Three days of camanchaca fog shut down every camera we own on the Humboldt Coast. The survey didn't stop — it changed tools. How to work a reef blind.
Tidewright field notes are our crew's own stories from project sites. Everything in them — sites, numbers, people — is fictional, part of the Tidewright demonstration. The oceanography is real.
The camanchaca arrived on the second night of our August fortnight on Chile's Humboldt Coast and stayed for three days — a fog so complete that from the boat console I could read the chartplotter and not see the bow. For a survey program built on cameras and drones, this should have been dead time. It wasn't. By lunchtime on day one we had changed tools, and by the end of day three we had produced one of the most useful datasets of the season. This note is about how to survey a reef you cannot see, and what fog teaches an operation that mostly runs on light.
Our Humboldt pilot site sits in the lee of a low desert headland where the Humboldt Current's upwelling meets warm inland air — a fog machine running since before there were words for it. The August plan was a drone canopy census paired with diver camera transects across nine hectares of Lessonia beds. The drone doesn't leave the case in fog: photogrammetry needs contrast, and wet optics on a half-million-peso aircraft over salt water is how budgets die. The camera transects fared no better — divers reported visibility of "one arm, sometimes two."
What the fog does not touch is sound.
Drifter 11, our survey vessel, carries a pole-mounted multibeam and a sidescan towfish — the full seafloor mapping kit — as standard fit. Fog is frustrating for the skipper and irrelevant to the sonar. So day one became a systematic acoustic resurvey of the whole site: 200-metre lanes, 50% overlap, combing nine hectares at walking speed. We know from sonar survey practice that backscatter hardness separates reef from sand; what we wanted this time was finer — whether the backscatter texture could pick out the kelped parts of the reef from the bare ones.
It half could. Dense Lessonia canopy softens and roughens the return in a way our analyst could pick out against the reference mosaic from May, at perhaps 60–70% agreement with the diver maps. Not audit-grade — but a genuine fog-day product, and a hypothesis to test properly in clear weather.
With cameras useless, we rebuilt the diver transects around tools that don't care about visibility:
By the end of day two we'd completed seven of the nine planned transects — slower than a camera day, but the stipe-count data were as good as any we have.
By Thursday we stopped treating the fog as an obstacle and started treating it as data. The camanchaca isn't just a nuisance — it's the water supply of the coastal desert behind the site, part of why fog defines these coasts ecologically. We repositioned Drifter 11's weather mast and ran a fog-day microclimate profile: sea-skin temperature, air temperature and dew point at two heights, photosynthetically active radiation at the surface and at five metres. The finding worth writing down: at the thickest hour, light at five metres depth had fallen to levels we'd expect at twenty metres on a clear day. Three days of that is nothing to a Lessonia bed; the light budgets of kelp forests run on weeks, not afternoons. But a twenty-day camanchaca event — which old fishermen here describe as happening some winters — is a light stressor nobody currently puts in the site risk model. It goes in this autumn.
The crew joke by Friday was that we should schedule a fog week into every season. It's a joke with a pulse. A survey program that only works in good weather is a fair-weather friend to a forest that lives through everything.
The in-water risk is manageable — divers navigate by line and beacon regardless of surface visibility — but the surface risk is real: a boat that can't see divers and can't be seen. Our rule is beacons plus a strict one-tender-per-two-divers ratio, and no operations at all in shipping lanes during fog.
Acoustics can detect canopy reliably in some settings — echosounder work on canopy is a mature method — but using hull-mounted backscatter texture to separate kelped from bare reef is site-specific and needs validation per bed. That's next season's job.
On the Humboldt Coast in winter, potentially a third of any fortnight; in Mendocino summers, we've written about long fog before. Any temperate upwelling coast should assume fog weeks exist and carry the kit that makes them productive.
The acoustic beacons. Everything else on the fog list has a workaround; straight, repeatable transect lines without them do not.
Tidewright's first Humboldt Coast season: working alongside local diver unions, learning huiro protocols, and finding out what the Chilean coast teaches back.
Jack Callum ·
Multibeam and sidescan sonar build metre-scale maps of reef relief, substrate and kelp canopy. How acoustic mapping works, and what it can and can't see.
Dr. Mara Ellison ·
A field note from three fog-bound months on California's north coast: working around the marine layer, and why the fog and the kelp are the same story.
Jack Callum ·