The EconomistMarch 14th 2020 Science & technology 732 ond, from a firm called Areté Associates,
has an aluminium shell, and uses wood for
buoyancy. Both models feature solar pan-
els. The third, made by a company called
Numurus, is made of lacquered cardboard,
and relies entirely on its batteries. All three
are designed to last for a year or so and are
made to be as environmentally friendly as
possible, with minimal use of plas-
tics. That is important because, at the end
of their mission, the floats are designed to
scuttle themselves.
Some of the instruments on offer are
common to any smartphone—gpssensors,
accelerometers to detect motion, a com-
pass, a microphone, temperature sensors
and a camera. Others are more directly tai-
lored for the job, such as an underwater mi-
crophone, a gizmo to measure the water’s
conductivity (and therefore its salinity),
and detectors to pick up radar and radio
signals, including transmissions from ma-
rine anti-collision beacons. Some data
from these instruments will be crunched
on board, but most will be sent back to land
in bursts, for onshore analysis. For now,
that connectivity is provided by the Iridi-
um network of geosynchronous satellites.
But the modems necessary to talk to those
satellites, says Mr Waterston, are the most
expensive and power-hungry devices on
the floats. He hopes that new, lower-flying
satellite networks, currently being built by
firms such as Spacexand OneWeb, will pro-
vide cheaper alternatives.
Having lots of different sensors will
help the floats build the best possible pic-
ture of what is going on around them. For
example, if the microphone picks up a
sound at the same time as the accelerom-
eter shows movement, it could mean that a
bird has landed on the float. Several birds
landing on several floats could show how a
flock is moving. Their presence, in turn,
might be an indicator of shoals of fish or
other biological activity.
Similarly, a ship sailing through a float
field will leave all sorts of traces. It might be
detected by its radio beacon, or its radar. It
might sail close enough for a float to take a
picture, or hear it on the hydrophone, or be
disturbed by its wake. Correlating data
from several floats will reveal the ship’s
speed and direction. By building a database
of such encounters, the project’s scientists
hope to learn quickly how to tell different
sorts of craft apart. Fishing vessels might
be using fish-finding sonar or noisy trawl
nets. A giant supertanker will sound differ-
ent from a naval frigate.
The range of sensors on a float will also
produce a mass of data of interest to ocea-
nographers, meteorologists and biologists.
The cameras and microphones on a field of
floats could, for example, detect and track
whales and dolphins. At the moment,
whenever a marine mammal is spotted in
the shipping lanes off Los Angeles harbour,
one of the busiest in America, traffic is
slowed down. Better tracking would allow
traffic to be rerouted, benefiting both crit-
ters and commerce. Float fields could
watch for illegal fishing, smuggling and
icebergs. They could monitor and track oil
spills and algal blooms.
That, at least, is the long-term goal. So
far,darpahas bought around 4,500 floats,
and has tested them only in small num-
bers. The next stage, starting this spring,
will see fields of 1,000 at a time deployed in
the Gulf of Mexico and in the waters off Cal-
ifornia. The plan is to deploy one float for
every three square kilometres of ocean.
The hope is that, as the technology ma-
tures, useful data could be gleaned from
densities as low as one float per 20 square
kilometres. With 361m square kilometres
of ocean on the planet, a true Ocean of
Things, monitoring everything on and un-
der the water, would require about 18m
floats. That will not happen for a while yet.
But Mr Waterston’s plans are a start. 7T
urtles havean unfortunate habit of
devouring plastic objects floating in the
sea. These then get snared in their alimen-
tary canals, cannot be broken down by the
animals’ digestive enzymes and may ulti-
mately kill them. It is widely assumed that
this penchant for plastics is a matter of
mistaken identity. Drifting plastic bags, for
instance, look similar to jellyfish, which
many types of turtles love to eat. Yet lots of
plastic objects that end up inside turtles
have no resemblance to jellyfish. Joseph
Pfaller of the University of Florida there-
fore suspects that something more compli-
cated is going on. As he writes in Current Bi-
ology, he thinks that the odour of marine
micro-organisms which colonise floating
plastic objects induces turtles to feed.
The idea that the smell of plastic flot-
sam might lure animals to their doom first
emerged in 2016. Researchers at the Uni-
versity of California, Davis, noticed that
certain chemicals, notably dimethyl sul-
phide, which are released into the air by
micro-organism-colonised plastics, are
those which many seabirds sniff to track
down food. These chemicals mark good
places to hunt because they indicate an
abundance of the algae and bacteria that lie
at the bottom of marine food chains. The
researchers also found that birds which
pursue their food in this way are five or sixtimes more likely to eat plastic than those
which do not.
Since turtles are known to break the sur-
face periodically and sniff the air when
navigating towards their feeding areas, Dr
Pfaller theorised that they are following
these same chemicals, and are likewise
fooled into thinking that floating plastic
objects are edible.
To test that idea, he and his colleagues
set up an experiment involving loggerhead
turtles, a species frequently killed by plas-
tic. They arranged for 15 of the animals,
each around five months old, to be ex-
posed, in random order, to four odours de-
livered through a pipe to the air above an
experimental arena. The odours were: the
vapour from deionised water; the smell of
turtle-feeding pellets made of shrimp and
fish meal; the smell of a clean plastic bottle
chopped up into ten pieces; and the smell
of a similarly chopped bottle that had been
kept in the ocean for five weeks to allow al-
gae and bacteria to grow on it.
Two of the smells proved far more at-
tractive to the animals than the others.
When sniffing both the odour of food pel-
lets and that of five-week-old bottles tur-
tles kept their nostrils out of the water
more than three times as long, and took
twice as many breaths as they did when
what was on offer was the smell of fresh
bottle-plastic or deionised-water vapour.
On the face of it, then, the turtles were re-
sponding to the smell of old bottles as if it
were the smell of food.
Though they have not yet tested wheth-
er dimethyl sulphide is the culprit, Dr
Pfaller and his colleagues think it is the
most likely candidate. In an unpolluted
ocean, pretty well anything which had this
smell would be edible—or, at least, harm-
less. Unfortunately, five-week-old plastic
bottles and their like are not. 7Why turtles eat plastic rubbishEcologySiren smell
Foul smells fair