combining midwater and bottom trawling with sonar technology. These surveys assess the distribution and abundance of walleye (Alaska) pollock and other fish and invertebrate species in Alaska in winter and summer. Our survey data are critical for the Eastern Bering Sea, Gulf of Alaska, and Bogoslof Island Region stock assessments. They are also used in the Ecosystem Status Reports and Ecosystem and Socioeconomic Profiles for these areas. Our work supports the Eastern Bering Sea pollock fishery, the largest single commercial fishery in the United States, and communities in the Gulf of Alaska who rely on the pollock fishery. Our data also contribute to an understanding of rockfishes, krill and other forage species in the Bering Sea, Gulf of Alaska, and Chukchi Sea ecosystems. We lead the Acoustics on Vessels of Opportunity project, which uses acoustic data collected by chartered fishing vessels during the Groundfish Assessment Program bottom trawl survey to calculate an index of pollock abundance for the Eastern Bering Sea stock assessment.
We conduct cutting-edge research on acoustic technologies, including sonar instruments on uncrewed ocean-going robots and stationary moorings. Mobile uncrewed systems, whether wind-powered like Saildrones or diesel-powered like DriX, can collect data at times or in locations that research ships are not present. These uncrewed systems also have the potential to serve as force-multipliers for existing data collection platforms and vessels. Fish are less likely to avoid approaching uncrewed vehicles as compared to large vessels, leading to more accurate density estimates. Moorings deployed on the seafloor provide long-term information about fish movement at specific locations, and have been used to measure pollock migrations across the U.S./Russia maritime boundary.
Optical system research and development is a core part of our work, and we are a world leader in this area. We design, build, test, and use camera systems that are:
Deployed in trawls (e.g. CamTrawl)
Deployed on commercial fishing gear (e.g. longline fishing gear and associated with catch protection devices)
Towed above the seafloor (e.g. CatCam for untrawlable habitat, CoralCam for coral habitat)
Lowered into acoustic targets in the water column
Mounted on remotely operated vehicles or autonomous underwater vehicles
Our camera systems and software collect stereo images to allow measurements of fish size, or video to enable continuous observation. We work to understand the utility, and potential limitations, of using optical data, including whether fish avoid or are attracted to the equipment.
Evaluating the performance of trawl gear is another key area of work. We deploy small recapture nets installed on the outside of a trawl to estimate what species and sizes of fish and invertebrates are escaping the net before reaching the codend. By quantifying which fish are not fully accounted for in our trawl samples, we can correct for potential biases by species or length in our catch information. We have a history of work on observations of fish behavior in trawl gear, and the development of excluders which enable non-target fish species to escape from the net during fishing.
Much of our at-sea hardware and software for data collection are designed and built in-house. The Catch Logger for Acoustic Midwater Surveys data acquisition system and Ichthystick measuring board are both used across NOAA Fisheries Science Centers, allowing the collection of detailed and accurate biological information. The CamTrawl Data Acquisition software, originally developed for our CamTrawl system, now runs many of the cameras we developed. It supports data acquisition by other Science Centers and collaborators.
Our processing of acoustic and optical data covers a wide variety of approaches. Fish and invertebrates reflect acoustic energy differently across acoustic frequencies (18, 38, 70, 120, and 200 kHz are used on research vessels). These differences are used to measure krill abundance during acoustic surveys. When fewer frequencies are available, such as on an uncrewed vehicle, this method can be adapted to having less information by applying a Machine Learning method. We are developing the use of acoustic signals over a broader frequency range to assess its potential to better identify sound-scattering animals. We have also developed an innovative approach to the long-standing challenge of using reflected acoustic energy to identify species or groups. This method is probabilistic as it allows us to assign an identity and abundance to likely species/groups from acoustic and other data, while providing a sense of how certain those classifications are. Finally, we are developing machine learning methods for identifying and tracking fish in video and stereo camera data from trawl-mounted camera systems. Machine Learning methods are also being used to classify seafloor habitat type with acoustic data.
We have a long history of supporting the NOAA Fisheries’ Teacher at Sea Program. The first Teacher At Sea sailed with us in 1994, and we continue to host teachers on many of our surveys. The benefits of our partnership with the Teacher At Sea Program go both ways, and we continues to learn from (and have fun with!) the Teachers that sail with us.
Contact: Sandra Parker-Stetter, Program Manager, Midwater Assessment and Conservation Engineering
