The Economist January 29th 2022 Technology Quarterly Defence technology 7
vering 16 squarekilometresata resolutionofonemetre—good
enoughtomakeouta car—for$500a timeundera CreativeCom
monslicencewhichletsbuyersdoastheypleasewiththeproduct.
Thecurrentpriceofanopticalimageofequivalentgranularity
wouldbearoundsixtimesasmuch,andbesoldunderrestrictive
conditionsthatlimithowitcouldbedisseminatedandused,he
says.Suchcompetitionshoulddrivedownthepriceallround—
andincreasethescopeforaddingallsortsofvalue.
Hugeamountsofsatelliteimagesatrelativelylowpricespro
videsa prettymuchidealinputforthesortofmachinelearning
algorithmsthathavepoweredrecentadvancesinai: lotsofwell
structureddatawithwhichtolearnhowtorecogniseallsortsof
differentthings.Suchautomationallowsbothnewinsightsand
analysisatscale.MrMaherofUrsasaysthathisfirmmightcheck
inon20,000facilitiesoncea weekfora singleclient.“It’slessthan
a fulltimejobforonepersontomonitorthosesites,”hesays.
Thecapabilitiesofsararejustoneexample,albeita dramatic
one,ofthesortofprogressbeingmadewithallsortsofadvanced
sensorsandthedataprocessingsystemsneededtointerprettheir
output.Satelliteprovidersofopticalimages,hyperspectraldata
andradiolocationserviceswhichkeeptabsonaircraftandships
arespringinguparoundtheworld.Theabilitytouseradartotrack
movingobjectsfromspaceinrealtimeisgettingcloser.Andpretty
mucheverythingwhichcanbedonefromorbitcanalsobedone
withdronesclosertothetargetandcapableofrespondingtonew
requestsalmostinstantly.Therearefewplaces,ifany,onthesur
faceoftheEarththatarefreefromsuchsnooping.Belowthesur
faceoftheseas,though,thereisstilla redoubt—fornow.n
See-through seas?
M
aritime mysteriesare not unknown on the ragged western
flank of the British Isles, buffeted by cruel seas and bitter
winds. But the tiny stranded vessel reported to the coastguard on
Tiree, the most westerly island of the Inner Hebrides, in October
2020 was particularly mysterious. It bore no markings, broadcast
no identifying signal and carried no running lights.
Once pictures showing its distinctive surfboard shape and so
lar panelling were put online, though, it was quickly identified as
a “Wave Glider”—an uncrewed surface vessel (usv) built by Liquid
Robotics, a California company owned by the aerospace giant Boe
ing. First developed to listen to humpback whales, the vessels’
ability to carry sensors slowly through the seas (they have an inge
nious way of turning the rise and fall of waves into forward mo
tion) has seen them put to all sorts of uses by researchers and na
vies alike; over 500 have been sold to date.
Whales are not the only things they listen to. When contacted
about the Wave Glider found in Tiree, and others that have washed
up on the coasts of Lewis in the Outer Hebrides, Unst in Shetland
and Donegal, Liquid Robotics simply said it would let the owners
know. But it is a matter of record that in 2016 Wave Gliders were
used to detect and track a Royal Navy submarine in an exercise;
and it is past those northwestern coasts that Britain’s nuclear
armed submarines (ssbns) make their way from their base at Fas
lane, near Glasgow, to the open ocean.
The idea that submarines loitering in the depths are undetect
able is fundamental to modern nuclear deterrence. America, Brit
ain, China, France, India, Israel and Russia act on the basis that
though a nucleararmed adversary could conceivably destroy
their landbased forces in a first strike, it could not wipe out their
submarines. A submarine at depth cannot be seen from afar—day
light is all but undetectable at depths of more than a couple of
hundred metres. Radio waves fare even worse, making underwa
ter radar a nonstarter. Sound carries, but the boats can be remark
ably silent; the noise made by a modern ssbnis less than a mil
lionth of the racket produced by the first such boats.
That leaves active sonar, which rather than listening for noises
made by its quarry emits sounds designed to bounce off them.
This lets an adversary locate a submerged submarine precisely
enough to attack it. But it is limited in its range, and by the ways
sounds are distorted and dissipated by changes in pressure, tem
perature and salinity.
This is all taken to mean that weapons on ssbns can be relied
on for retaliation if a first strike devastates everything else. That
guaranteed capability to respond is held to keep deterrence stable.
From surfboards to Red October
America’s most recent review of nuclear policy concluded that
“There are no known, nearterm credible threats to the survivabil
ity of the ssbnforce.” But what is nearterm, and for that matter
credible, is a matter of opinion. In 2020 a panel of experts assem
bled by the National Security College of the Australian National
University concluded that, “The oceans are, in most circumstanc
es, at least likely...to become transparent by the 2050s.” Trends
making it harder to hide above the waves—more numerous and
more capable sensors and more powerful ways of sifting through
their output—are at play underwater, too.
To put a submarine at risk you must first detect it, then track it.
Since the 1980s America and its allies have relied on the hydro
phone arrays of the Fixed Distributed System (fds) for such detec
tion. The sensors, tethered to the sea floor, float at a “critical
depth” where acoustical geometry shields them from extraneous
noise and makes the faint sounds of a submarine passing over
head stand out. America now has fdssystems in both the Atlantic
and the Pacific, with ships, submarines and aircraft available to set
about tracking anything they detect.
Owen Cote, a submarine expert at mit, says that such listening
arrays remain, for now, the only truly effective method of detect
ing a very quiet submarine. Their drawback is that, being quite
short range, they cannot cover whole ocean basins. They just cover
choke points, such as the gaps between Greenland and Iceland and
Iceland and Britain, or the entrances to the Philippine Sea.
Arrays that can be towed behind ships are crucial to tracking
what has been thus detected. If they could be made far more nu
Hiding submarines may get harder
Finding Captain Nemo