We are in the midst of a disturbing trend – states are developing nuclear-powered drones using fully autonomous reactor control systems. The ecological and environmental impacts, even of successful testing, may be severe but the assumption that the ocean is an acceptable dumping ground for “disposable” nuclear reactors is even worse.
Fission reactors provide these subsurface or aerial munitions (with either conventional or nuclear warheads) with such enormous range that they don’t have to follow predictable trajectories and can duck and weave through sensor bubbles to surprise an adversary.
To be clear, I don’t think that we will see them deployed in the near term – the technical challenges are extreme – but that hasn’t stopped Russia from conducting tests of the nuclear-powered nuclear cruise missile, Burevestnik, and developing its nuclear-powered nuclear torpedo cousin, Poseidon. Chinese researchers have also announced the completion of a conceptual design for a nuclear-powered conventional torpedo that could strike at a target 10,000 km away. The project lead, Guo Jian, speaks of using them in swarms that deploy AI-targeting to recognise and prosecute a maritime target in a shipyard or base before it can deploy.
This is dangerous in so many ways: How will operators ensure the torpedo is on target if they can’t communicate with it? Remote undersea communications are too low-bandwidth at this range, hence the reliance on autonomous targeting and reactor control. How will an ‘AI’ reliably discern the difference between civilian and military targets at port from beneath the waves? The prospective torpedo would certainly have active and passive sonar for targeting, just like current regular torpedoes, but the smaller housing necessitates lower performance sensors and data processing. What will happen to the reactors that are ‘dumped’ by the torpedo before its attack run? Presumably, this will occur in shallow coastal waters where trawling and other commercial activities may take place. While the torpedo’s operators may not care about the environmental and human consequences of a reactor drop on the other side of the ocean, they’d still have to worry about it during the testing stage. How will those operators retrieve reactors safely from the seafloor during testing? Rosatom engineers and technicians were killed when a test Burevestnik missile exploded during retrieval, in what was widely reported to be a criticality event.
The list of reasons not to develop and test these technologies is lengthy.
The general argument for this being a “safe” activity is that water is a good shielding material from radiation, and also very good at cooling hot things down, making the oceans the ultimate decay heat sink. And, while it’s true that water is very, very good at slowing – and eventually stopping – the radiation released by fission, reactors are so finicky that without precise knowledge of the environment in which a reactor finds itself, predicting its behaviour is difficult, to say the least. We cannot assume that a reactor “dumped” in an uncontrolled seabed environment won’t ever reach criticality, the point at which the fission reaction is self-sustaining, or even go beyond it. Neither can we assume that the fission products – the physical leftovers of a fission reaction that are highly radioactive – will not cause ecological damage downstream.
Russian prototype development of Poseidon, the nuclear-powered nuclear torpedo, and Burevestnik, the nuclear-powered nuclear cruise missile, have been ongoing for a number of years. They, too, are driven by autonomous reactors designed to provide huge amounts of power for a long time, allowing for (in the case of Burevestnik) an unknowable trajectory avoiding air defences, or (in the case of Poseidon) an “undefeatable”, fast moving nuclear torpedo for destroying aircraft carrier strike groups or coastal targets.
Grouping these three concepts together, Poseidon, the new Chinese design, and Burevestnik, have similar technical challenges and environmental impacts, despite their differences in warhead, medium and provenance.
The first of these challenges is autonomy. You cannot communicate with, and therefore control, any of these reactors once they’re launched – for Burevestnik, radio communications would defeat its stealthy purpose, for the torpedoes, you simply can’t communicate underwater at the bandwidth necessary to control a nuclear reactor over the ranges we’re discussing, and to surface and transmit information would sacrifice its stealthy advantage. As a result, all the reactor and navigational control systems must be autonomous and onboard, and capable of operating flawlessly in intense radiation.
We’re not there yet.
The second of these challenges is the safety of operators and handlers. We know more about this with respect to the new Chinese torpedo design, as the SCMP reports “Since the mini reactor would produce no radioactivity, service personnel could handle it as a “clean asset” without the need for protective gear”. This is misleading: the reactor fuel would produce low-level but still dangerous levels of radiation until it’s turned on – when it would become extremely dangerous. Given that the designers are quite open about stripping out much of the shielding and cheapening of the components to make the reactor “disposable”, contamination, accidental start-up, and general handling risks are unknown but orders of magnitude above safe levels.
It’s certainly not a “clean asset” as Guo says.
Thirdly, the elephant in the room. The environmental impact of a seafloor littered with the test carcasses of nuclear-powered torpedoes and cruise missiles. Water is a very good shielding material but hot, salty seawater will corrode the weapon housing. The gases produced by this corrosion could displace the cooling water leading to criticality, or build up only to be ignited in a chemical explosion once the reactor is brought to the surface and comes into contact with oxygen. It’s not hard to imagine a trawler dredging up a failed and missing test torpedo in the waters of the South China Sea. The torpedo may also remain neutrally buoyant after failure, which could lead to it being caught in fishing nets, or washing up on a beach, as has happened before.
So perhaps we should leave these reactors where they fall? They may not even be physically retrievable, given that there are political, economic, legal, or engineering factors that might prevent authorities from doing so. The ecological impact of leaving a possibly critical or supercritical reactor on the seafloor are unknowable at present, but they are unlikely to be good – and will be substantially worse if the asset is ever retrieved deliberately, or unknowingly.
It’s easy to see how leaders across the planet get excited by ‘invincible’ weapons, as the pursuit of national security is paramount for many, but we have to do more to counter the lure of reactor-driven platforms with unlimited range. These kinds of weapons are extremely resource intensive to develop, let alone test and field, and so it’s reasonable to assume that one of the drivers of this trend is changing threat perceptions. States are seeking new and sophisticated ways of achieving security, as they sense the old ones are no longer fit for purpose.
To counter this, a twin-track approach should be carried out, consisting of studies to determine the ecological and environmental impact of damaged reactors on the seabed (of which there are some that are mostly retrospective and incident specific), and multilateral, state-level discussions on what the rules should be regarding retrieval, transport, and decommissioning of such reactors.
The latter is crucial, as in the current climate, the West is unlikely to convince either China or Russia to stop testing and developing such technologies, so we must prepare for damage limitation. I’m not convinced anyone will make this technology work in the near term, but I am convinced that they shouldn’t try – at least until we know more about the second- and third-order consequences of doing so.