BRENNEN

Industries / Food, farming and forestry / Fishing and aquaculture / Shellfish farming

Water quality

This is the work of watching the water shellfish grow in for sewage, toxic algae and changes in its chemistry, and of closing beds when the water makes shellfish unsafe.

7

Milestones achieved, from 1925 to 2022.

2

To watch for: announced, not yet confirmed outside.

2

Open deliverables. Nobody has shipped them.

The road

AchievedWatch forOpen

Achieved

  1. 1925

    Certifying clean growing waters

    In the winter of 1924–25 a typhoid fever outbreak, with cases in New York, Chicago and Washington, D.C., was traced to oysters from sewage-polluted water. The US Surgeon General called a conference on 19 February 1925 that founded a national certification programme, now the National Shellfish Sanitation Program (NSSP). States approve a growing area only after a sanitary survey, a study of the sewage sources near it plus bacteria counts in the water, and certify only shippers who meet agreed standards.

    Lessoneach producing state was made answerable for the shellfish it sent to other states, not only for its own residents.

    fda.govfda.gov

    Credit

    US Public Health Service, with state health authorities and the shellfish industry

  2. 1991

    Europe grades growing areas

    Council Directive 91/492/EEC made EU authorities class growing areas by the count of E. coli, a gut bacterium that signals sewage, in the shellfish flesh rather than in the water. Shellfish from the cleanest class could be sold directly; those from the middle class had to be purified or relayed (moved to clean water) first; those from the dirtiest class had to be relayed for at least two months. The same three classes, A, B and C, still apply under Regulation (EU) 2019/627, in which Class A allows 230 E. coli per 100 g in 80% of samples and no more than 700 in the rest.

    eur-lex.europa.eueur-lex.europa.eu

    Credit

    Council of the European Communities (1991); European Commission (2019)

  3. 2008

    A seasonal toxic-bloom forecast

    Alexandrium, the alga behind paralytic shellfish poisoning, rests on the seafloor as dormant cysts. In spring 2008 Don Anderson and Dennis McGillicuddy of the Woods Hole Oceanographic Institution (WHOI) counted about 30% more cysts in the Gulf of Maine than before a large bloom in 2005, and a computer model run on the counts predicted a big bloom. It came and closed shellfish beds from Canada to Massachusetts. A 2014 paper by Kleindinst and colleagues calls it the first seasonal forecast for paralytic shellfish poisoning in the Gulf of Maine and says such forecasts were by then issued every year.

    LessonMaine's head of biotoxin monitoring said the forecast gave the state "a bit of an advance warning" to combine with local knowledge.

    whoi.edupmc.ncbi.nlm.nih.gov

    Credit

    Don Anderson and Dennis McGillicuddy (WHOI); Ruoying He (North Carolina State University), who built the model with McGillicuddy

  4. 2008

    Plankton camera catches toxic bloom

    The Imaging FlowCytobot is an underwater instrument that photographs each plankton cell as seawater flows through it. One had run at the mouth of the Mission-Aransas estuary in Texas since September 2007. In early February 2008 researchers spotted cells of Dinophysis, a toxic plankton genus, in its images; by late February counts passed 200 cells per mL, tests found okadaic acid, the toxin behind diarrhetic shellfish poisoning, and shellfish harvesting was closed. The authors' paper title calls it the first harmful Dinophysis bloom in the US.

    ebi.ac.uk

    Credit

    L. Campbell, R. J. Olson, H. M. Sosik and colleagues (Campbell et al., 2010)

  5. 2014

    Acidity data for hatcheries

    NOAA's Integrated Ocean Observing System (IOOS) launched a Pacific ocean acidification data portal showing real-time seawater chemistry for the coast and for shellfish hatcheries and farms in Washington, Oregon, Alaska and California. Many readings come from Burke-o-Lators, instruments developed by Burke Hales of Oregon State University that measure how easily larvae can build shell from the water. Growers use the data to decide when to raise larvae, when to set seed out and when to fill their tanks.

    washington.edu

    Credit

    NOAA IOOS; Jan Newton (University of Washington), who led the portal; Burke Hales

  6. 2019

    Norovirus counted across Europe

    The European Food Safety Authority (EFSA) analysed two years of EU-coordinated sampling of raw oysters: 2,180 samples from growing areas and 2,129 from dispatch centres. It estimated that norovirus was present in 34.5% of growing-area samples and 10.8% of batches leaving dispatch centres, with higher levels from November to April and lower levels in Class A areas, and concluded that the E. coli rules "might be complemented by a norovirus criterion".

