If you think U.S. Census Bureau officials have their hands full locating and counting everyone who lives in the United States, consider the men and women at the National Oceanic and Atmospheric Administration’s National Marine Fisheries Service (NMFS) who are charged with counting fish in U.S. waters.
Their mission is to prevent the decline of fish stocks and the loss of habitats that fish need to breed, spawn, feed, or grow. But a whole lot of water gets in the way of their ability to locate those essential habitats and to estimate fish stocks.
In recent years, scientists at Woods Hole Oceanographic Institution have been lending a hand, applying new technologies and methods to see into the sea and help assess and maintain fisheries.
How slow can you go?
For example, to survey seafloor populations of rockfish, NMFS has relied on sampling with bottom trawls. The method has inherent limitations.
“Rockfish live on rocks, and trawls get stuck on rocks,” said Elizabeth Clarke, director of the NOAA Fishery Resource Analysis and Monitoring Division in Seattle, Wash. “We have lots of untrawlable areas, and our bottom trawl surveys cannot go into those areas.”
Another option, sonar, doesn’t work well, because it is hard to distinguish between sound signals reflected off rocks and off fish buried amid rocks. So NMFS tested another alternative: SeaBED, the underwater autonomous vehicle developed at Woods Hole Oceanographic Institution (WHOI) by Associate Scientist Hanu Singh and colleagues in the WHOI Applied Ocean Physics and Engineering Department.
SeaBED flies about 8 feet (2.5 meters) above the seafloor, flashing its strobe light and snapping a photo every three seconds. It maintains a constant altitude and speed of a ½ knot.
“Most people ask us, ‘How fast do you go?’ Singh said. “We’re proud about how slow we can go.”
A treasure trove of seafloor images
In 2005, NMFS dispatched SeaBED and Singh’s team to survey seafloor fish proposed for protection off the coasts of Oregon and California. SeaBED collected 30,952 high-quality photos that WHOI scientists blended together into large photomosiacs of rockfish habitats.
“The images were great,” Clarke said. “The fish were easy to identify. We saw invertebrates tucked in all around the rocks, as well as the fish. The photomosaics give us a large perspective, so we can see the relationships between geological features, fish, and invertebrates.”
To store, manage, and analyze this overflowing treasure trove of images, Vicki Ferrini, data manager of the National Deep Submergence Facility at WHOI, created FISH_ROCK, a computer software tool that gives scientists the ability to identify and count organisms in digital seafloor images. Using FISH_ROCK, scientists can annotate each image with information about the numbers and kinds of fish, invertebrates, and seafloor characteristics (boulders, sand, pebbles, mud, etc.). The software also marries SeaBED’s precise navigation data to each photo. So with a few keystrokes, scientists can automatically search and sort the photo database—to count the number of specific rockfish such as hake or bocaccio in a given area, for example, or determine the particular environment they prefer.
“We’ve never had anything like this tool before,” Clarke said. “It’s a lifesaver that is essential to getting our work done.”
Clarke said SeaBED is cost-effective. It can cover a lot of territory, photographing 3 to 6 miles (5 to 10 kilometers) on a typical six-hour dive. It runs autonomously, so that a research vessel can simultaneously conduct other research while SeaBED completes its mission. And it is controlled by a laptop computer, so SeaBED can also be used on less-expensive vessels, such as fishing boats.