> ## Content Index
> Fetch the complete content index at: https://bluefinance.ca/llms.txt
> Use this file to discover other available public pages before exploring further.

# How Do We Know What We Are Doing to the Ocean?
- URL: https://bluefinance.ca/what-we-are-doing-to-the-ocean/
- Published: 2026-09-18T18:48:28.000Z
- Updated: 2026-09-18T18:48:28.000Z
- Description: Collecting and processing ocean data has got cheaper and faster. Knowing what we are doing to the ocean has not. Observation, interpretation and assurance are three different problems.
- Author: Brian Rogers
- Tags: Where Things Stand, #policy-then-explore, #ocean-begin-here, #ocean-begin-1, #policy-explore-1

Line P runs west from Vancouver Island for almost 1,500 kilometers, out past the continental shelf and into water several thousand meters deep. Ships have been working hydrographic stations along it since 1956, and much of what is known about the North Pacific rests on that record. A ship goes out and occupies the line three times a year.

Three cruises a year is not a failure of ambition. It is what a research vessel costs, multiplied by the distance, in a country with the longest coastline on earth and an exclusive economic zone of roughly 5.75 million square kilometers. Line P is among the best-observed stretches of open ocean Canada has, and most of the rest gets less.

A lender in Vancouver can establish, in an afternoon, who owns a property in Prince George, what its boundaries are, whether the building on it was inspected, what an appraiser thinks it is worth, and what the business inside it earned last year. The lender may never see the building. None of that feels like infrastructure. It feels like looking things up.

Every one of those answers rests on a system somebody built. [British Columbia's survey record reaches back to 1851](https://ltsa.ca/property-owners/about-land-records/history-of-bcs-land-title-system/?ref=bluefinance.ca), and the survey monuments, plans and regulations that followed made a boundary a legal object instead of a description. The province's modified Torrens system makes the land title register conclusive, so a lender does not have to reconstruct the history of ownership to rely on it. [Canadian appraisal standards](https://www.aicanada.ca/about-aic/cuspap/?ref=bluefinance.ca), introduced in 2001, set common requirements for how AIC members develop and report an opinion of value.

The lender pays for the appraisal. It does not pay for the century and a half of institution-building that makes an appraisal mean something, and it does not have to invent survey law, title registration, professional accreditation, building codes, or accounting standards before it can advance the money. That machinery is established and it is shared, which is why nobody looks at it.

The ocean does not have an equivalent system, but decisions still have to be made about what is happening in it. A regulator may need to know whether a discharge is causing harm, a community whether a closure is helping a stock recover, or a lender whether an ecological indicator has actually changed. Somebody has to establish what happened in the water. That question breaks into three parts. Observation is putting an instrument, a person, a vessel, a sample bottle, or a satellite [where the ocean can be measured](https://bluefinance.ca/scientists-and-the-ocean-economy/). Interpretation is turning what comes back into a claim about what happened, what caused it, and whether an intervention worked. Assurance is what lets somebody who was not there rely on that claim.

In 2020 a team led by [Manuel González-Rivero](https://doi.org/10.3390/rs12030489?ref=bluefinance.ca) compared what it cost to have an expert label coral reef images by hand with software doing the same job. The expert came to USD 5.41 an image. The software came to USD 0.07\. It read about 1,200 images an hour. An experienced ecologist needed 16 hours to get through the same number. The software introduced errors of 2 to 6 percent. The human comparisons cited in the paper ranged from 2 to 5 percent, including differences when the same observer did the work twice.

In 2021 [Katie Kowarski and colleagues](https://doi.org/10.1121/10.0004851?ref=bluefinance.ca) published a Canadian method for working through acoustic recordings. The software screened the recordings first, and an analyst then checked a selected subset rather than listening to everything. On the test dataset, the two approaches agreed on whether a species was present in a given hour between 73 and 100 percent of the time. A person still checks the work; there is simply much less of it to check.

