Deep Sea Mining Technology
Controversy and Outlook

An oceanographer gets hired to write the monitoring section of an ISA mining application. Knows sensors biofoul. Knows wipers jam. Writes the plan describing continuous monitoring, adaptive management, trigger thresholds. Does not mention that the sensors will fail and the ships may not come and the taxonomy bottleneck runs through a few dozen specialists worldwide and temporal baselines do not exist. The mining company did not pay for a document about why its application should fail. ISA review evaluates completeness. No procedure for checking deliverability.

Biofilm colonizes the sensor window within days. Wiper sweeps every few hours. Extends useful life from weeks to months. Wiper jams. Most common failure mode for long-duration optical instruments at depth. Anti-fouling research since the 1980s. Copper coatings, UV-C LEDs, chlorine electrodes, hydrophobic treatments. Lab results promising, field results at 450 atm in near-freezing water over years: biology colonizes surfaces faster than countermeasures can prevent it. The delay anti-fouling provides is weeks to months. Monitoring requirement is years to decades. Four decades of research have not changed that ratio. Progress has been made in the sense that a sensor that lasted two weeks in 1990 might last three months now. The monitoring plan calls for years.

A mooring deployed from a German research vessel during a 3-week CCZ window between Manzanillo and Honolulu. Eight months later the optical data is noise. Wiper jammed month four. Servicing cruise postponed, ship reassigned after Tonga earthquake.

Normal.

Replacement cruise month eleven.

Normal.

Box corer biological sampling. Fraction of a square meter. Plume may have affected tens of km². Samples to taxonomy lab in Europe. Months to a year for results. Collector kept moving.

Changing it would require the ISA to build a technical assessment capability. Nobody is proposing this. The ISA Secretariat has limited budget and staff. Hiring instrument engineers to evaluate monitoring deliverability would be an admission that the current review process cannot evaluate monitoring deliverability, which is an admission that the adaptive management framework the entire Mining Code rests on has not been validated, which is an admission that would complicate every application currently in process and every application previously reviewed. Bureaucracies do not voluntarily surface findings that invalidate their previous work. This is not specific to the ISA. It is how bureaucracies work. The U.S. Minerals Management Service did not voluntarily surface the finding that its inspection capacity was a fraction of what its regulations assumed, either, and that ended with Deepwater Horizon.

Washburn and colleagues, Marine Policy: CCZ baselines too sparse for impact attribution. Temporal baselines, repeat visits over multiple years, barely exist. A detected change near a collector track: mining or natural variation? Unanswerable without temporal replication. The monitoring data stays ambiguous while additional cruises accumulate over years while the operation continues because the ambiguous data has not triggered the threshold that the threshold was calibrated against a baseline that was not adequate to detect its own crossing.

That sentence was convoluted. It should be, because the situation is convoluted. The circularity, assess impacts before permitting operations that generate the data needed to assess impacts, is not a rhetorical device. It is the operational reality of every ISA environmental impact assessment for nodule mining. The GSR and BGR experiments tracked plumes from hectare-scale disturbances. Commercial operations produce overlapping plume events across orders of magnitude more area. Plume physics is not scale-invariant. The coefficients needed for commercial-scale models have not been measured. The circularity has no exit within the current framework.

Here is where the monitoring problem connects back to the DISCOL threshold.

The ISA framework treats environmental impact as proportional and manageable through adaptive management. DISCOL data shows impact is binary and permanent. Adaptive management requires monitoring. Monitoring cannot be delivered. So the framework is wrong about the nature of the impact AND cannot detect the impact it is wrong about. These are independent failures that happen to compound. If the framework had the impact model right and the monitoring wrong, you would have good intentions defeated by logistics. If the framework had the monitoring right and the impact model wrong, you would have good data feeding into bad analysis. Both wrong simultaneously means the governance system is disconnected from the physical and ecological reality it governs on two separate axes.

The ISA Council delegates negotiating the Mining Code in Kingston are, in a specific and non-metaphorical sense, designing a management system for a process they cannot observe, based on an impact model that the field data contradicts, reviewed by a process that cannot evaluate whether the management system can function, and all of this has been proceeding for over a decade.

Sediment plume scaling: GSR and BGR hectare-scale data, commercial operations at orders of magnitude larger scale, plume physics not scale-invariant, parameters unmeasured. Return water from dewatering at 1,000-1,200m, mesopelagic, biological pump, several Gt C/yr. Discharge for decades. Carbon flux interaction at commercial scale: unmodeled. The mining pitch is decarbonization minerals and the carbon pump interaction has not been quantified. These are recognized as problems in the scientific literature. They are not the focus of this article because they are discussed elsewhere. The monitoring deliverability problem is not discussed elsewhere and is, in operational terms, more fundamental: you can in principle model plume dispersion and you can in principle model carbon flux impacts, you cannot in principle monitor abyssal mining impacts in real time with instruments that biofoul in weeks.

All of this has been written about. Extensively. The positions are stable. The geopolitical landscape of deep sea mining is one of the most thoroughly analyzed topics in ocean policy and there is nothing useful to add about it here.

What has not been written about, and the reason this article exists, is the wiper blade.

The wiper blade and the biofilm and the ship that was late and the taxonomist in Bremen and the box corer that covers a fraction of a square meter and the baseline that does not exist and the review form that has no field for "can this be physically done." The aggregate of these small, boring, technical, logistical, institutional failures is a monitoring system that does not monitor. An adaptive management framework that does not adapt. A regulatory architecture that exists on paper in Kingston and does not exist on the seabed in the CCZ. The architecture governs an activity whose ecological consequences are permanent on geological timescales, and the architecture's connection to the physical environment it governs runs through a rubber wiper blade on a glass window at 450 atmospheres that jams at month four.

Whether a collector reaches the CCZ seabed at all depends on cathode chemistry. If CATL's sodium-ion cells or whoever's solid-state lithium cells or whoever's whatever-comes-next cells scale to automotive range, the cobalt and nickel demand case for nodules weakens to the point where 7-10 years of capital expenditure does not make financial sense. Indonesian laterite nickel is already depressing prices. Manganese, the fallback value story, is the 12th most abundant crustal element with 70% of reserves in South Africa. The industry may die before it deploys. The Mining Code becomes a seminar. The CCZ stays intact.

If the industry does not die, a collector reaches the CCZ. Under the framework described above. With the monitoring described above. Proportional-impact model contradicted by DISCOL. Adaptive management dependent on instruments that biofoul and ships that are late and baselines that do not exist. Approved through a process where the knowledge about why the monitoring will fail and the authority to approve the monitoring exist in professional worlds that do not intersect.

The ISA Council could commission an independent technical assessment of monitoring deliverability. One study. Hire three or four ocean instrument engineers, give them six months, have them evaluate the monitoring plans in existing applications against published data on instrument performance, ship availability, and baseline adequacy at abyssal depth in the CCZ. The study would cost less than a single day of research vessel time.