Industries / Food, farming and forestry / Fishing and aquaculture
Keeping the catch cold
Keeping the catch cold is the work of chilling or freezing fish from the moment it is caught until it is landed, using ice, chilled seawater or freezers on board and at the landing site.
5
Milestones achieved, from About 1840 to 2011.
2
To watch for: announced, not yet confirmed outside.
2
Open deliverables. Nobody has shipped them.
The road
AchievedWatch forOpen
Achieved
About 1840
Ice taken to sea
Samuel Hewett (1797–1871) of Barking, England, packed the catch of his Short Blue fishing fleet in ice so it stayed fresh longer. His boats could then fish grounds farther away, while fast boats carried the iced fish to market. Ice came from Norway or from Barking marshes flooded each winter. A local history manuscript dates the start of ice use to about 1840.
Credit
Samuel Hewett, whom the Thames Festival Trust calls "the first person to use ice for the preservation of fish at sea."
1954
Freezing the catch at sea
The stern trawler Fairtry (see Vessels and gear), completed in Aberdeen on 29 April 1954, quick-froze its catch on board, so fish no longer had to reach port before the ice ran out. Its launch report called the ship "the first to be specially designed and built for this service." Factory freezer ships soon followed in other countries.
Credit
John Lewis & Sons (builder); Christian Salvesen (operator)
2003
Chill to near 0 °C
The Codex Alimentarius Commission, the joint FAO and World Health Organization food standards body, adopted its Code of Practice for Fish and Fishery Products. It says chilling should start "as soon as possible" and that fresh fish should be held "as close as possible to 0 °C," in shallow layers surrounded by finely divided melting ice, or in chilled or refrigerated seawater.
Credit
Codex Committee on Fish and Fishery Products
2004
Hold temperatures become law
EU Regulation (EC) No 853/2004 requires fish to be chilled as soon as possible after it comes on board. Tanks of refrigerated seawater must bring the fish and water to no more than 3 °C six hours after loading and no more than 0 °C after 16 hours, and freezer ships must keep holds at −18 °C or colder with a temperature recorder.
Credit
European Parliament and Council of the European Union
2011
A third lost, roughly counted
An FAO report estimated that 35 percent of aquatic food is lost or wasted worldwide. WorldFish researchers note that the estimate assumed all losses came from wild catch and was not based on direct measurement.
Lessonthe scale of loss is still an estimate, because few landing sites measure it.
Credit
FAO
Watch for
Solar ice for tuna fishers
The International Pole & Line Foundation (IPNLF), the German development agency GIZ and partners put a container-sized, solar-powered ice maker in the village of Kawa, Maluku, Indonesia, in late 2024; before that, fishers used home freezers or travelled far for ice. The partners report 176 tonnes of ice in the first year and 4,290 litres of diesel not burned, and say the quality of landed fish improved.
Achieved whenan outside evaluator publishes before-and-after measurements of fish quality or loss at Kawa, or similar units run at other sites without project support.
Claimed by
IPNLF and GIZ, October and December 2025
Timing each fish to ice
Tryolabs and The Nature Conservancy fitted a tuna longliner in the eastern tropical Pacific with RFID readers (radio tags read without contact) at the deck and the ice-well hatch, plus a temperature logger in the ice well. On two trips in late 2025 the crew tagged 39 tunas, and the system recorded each fish's time from deck to ice and the ice well's temperature automatically. The report says the logger reads air, not fish, and that the scores still need checking against quality grades at landing.
Achieved whenan outside party shows the onboard scores match quality grades at landing and buyers use them routinely.
Claimed by
Tryolabs and The Nature Conservancy, April 2026
Open
01
Fish temperature from air readings
a tested conversion, as a small model and a printed lookup table, that turns readings from a cheap air-temperature logger in an ice well or cool box into an estimate of the temperature at the centre of the fish.
The briefCredit · unclaimed02
Fish temperature from thermal photos
a phone method that estimates the centre temperature of a whole fish at landing from a thermal-camera photo of its surface plus its size, without pushing a probe into the flesh.
The briefCredit · unclaimed
Open 01 · the brief
Fish temperature from air readings
The first to publish is credited on this line, and the deliverable moves from Open to Achieved.
What to build
a tested conversion, as a small model and a printed lookup table, that turns readings from a cheap air-temperature logger in an ice well or cool box into an estimate of the temperature at the centre of the fish.
Why it is still open
the 2026 Tryolabs and Nature Conservancy pilot found its ice-well sensor "measures air temperature, not the temperature of the fish in the ice," and named comparing onboard readings with landing grades as the next step. Iceland's Matís and others have logged fish temperatures in transport boxes, but searches found no published calibration for whole fish under a single air logger in an onboard ice well.
Built with what exists
two or three waterproof air loggers (for example Elitech RC-5 or Onset HOBO), a logging thermocouple probe, a household cool box, shop ice, whole fish bought at a market, FAO's ice-to-fish ratios (1 to 1 as a minimum in the tropics), and Python with pandas and scikit-learn to fit the model.
First step
run one pilot trial. Put a probe in the thickest part of one whole fish, tape a logger to the box wall above the ice, record both every 5 minutes for 6 hours, and open the lid twice.
Done looks like
at least 20 recorded trials covering 2 species, 2 fish sizes, ice ratios of 1 to 1 and 3 to 1, and lid openings; the dataset and code public; a published error table for trials held back from fitting, with 90 percent of estimates within 1 °C of the probe as the target.
Open 02 · the brief
Fish temperature from thermal photos
The first to publish is credited on this line, and the deliverable moves from Open to Achieved.
What to build
a phone method that estimates the centre temperature of a whole fish at landing from a thermal-camera photo of its surface plus its size, without pushing a probe into the flesh.
Why it is still open
FAO guidance says fish should be between 0 °C and 2 °C when landed, and checking that means probing fish one at a time. The 2026 Tryolabs and Nature Conservancy pilot reports it is still seeking "additional sensors and more accurate proxies" for fish quality on board. Searches found thermal-camera work on beef carcasses and core-temperature work on cooked fish cakes, but no tested method for whole chilled fish.
Built with what exists
a phone thermal camera (for example InfiRay P2 Pro or Teledyne FLIR ONE Edge Pro), a probe thermometer, a kitchen scale and ruler, bought fish and ice, Meta's open Segment Anything model to outline the fish in each photo, and Python to fit and test the estimate.
First step
chill one whole market fish in ice, then let it warm on a bench, taking a thermal photo and a probe reading every 10 minutes for 3 hours.
Done looks like
at least 300 paired thermal photos and probe readings from at least 3 species, taken while fish chill in ice and while they warm in air; the dataset published; an error table for fish held back from fitting, with 90 percent of estimates within 1.5 °C of the probe as the target.