IEEE Spectrum August, 2011 - 53

information, we can form a pretty good picture of what the
ocean contains up to about 100 km away.
By taking frequent sonar images of the ocean over time,
we're essentially gathering frames of a video; when we play
it back, it reveals what hundreds of millions of fish are doing
across an area that's hundreds of kilometers across. Modern
computers have no problem handling the signal- and imageprocessing calculations needed to generate these moving
images in real time, so we can watch all the subsurface action
unfold in front of our eyes.
Surveying an area that size with traditional fish-fi nding
sonar would be nearly impossible. You'd have to send out a
fleet of fast survey ships working around the clock. Nobody,
not even the navies of the world, has money for that. That's
a c o N v E N T i o N a l F i s h - F i N d i N g s o N a r w o r ks aT
why after nearly a decade of surveys near the Georges Bank
ultrasonic frequencies, well above the audible range. It directs
fishing grounds, no one in the National
an ultrasonic beam, like a flashlight's, on marine
Marine Fisheries Service ever knew that
life. Such systems can provide high-resolution
herring were amassing on such a large scale
images of things in the immediate vicinity, often
Fish species
there each night during spawning season,
letting you pick out the echoes of individual fish.
Found on the
something we quickly discovered with our
The main drawback of these systems
continental shelf
long-range sonar.
is that they work only at short range, typisouth of New york
cally over distances that are about equal to
the depth of water you're in. To detect things
sT r o N g a u T u M N sTo r M s c u r Ta i l E d
farther away, you need to use lower frequenour 2006 sonar survey of Georges Bank to just
cies. For our sonar system, we use audio fretwo weeks, chilling us with stiff breezes and
quencies that go from roughly the G above
gale-force winds while we were at sea. During
aT L a n T i c H e r r i n g
middle C (392 hertz) to about two octaves
the first week my colleague Purnima Ratilal,
above that (to 1570 Hz).
of Northeastern University, in Boston, noticed
Our fish-sensing sonar requires two
a curious pattern. Each day a shoal of herring
instruments-one to transmit sound waves
would
form at roughly the same place and
aT L a n T i c M a c K e r e L
and one to receive their echoes. The source
time. To prove it to the rest of us, she said she
of the sound waves is a string of what are
could tell us precisely when we would see a
essentially loudspeakers, which hang vertimassive shoal emerge on the northern flank of
cally below the first ship. The speaker array
Georges Bank. At around 5 p.m. that day, she
sends out short bursts of sound that travel in
gathered several of us to the sonar display and
BLacK Sea BaSS
all directions.
said, "Just watch, just watch."
These acoustic waves can go hundreds or
From a sea of acoustic darkness, little
even thousands of kilometers. On this scale
specks of high-intensity scattering began to
the ocean can't be thought of as an infinite
show up on the display, just where she said
DogFiSH
abyss-its average depth is just a few kilothey would be. These specks quickly grew into
meters. And over the continental shelves,
lines, their endpoints spreading about 10 times
the water is only a couple of hundred meters
as fast as even the most athletic herring can
deep at most. So the overall geometry resemswim. With our long-range sonar, we were
bles a puddle, only bigger. Any sound waves
able to show in just a few days that these herreD HaKe
traveling upward or downward quickly hit
ring shoals formed rapidly and like clockwork.
a boundary-the surface or the bottom-
Once the fish reached a critical population
and bounce back. In this way, the thin veneer
density, they began slowly to migrate toward
of ocean horizontally channels the sound
shallow spawning grounds on Georges Bank.
energy we inject into it, guiding the acousThese observations-part of the Alfred P.
tic waves just as effectively as an optical fiber
Sloan Foundation's Census of Marine Life-
ScuP
guides the light sent through it.
proved a fundamental theory about the
The sound waves we send out reflect or
behavior of many animals that travel in
scatter off the objects they encounter, and a
groups. They follow a simple rule: Match the
long line of hydrophones (underwater microaverage velocity of your fellow fish or birds or
phones), towed horizontally behind the seccaribou or whatever, at least the ones you can
S i Lv e r H a K e ( W H i T i n g )
ond ship, pick up these echoes. Careful prosee. When population density is low, the anicessing of the received signals allows us to
mals' spheres of perception don't intersect, so
figure out which direction the echoes are
they don't move coherently. But as the popucoming from and how long the sound waves
lation density increases, they start to notice-
SPoTTeD HaKe
are taking to make the round trip. With that
and to mimic-one another. Chain reactions
It took us years of detailed survey work and analysis to
show the Navy that there was no more than a random correlation between river channels and UFOs, so the former couldn't
cause the latter. This conclusion disappointed many researchers with an interest in studying seafloor geology-and put us
in the hot seat. If not buried geology, what was causing these
strange echoes to form? For a while, we had only our hunches.
But the next time we went to sea with our equipment, we
came prepared: We brought an extra ship with conventional
fish-fi nding sonar to collect hard evidence. And our discovery in 2003 that shoals of fish were responsible for UFOs put
an end to the debate.

u.s. Fish and WiLdLiFe service

spectrum.ieee.org

AuguSt 2011 * iEEE SpEctrum * NA

http://spectrum.ieee.org

Table of Contents for the Digital Edition of IEEE Spectrum August, 2011