None of that makes the pictures cheaper to take. [Pacific Northwest National Laboratory](https://tethys.pnnl.gov/sites/default/files/publications/Fu%5Fet%5Fal%5F2021.pdf?ref=bluefinance.ca) priced three hypothetical ways of surveying fish at the same four sites in Sequim Bay over a single day. Using senior-researcher labour rates, eDNA came to about USD 4,450, against about USD 14,950 for beach seining and USD 11,580 for a scuba video transect. The labour estimates were 13 hours, 84 and 68 respectively. At student labour rates the ranking changed, with the scuba survey marginally cheaper than eDNA. More than half the eDNA figure was the USD 200 a sample charged by the laboratory for extraction and sequencing, and that cost does not change with the labour rate. These were constructed scenarios built from published costs and expert input, and they excluded vessel use and operation. It is not a receipt.

Gliders do reduce some of the cost of observation. A [2023 paper](https://doi.org/10.3389/fmars.2023.1108326?ref=bluefinance.ca) drawing on nearly two decades of Atlantic Canadian operations describes a significant reduction in cost per profile compared with ship-based surveys. But running one still means batteries, maintenance, communications, calibration, insurance, staff and the logistics of getting it into and out of the water. The paper's estimates exclude institutional overhead, which the authors say can add more than 50 to 60 percent in some cases. Remote missions may still need a large vessel for deployment and recovery.

The classifier that reads an image for USD 0.07 was trained on annotation that cost USD 5.41, and its results still have to be checked against labelled observations somebody produced by hand. The cheap layer runs on the expensive one. When [Peter Mumby's team](https://doi.org/10.1006/jema.1998.0255?ref=bluefinance.ca) costed coarse habitat mapping from satellite imagery in 1999, field survey came to about 20 percent of total cost, and to more than 80 percent where a remote sensing facility already existed and the set-up costs fell away. Field survey took around 70 percent of the project's duration. A [2023 study](https://pmc.ncbi.nlm.nih.gov/articles/PMC10622156/?ref=bluefinance.ca) mapped seagrass at eleven United States coastal sites from WorldView imagery and checked the classifications against reference data at every site. The satellite was much better at identifying where seagrass was absent than where it was present.

Sensors drift and instruments foul. The oxygen optodes on some of the Atlantic Canadian gliders are calibrated in a laboratory at Dalhousie, CTDs are normally returned for calibration every two or three years, and keeping a sensor calibrated means either holding a spare or accepting that the instrument is out of service while it is away. A glider collects without a vessel alongside it, but somebody launches it from a small boat, somebody pilots it for two or three hours a day on the shelf, and somebody goes out and gets it back. One glider off Newfoundland lost its main battery power 25 miles from the recovery point, drifted for more than a day, and was picked up by a fishing boat and brought into Old Perlican. eDNA moves the work out of the water and into a laboratory. The samples still have to be filtered, extracted and sequenced, with reference databases and contamination controls behind the result.

Replacing a cruise with a sensor solves one problem. Keeping a series going since 1956, through the end of the weathership program and into today's cruises, is another. MEOPAR's [National Research Vessel Task Team](https://meopar.ca/partnerships/national-research-vessel-task-team/?ref=bluefinance.ca) surveyed more than 50 marine researchers and found that 86 percent did not have sufficient access to research vessels. Only 30 percent had recently had an opportunity to propose and lead an expedition as chief scientist. One response was [Modular Ocean Research Infrastructure](https://meopar.ca/partnerships/modular-ocean-research-infrastructure-mori/?ref=bluefinance.ca), or MORI: containerized laboratories that can be put aboard vessels that were not built for research. MORI has completed six cruises in its initial development and demonstration phase. In 2025 it put a laboratory aboard HMCS Margaret Brooke for Canada's first all-Canadian Antarctic expedition.

Somebody has to be able to make a claim that somebody else is willing to act on. That can take several forms, including independent verification, professional certification, scientific peer review, regulatory assurance, or an audit opinion, and they differ in how binding they are and who carries the consequences. What they share is a standard the claim is made against, some distance between the person making the claim and the person checking it, and something that happens when the claim turns out to be wrong. Cheaper measurement does not create any of that. It does not establish who has the authority to attest to a claim, or what follows if the attestation fails. Blue finance has its own version of the question: [who is positioned to do that work in Canada](https://bluefinance.ca/who-checks-that-blue-finance-works/).

