What Subsea Research Brought to Fish Farming

Most of the technology inside a modern fish farm was developed for something else. Offshore oil and gas paid for the vehicles, ocean science paid for the sensors, and aquaculture inherited both at a fraction of the development cost.

What transferred directly

Small ROVs came across almost unchanged. An observation-class vehicle built for harbour inspection works fine inside a net pen, and the aquaculture industry adopted them for net and mooring inspection, fish monitoring and installation work without needing anything redesigned.

Oceanographic sensors transferred equally cleanly. Dissolved oxygen, temperature, salinity and current profiling instruments developed for research moorings do the same job on a farm site. The measurement problem is identical. Only the reason for measuring changed.

Acoustic positioning came over too, with USBL systems built for offshore construction now tracking vehicles around net pens.

What had to be rebuilt

Navigation. This is the interesting one, because the assumption that broke was so fundamental nobody had noticed making it.

Subsea navigation assumes a rigid reference. A Doppler velocity log measures speed by bouncing sound off the seabed, and the seabed does not move. Inside a net pen there is often no seabed within useful range, and the surface you actually care about is a flexible net that moves with the current.

Work on hydroacoustic instruments for net pen navigation had to establish which sensors could track against the net itself, and how to combine DVL and USBL data to get reliable autonomous coverage. That research had no offshore equivalent to borrow from.

Scale broke things in the other direction. Offshore vehicles are built for standoff distances measured in metres. Inside a pen, everything happens within touching distance of a net that will entangle a thruster, and around animals whose behaviour changes when a vehicle approaches.

The fish are part of the environment

No offshore discipline prepares you for this. A survey vehicle on a pipeline does not affect the pipeline. A vehicle in a net pen changes fish behaviour, and stressed fish stop feeding, grow more slowly and become more susceptible to disease.

Research on farmed Atlantic salmon avoidance responses to artificial sound and light exists precisely because the instrument is part of the system it measures. A thruster that spooks a pen full of salmon has cost the farmer more than the inspection was worth.

That constraint pushes toward quiet, slow, resident systems that fish habituate to, rather than the powerful, fast vehicles offshore work favours.

Data flowing the other way

Aquaculture now generates subsea datasets at a volume ocean science rarely matches, because a farm site is instrumented continuously for years while a research mooring is a funded campaign.

SINTEF Ocean's SOLAQUA dataset, covering aquaculture robotics, is an example of that flow reversing. Long-duration imagery and sensor records from working farms feed algorithm development that then benefits other subsea sectors.

Biofouling is the clearest case. Farms deal with it continuously, at scale, with real consequences for measurement accuracy. Antifouling strategies developed for aquaculture instruments apply directly to any long-deployment ocean sensor, and that knowledge moved from industry to science rather than the usual direction.