    Lessongrading areas by a sewage bacterium does not screen out norovirus, which FAO calls the most common illness from eating bivalves in many developed temperate countries.

    pmc.ncbi.nlm.nih.govfao.org

    Credit

    EFSA, with EU member-state laboratories

  7. 2022

    Rain-closure forecasts per lease

    US rules let states run some growing areas as "conditionally approved", closing automatically when a set trigger is passed; the NSSP Guide's example trigger is 5 cm (2 inches) of rain in 24 hours. From early 2022 North Carolina State University's ShellCast app combined National Weather Service rain forecasts with each grower's lease to send a daily closure-risk alert for that day and the next two, in North Carolina and, from 2023, South Carolina.

    ncseagrant.ncsu.edusecoora.orgfda.gov

    Credit

    Natalie Nelson's team at North Carolina State University, funded by North Carolina Sea Grant and the Southeast Coastal Ocean Observing Regional Association (SECOORA)

Watch for

  1. Copepod signals forecast toxins

    Copepods are tiny crustaceans that graze on algae and release chemical cues into the water. A team sampled these cues for 28 months at the Santa Cruz Municipal Wharf, California, with resin samplers that monitoring networks already use, and reports that they predicted toxic Pseudo-nitzschia blooms six weeks ahead and domoic acid in mussels above safety limits seven weeks ahead; models that tracked the toxin itself did best only one week ahead.

    Achieved whena state or national biotoxin programme adds copepod cues to its routine sampling and later flesh tests at sites beyond Santa Cruz bear out the warnings.

    globalseafood.org

    Claimed by

    Aubrey Trapp (Northwest Indian College), Raphael Kudela (University of California Santa Cruz) and colleagues, in PNAS, October 2026

  2. Satellite bloom prediction in Portugal

    Researchers trained machine-learning models on satellite data from 2013 to 2023, including sea temperature, upwelling and chlorophyll, to predict blooms of Pseudo-nitzschia in two of Portugal's shellfish production zones. Tested on whole years and areas held back from training, the blooms proved "moderately predictable", which the authors call useful for early warning.

    Achieved whenthe work passes peer review and a monitoring authority issues warnings from it that later flesh tests confirm.

    arxiv.org

    Claimed by

    Ayman Bnoussaad, Alexandre Bernardino and colleagues (arXiv preprint), July 2026

Open

  1. 01

    Count toxic algae by phone

    A low-cost microscope-and-camera kit, with software, that counts Dinophysis (which makes the toxins behind diarrhetic shellfish poisoning) and Pseudo-nitzschia (which makes domoic acid) in seawater samples, tested against a monitoring laboratory's counts over a season.

    The brief
    Credit · unclaimed
  2. 02

    One feed of harvest closures

    A public, machine-readable feed of whether each shellfish harvest area is open or closed, across many jurisdictions, refreshed daily from the regulators' own pages, with every status linked to the page it came from and an archive kept.

    The brief
    Credit · unclaimed

Open 01 · the brief

Count toxic algae by phone

Credit · unclaimed

The first to publish is credited on this line, and the deliverable moves from Open to Achieved.

What to build

A low-cost microscope-and-camera kit, with software, that counts Dinophysis (which makes the toxins behind diarrhetic shellfish poisoning) and Pseudo-nitzschia (which makes domoic acid) in seawater samples, tested against a monitoring laboratory's counts over a season.

Why it is still open

HABscope, a US$500 kit from Texas A&M and the Gulf of Mexico Coastal Ocean Observing System (GCOOS) that sends a short microscope video to cloud software for counting, is trained on Karenia brevis, Alexandrium and Pyrodinium; its page names neither Dinophysis nor Pseudo-nitzschia and reports no accuracy figures. A UCLA portable imaging cytometer, with parts under US$2,500, matched California Department of Public Health Pseudo-nitzschia measurements at six Los Angeles beaches on one day in 2018, as a proof of principle. Searches found no kit in routine use for both genera.

gcoos.orginnovate.ee.ucla.edu

Built with what exists

A HABscope-style kit (a compound microscope, a 3D-printed case, a Raspberry Pi and camera); labelled plankton images from monitoring laboratories and from WHOI's public Imaging FlowCytobot archive; an object detector such as Ultralytics YOLO.

  1. First step

    Ask a state or volunteer phytoplankton monitoring programme for 20 split samples and film them on a borrowed microscope with a phone adapter.

  2. Done looks like

    200 seawater samples split with a monitoring laboratory over a season; the kit's counts within a factor of two of the laboratory's for at least 80% of samples, and at least 90% of samples above the laboratory's alert level flagged; results published for each genus.

Open 02 · the brief

One feed of harvest closures

Credit · unclaimed

The first to publish is credited on this line, and the deliverable moves from Open to Achieved.

What to build

A public, machine-readable feed of whether each shellfish harvest area is open or closed, across many jurisdictions, refreshed daily from the regulators' own pages, with every status linked to the page it came from and an archive kept.

Why it is still open

Oregon posts its shellfish biotoxin results as a PDF in which "Results in RED means there is a closure". Canada's national map of classified harvest areas is updated twice a year, warns that areas "may be superseded at any time by regulatory orders", and tells users to contact their local Fisheries and Oceans Canada office for current status. Searches found no single feed of current status for either country.

oregon.govcatalogue.ec.gc.ca

Built with what exists

Python scrapers; pdfplumber, which reads text and its colour from PDFs; a small, low-cost model for repeated extraction work, such as Meta's Muse Ember (Note: "The jobs agents repeat"); GeoJSON for area outlines; GitHub Actions to run it daily.

  1. First step

    List every US state and Canadian regional page that posts current shellfish closures, with its format: web page, PDF, map or phone line.

  2. Done looks like

    Ten jurisdictions covered for 90 days; a daily hand check of a random 5% of areas agrees with the official source at least 99% of the time; feed, archive and code public.