May 2022, ScientificAmerican.com 53salmon can tolerate soy, which in large quantities can
damage their gut and immune system.
Experimental fish feeds aim to solve this problem
and address yet another concern: that waste from fish
pens will pollute natural waterways. New “low-pollut-
ing” feeds contain far less nitrogen and phosphorus and
provide nutrients in a form that is easily digested and
absorbed to minimize the amount of these elements
excreted by fish. Emerging feeds include algae oil, yeast
protein and black soldier fly larvae, commonly known
as maggots. Scientists are also designing proteins with
amino acid compositions similar to those in feed made
of fish but created from sawdust and other forest resi-
due plentiful in the sylvan state of Maine. Whether algae,
maggots or reconstituted sawdust will ultimately replace
fish in the aquaculture food chain is unclear, but scien-
tists agree that these experimental feeds show substan-
tial promise. Still, growing fish in pens in the open sea
poses risks that some believe outweigh the benefit.
Giant RAS tanks are meant to minimize several of the
environmental threats posed by net pens. The approach
dates back to the 1950s, but recently scientists and engi-
neers have supercharged the technology into a mind-bog-
glingly efficient protein-generation machine capable of
growing more fish in less time and space than is possi-
ble in nature. In contrast to offshore net pens, land-based
RAS are designed to maintain tight control over water
quality, temperature, circulation and other environmen-
tal variables. Bacteria convert ammonia excreted by the
fish into typically harmless nitrate; gas-exchange systems
add oxygen and remove dissolved carbon dioxide; and
wastewater is microfiltered and zapped with ultraviolet
light to remove phosphorus, lingering bacteria and
viruses. Some of the newest systems feature machine-
vision sensors and artificial intelligence to optimize feed-
ing by monitoring fish biomass and behavior in real time.
Several multinational RAS companies are currently
vying for a stake in Maine. Notable among these is Nor-
dic Aquafarms, a Norwegian firm that has proposed a
major tank facility in the small coastal city of Belfast. Nor-
dic’s $500-million plan to generate nearly 73 million
pounds a year of Atlantic salmon would make it the sec-
ond-largest RAS salmon farm in the world. The largest
is an operation run by Atlantic Sapphire, another Nor-
wegian-owned firm based in the unlikely location of
Homestead, Fla. That facility has suffered a series of cri-
ses, including a massive “mortality event”—elevated gas
levels that wiped out 500 tons of salmon. In the first half
of 2021, the company reported losses of more than
$50 million and faced accusations of criminal ani-
mal cruelty.
Nordic Aquafarms founder and president Erik Heim
insists that his company will avoid these problems, but
researchers are cautious. Jon Lewis worked for 23 years
as a scientist and diver, and he recently retired as direc-
tor of the Division of Aquaculture in Maine’s Depart-
ment of Marine Resources. “Land-based systems have
real advantages in that they don’t rely on Mother Nature
for water treatment,” he says. But even if the filtration
systems remove most of the discharged waste, he adds,
some still flows to the bay. “Given the scale of these sys-
tems, that’s significant.”
At the Belfast facility, salmon eggs will be incubated
in specially designed hatching cabinets, and the emerg-
ing baby fish, known as fry, will spend the first few weeks
of their lives eating their yolk sacs before entering a quar-
antine tank. The young fish will then move to a series of
tanks that simulate the natural life cycle of salmon—
swimming from fresh water to increasingly salty water
over a period of 18 to 20 months, when they reach mar-
ket size. The system circulates 5,200 gallons of water a
minute: 4,700 gallons of salt water sucked from the Gulf
of Maine and 800 gallons of fresh water pumped from
groundwater wells and aquifers. Filtered waste will be
converted into animal feed or fertilizer—in theory
enough to fertilize about seven and a half pounds of
vegetables per pound of fish produced.
Unfortunately, pumping and filtering all that water