In 1970 [Stephen Blaber](https://doi.org/10.1093/oxfordjournals.mollus.a065097?ref=bluefinance.ca) reported something strange in female dogwhelks from Plymouth Sound: a penis-like outgrowth behind the right tentacle. He had an observation and no explanation for it. By the late 1970s another problem was unfolding in [Arcachon Bay](https://pubmed.ncbi.nlm.nih.gov/15091358/?ref=bluefinance.ca) on the French Atlantic coast. Pacific oysters stopped reproducing normally, their shells thickened and deformed, and from 1977 to 1981 no oyster spat were collected as the local shellfish industry collapsed. Water in the open bay had seldom carried more than about one nanogram of tin per litre through the 1960s and probably passed 100 during the collapse years. A nanogram per litre is roughly one part per trillion in water. [The source was TBT leaching from antifouling paint](https://doi.org/10.1016/S0048-9697%2800%2900510-6?ref=bluefinance.ca) on the boats moored around the oyster beds. The dogwhelk response was later shown to vary with exposure and became a way of monitoring the contamination itself.

France moved first, banning TBT paints on boats under 25 metres in January 1982, and Arcachon came back quickly. Spatfall was normal by 1983 and the shell anomalies had gone by 1985\. Elsewhere it carried on. Bryan and colleagues found imposex widespread along the south coast of England in 1986\. IMO took up the question in 1988\. The [convention prohibiting organotin antifouling systems](https://www.canada.ca/en/environment-climate-change/corporate/international-affairs/partnerships-organizations/anti-fouling-systems-ships.html?ref=bluefinance.ca) was adopted in October 2001, entered into force in September 2008, and entered into force in Canada on July 8, 2010\. Forty years from a dogwhelk in Plymouth Sound to Canadian law.

In commercial lending, a property with an industrial history could trigger a Phase I environmental assessment, and a Phase II if the first report found something that needed closer investigation. Lenders relied on those reports because a qualified professional signed them and the findings had consequences. Remediation steps could become conditions of the financing, and in more serious cases we might need a remediation plan approved by a ministry before proceeding. I remember environmental engineers coming in to walk relationship managers through the issues so they understood what they were being asked to rely on.

Canada already has [substantial observing infrastructure](https://bluefinance.ca/marine-technology-and-ocean-monitoring-science-sovereignty-and-the-commercialisation-gap/). [Ocean Networks Canada](https://www.oceannetworks.ca/observatories/?ref=bluefinance.ca) operates observatories on the Pacific, Atlantic and Arctic coasts. The Canadian Integrated Ocean Observing System does a different job: making measurements collected by different organizations discoverable and usable together. DFO renewed CIOOS funding through March 2027\. In March 2026, [CIOOS received another CAD 1 million over two years](https://cioos.ca/new/data-space-pilot-for-the-canadian-integrated-ocean-observing-system/?ref=bluefinance.ca) through the National Data Spaces Pilot to work on data management, governance and interoperability.

There is no single defensible figure for how much of Canada's ocean is monitored. A satellite reads surface temperature across enormous areas every few days. A cabled observatory watches a few square meters without stopping. A ship reaches Line P three times a year. One percentage would hide all of it. The [Council of Canadian Academies found in 2013](https://www.cca-reports.ca/reports/ocean-science-in-canada-meeting-the-challenge-seizing-the-opportunity/?ref=bluefinance.ca) that Canada had several world-class observing systems alongside problems of geographic coverage and data integration. Its [June 2026 report](https://www.cca-reports.ca/reports/mobilizing-ocean-research-for-canada/?ref=bluefinance.ca) found that Canada's ocean research strengths still do not consistently translate into benefits for the environment, economy and communities.

[Norway monitored dogwhelks along its coastline from 1991 to 2017](https://doi.org/10.1016/j.marenvres.2018.11.011?ref=bluefinance.ca). In 2017 imposex was not detected at any of the eight monitoring stations. The recovery became visible because somebody kept going back and looking. The [Global Ocean Observing System's 2025 status report](https://www.ocean-ops.org/goosreport/?ref=bluefinance.ca) calls for long-term, coordinated investment in ocean observing as shared global critical infrastructure. Canada did not make property financeable by inventing a clever verification technology. It built registries, survey systems and professional standards over a hundred and fifty years, until a lender in one city could establish the facts about a building in another and think nothing of it. The equivalent work on the water is partly done, unevenly funded, and not yet joined up. The ships still have to go out.