The Future of New England Seafood

Sounding out on fish assessment technology

There’s been some buzz lately from luminaries like Senator John Kerry and Legal Seafoods’  Roger Berkowitz about new methods for measuring fish populations using acoustic remote sensing that are being pioneered by a group of scientists from MIT and Northeastern University and that may have the potential application for dramatically improving fisheries science and management. Talking Fish had a chance to find out more about this initiative through a visit last week to the group’s lab – and it was very exciting!

Getting real time, accurate fish population data is a weak spot in fisheries management science. Fisheries data coming from fishermen has problems with quality control and bias; fisheries data from the government’s trawl surveys is expensive, episodic, and selectively samples small swaths of the ecosystem. The data issue is nonetheless hyper-critical since the fisheries stock assessment models on which all fishery management decisions are based, even the best ones, are only as good as the data inputs.  Given data limitations, the current ability to cost-effectively and accurately understand what is happening to fish populations in response to management changes and environmental or other factors is characteristically uncertain.

Many fishermen say that the scientists don’t see as many fish as the fishermen do. There are legitimate and sound reasons why scientists and fishermen see different abundances of fish; it doesn’t mean the science is “bad.” In fact, it is quite good in New England. The trend lines for stock biomass are accurate in projecting relative upward or downward changes in overall stock abundances. But these differing perspectives as to what fishermen and scientists “see” out on  the water does produce a confidence gap that is a major problem and science uncertainties force managers to set lower catch limits than might otherwise be possible.

A visual acoustic image of a herring shoal forming near sunset on October 3, 2006. The top panel shows that 1 hour and 45 minutes before sunset, no shoal is present, and the bottom panel shows that 10 minutes after sunset, a large shoal has formed (click to enlarge). (Credit: Jagannathan et al., "OAWRS of marine ecosystems," Marine Ecology Progress Series 395: 137-160, 2009)

The MIT/Northeastern work, led by their Mechanical Engineering, and Electrical and Computer Engineering Departments, respectively, promises a new, potentially cost-effective approach to developing large quantities of practically real time spatial and temporal data on fish populations at ecosystem scales. They are using a technology they call Ocean Acoustic Waveguide Remote Sensing (OAWRS). OAWRS involves sending out relatively low frequency sound waves horizontally—either from equipment based on research vessels or on fixed underwater stations—and then using the echoes that come back from fish to form a visual acoustic image.

The strongest returns come from echoes from a fish’s gas-filled swim bladder (an internal organ many fish have that allows them to maintain buoyancy), which acts like a mini-drum in response to the sound-waves.  This allows the research team to “see” groups of fish in near real-time over a large area, some 10,000 square kilometers per station. The OAWRS team claims that this technique may even be able to differentiate between different types of fish and identify fish without swim bladders like flounders. They liken the differences between data received from existing sonar and trawl research and data received from OAWRS methods to the differences between seeing one pixel on a television screen versus seeing the whole movie at once. The sound waves essentially “illuminate” a huge underwater landscape so that the fish can be detected over a very large area, instead of in very limited areas as with current methods.

The OAWRS technology is already proving its usefulness for determining real time estimates of  fish populations that have swim bladders and school together periodically, like Atlantic herring, and perhaps equally valuable, the technology is also revealing behavioral characteristics for some of the species they have examined that enrich, if not challenge, the conventional wisdom on fish behavior. And the technology is relatively cheap and can be added on to existing research platforms.

During a “side-by-side” experiment in 2006, herring research vessels equipped with conventional downward-looking sonar and survey nets and a vessel equipped with the OAWRS system both traversed the same area of the ocean on the western side of Georges Bank, trying to sample and estimate the populations of Atlantic herring in the area. The OAWRS technology characterized a massive shoal of Atlantic herring that was the size of Manhattan Island, comprised of approximately 250 million herring. By watching the conventional survey vessels move through this same massive shoal of fish on their computers, they were able to document how the traditional sampling techniques actually missed most of this shoal even though the vessels were in very close proximity. In fact, in order to further calibrate what the OAWRS equipment was seeing with some net surveys, the MIT/Northeastern team had to direct the herring research trawl vessel to the shoaling fish based on their pinpoint knowledge of where the fish were in time and space before the trawl vessel could actually get on top of the more dense concentrations.

Map showing a single OAWRS transmission (red circle) and the line transects of a NMFS trawl survey (click to enlarge). (Credit:Jagannathan et al., “OAWRS of marine ecosystems,” Marine Ecology Progress Series 396: 137-160, 2009)

These new technologies to assess stock size and improve understanding of fish behaviors should be strongly supported. OAWRS “sees” about a million times more of the ocean at any point in time than conventional sensing “sees.” This new tool, when coupled with the region’s excellent fisheries modeling, may reveal that populations of the fish that are being managed are significantly larger or smaller than currently estimated. Regardless, these methods certainly add tremendous value to the models and the stock assessment process. At the same time,  certain applications of OAWRS technology are still in the “proof-of-concept” stage, and Talking Fish remains cautious about claims that OAWRS can distinguish one similar species from another (particularly those species without swim bladders).

However, with respect to estimating populations of fish like Atlantic herring, which are the primary food for very important populations of larger predator fish like Atlantic cod, this technology may now be a component of the best available science – the use of which is required by the Magnuson-Stevens Act—for stock size assessments, so we should figure out how to get it deployed. And acoustic remote sensing technologies like OAWRS certainly may be the only technologies that would make large-scale ecosystem fisheries management possible.

Good conservation, just like healthy fisheries, is dependent on good science. Anything that allows scientists to “see” the dynamics and numbers of fish populations more accurately and more cost-effectively over the scale of the Gulf of Maine serves both ends. Thanks for the tip, Senator Kerry and Roger.

To learn more about OAWRS and the MIT/Northeastern researchers’ work on acoustic remote sensing, visit the MIT lab website, the Northeastern lab website, and read this article in the Marine Ecology Progress Series.


Talking Fish reserves the right to remove any comment that contains personal attacks or inappropriate, offensive, or threatening language. For more information, see our comment policy.

Leave a Reply

Your email address will not be published. Required fields are marked *