The cells clear first. The toxin does not.

Not every bloom is toxic and not every toxic bloom is visible. Those two facts together are why bloom management is run on assays rather than on photographs, and why a system that only reduces biomass is only doing half of the work.

Three toxin families account for most of the public health and regulatory pressure on the coastlines we work. They behave differently, they are measured differently, and they close different things.

Brevetoxin — the one you breathe

Florida red tide, Karenia brevis, is unusual in that its toxin becomes an air quality problem. Wave action aerosolises brevetoxins and onshore wind carries them inland. What the clinical literature describes is an apparently reversible upper-respiratory syndrome — coughing, throat irritation, chest tightness — and a markedly harder version of it in people who have asthma.

We used to state a recovery time here. We have taken it out. The Fleming study we were citing records asthmatic participants reporting symptoms for up to at least five days after a single one-hour beach exposure, which is not a number that belongs anywhere near the word "brief". If your public health authority asks us how long the effect lasts, the answer they get is that it depends on who is standing on the sand.

The commercial consequence follows from that, not from the severity. A beach where a proportion of visitors start coughing is a beach that empties, and it stays empty longer than the aerosol does. This is the mechanism behind the tourism losses on the Florida record, and it is why resort groups treat a bloom as a reputational event before an environmental one.

Domoic acid — the one with no antidote

Pseudo-nitzschia blooms produce domoic acid, which causes amnesic shellfish poisoning. At low doses that means nausea, vomiting and diarrhoea. At higher doses it means seizures, coma, irreversible short-term memory loss and death, with elderly people at highest risk.

Two properties make it a regulator's nightmare. There is no antidote. And cooking or freezing does not destroy it, so there is no step downstream of harvest that makes contaminated tissue safe.

That is the whole explanation for why a shellfish closure holds after the water has cleared. It is also why the 2015 West Coast event cost $97.5 million in Dungeness crab landings and roughly $40 million more in razor-clam tourism, against a bloom most people never saw.

Microcystin — the one in the tap

Cyanobacterial blooms in fresh and brackish water produce microcystins, of which microcystin-LR is the most studied. The WHO provisional drinking-water guideline for it is one microgram per litre. Human poisoning presents primarily as liver damage, gastrointestinal illness and hepatic toxicity.

Toledo, Ohio, is the case everyone in this field cites, because it was not a fishery or a beach. In 2014 a Lake Erie bloom contaminated the municipal supply and roughly 500,000 residents were advised not to drink the tap water. A city does not recover its confidence in the water system as quickly as it recovers the water system.

Four tiers of claim, and where ours sit

Toxin numbers are the ones a health authority will actually check, so it is worth setting out plainly how much weight each part of our evidence can carry. Sorted by who did the work.

Tier one — physics, published by other people. Bubbles below about a micron behave as a stable suspended phase rather than as something rising to burst, and the measured advantage in how fast gas crosses into water is in the peer-reviewed record. None of that is ours and none of it is contested.

Tier two — the approach, tested on somebody else's kit. An eight-acre Florida pond in 2018, an ozone nanobubble system, algae gone inside two days with reoxygenation held and no apparent harm to the life in the water, all reported by NOAA's coastal science centres. A separate ballast-water evaluation two years later found no statistically significant adverse residual toxicity downstream. Different equipment, different partners. Those releases say the approach is workable. They say nothing about our hull.

Tier three — ours, and mostly unpublished. One Florida DEP permit, FLOA00062, granted to Alarivean, Inc. in September 2024 under Chapter 403 of the Florida Statutes on a five-year window, to trial an ozonated-seawater craft against red tide with live Karenia brevis above 100,000 cells per litre as the gate. Logged independently by the Sarasota Bay Estuary Programme and by START. The estuary programme's stated view is that efficacy in open water is undemonstrated, which is accurate and is also an offer. It supported the permit application technically, took no funding, and would like to be the party holding the instruments when open-water measurements finally get taken. Then an endorsement past the bay would rest on their data rather than ours. We want that too, and for the same reason they do.

Also tier three, and readable by nobody: a Cooperative Research and Development Agreement with NOAA, under which work with Dr Peter Moeller established that this treatment sterilises inorganic toxins. CRADAs are private, so there is no link to hand you and there never was going to be. It belongs in this tier — the company's own account, stated by the company — rather than in tier two, because a research agreement is not an endorsement and NOAA has endorsed nothing.

Tier four — genuinely open. Whether the bubbles themselves generate hydroxyl radicals is argued in the literature and we assume they do not. Bromate in ozonated seawater is a live constraint with a real efficiency cost attached to controlling it. And nobody has shown that a duty proven in a pond and a ballast tank transfers cleanly to open coastal water with tide, wind and dilution working against it. That last one is precisely what a calibration phase is for, on your water, before capacity is signed for.