Fukushima ten years on: Lessons learned
By Ian McKinley, Shinichi Nakayama and Susie Hardie
Ian McKinley, Shinichi Nakayama and Susie Hardie consider how recovery has progressed at Fukushima Daiichi and what lessons can be learned for the future
The great Tohoku earthquake and tsunami of 11 March 2011, which devastated north-eastern Japan, with the direct loss of over 18 000 lives and a legacy of displaced populations, damaged infrastructure and environmental pollution, has been ranked as the costliest natural disaster (≈ $300 billion) in world history. The damage to coastal infrastructure was immense, with impacts from loss of services, costs of replacement and remediation of resultant pollution – including that from the destruction of over 70 sewage treatment plants and damage to over 3300 facilities handling oil and potentially toxic materials from the petrochemical industry. However, internationally this is not well known as most focus has been on only one of these incidents – the reactor core melts at Fukushima Daiichi nuclear power station.
The huge, magnitude 9, undersea megathrust earthquake (largest recorded in Japan and fourth largest worldwide since records began in 1900) did little damage to the working reactors on the Fukushima Daiichi site, but it did knock out off-site power supply and extensively damaged regional communication, transport and monitoring infrastructure. In addition, the associated tsunami resulted in wave heights estimated to be around 15m, far larger than the reference maximum value (5.7m) and sufficient to flood the site and knock out all backup generators for units 1-4. The combination of the earthquake and tsunami resulted in a “perfect storm”, leading to core meltdown of three reactors (units 1-3) and subsequent hydrogen explosions generated by water-cladding-fuel interaction (including in the linked unit 4, which was defueled at the time), resulting in the release of volatile radionuclides that were subsequently deposited over land and sea.
After decay of short-lived isotopes of iodine, regional contamination is dominated by caesium radioisotopes Cs-134 and Cs-137, half-lives two and 30 years respectively, with the latter becoming more important over time. Decontamination activities have been continuous since the accident, supported by extensive R&D to improve understanding of the environmental behaviour of Cs in relevant ecosystems (https://fukushima.jaea.go.jp/en/). Although a wide range of approaches to surface decontamination were studied, low tech options have proven to be at least as effective as the high tech options investigated. Coordination of remediation planning and research carried out by numerous organisations has benefited from advanced knowledge management and communication tools that had already been developed by the Japan Atomic Energy Authority (for deep geological disposal) and which were quickly adapted for this application.
A consequence of remediation has been the generation of huge volumes of contaminated waste (≈14Mm3), much of which is unstable (eg, vegetation and soil subject to biodegradation). Many temporary waste storage sites throughout the region were established based on the assumption that better centralised interim storage facilities (ISFs) would be available within three years. Technical and political delays resulted in them becoming operational only in late 2017, resulting in many problems associated with waste recovery from the old temporary stores. The ISFs include incineration and soil sorting facilities and are included within a plan to recycle materials to the extent possible and finally dispose of the rest at an undefined site outside of Fukushima prefecture (http://josen.env.go.jp/en/storage/). With the value of hindsight, it is clear that waste management has been the least cost-effective aspect of remediation and is still particularly constrained by decisions made for political reasons and associated permitting problems.
Taken together with natural dose reduction processes like mobilisation and dispersal of Cs containing soil particles (so called “self-cleaning”), remediation actions have resulted in average dose decreases by a factor of about two more than radioactive decay alone (Figure 1).
Indeed, dose reductions have been higher in urban areas, thus supporting a more rapid return to evacuated zones. Apart from areas close to the Fukushima Daiichi site, highest doses are now in forested areas where decontamination is impractical and constraints on use (particularly harvesting of mushrooms and game) are expected to continue for a considerable time.
On a regional basis (eastern Tohoku) there are two coupled issues – the impacts of the earthquake/tsunami and those from the reactor accident. In terms of the former, which was responsible for most of the displaced populations and all of the direct deaths, there has been a clear reluctance of evacuees to return to coastal towns and villages. Partially this is due to the slow rate of recovery from the extensive damage to infrastructure, but also the trauma of the event and the large number of deaths of friends and family. A significant proportion of those displaced have already re-established themselves in other parts of Japan and will not return.
For the area around Fukushima Daiichi there has been significant progress in the return of populations to the areas that were less contaminated as remediation progressed. Certainly, this was helped by new sources of employment in these areas, such as off-site remediation work and associated R&D to support radiation monitoring, decontamination activities and 1F decommissioning (eg, CLADS - Collaborative Laboratories for Advanced Decommissioning Science, see https://fukushima.jaea.go.jp/en/ for further details). Other sources of work in this area are 1F itself (stabilisation and clean up activities) and other related facilities (eg, ISF – interim storage facility – Figure 2). On balance, the little recognised conclusion is that recovery in the areas around 1F has been more extensive (and much better funded) than other areas impacted by the earthquake / tsunami.
On site recovery
The contrast with Chernobyl is marked, both off and on site, emphasising the very different nature of these accidents. Despite the dramatic images on television, most radioactivity – and almost all of the more toxic radionuclides – were contained within the reactor buildings. This made off-site recovery simpler but does, however, make on-site remediation particularly challenging.
In the aftermath of the accident, the initial primary concern was removing decay heat from the three scrammed reactors, after which stabilising the damaged reactor buildings was the key priority. Based on a US design, the fuel storage pools were positioned above the primary containment, meaning they were particularly vulnerable to further disturbances such as aftershocks or further earthquakes. Additionally, to facilitate access and reduce worker doses, contaminated debris distributed all around the site had to be removed before stabilisation work could be carried out.
Following indecision due to the inevitable acceptance of loss of reactors, the reactors were flooded with seawater only after core melting had commenced – a brave decision driven by the reactor operators. Pumping of cooling water through the primary containment since then has led to continuous leaching of radioactivity from the damaged fuel. Leakage of groundwater into the damaged reactor building basements has resulted in continually increasing volumes of contaminated water (currently ≈ 1Mm3) that is being stored on site – giving it the appearance of a tank farm (Figure 3).
Although the implementation of a decontamination system (ALPS) to allow recycling of cooling water was successful, attempts to restrict water inflow have been associated with serious problems. The main solution chosen, construction of an underground “ice-wall” around the reactors (with refrigerant flowing through steel pipes down to a depth of 30m), was an unproven, very expensive, high tech option that was difficult and dangerous to construct, requiring continuous upkeep and did not approach promised performance levels.
Here a contrast is seen with the off-site options that were tested and critically evaluated before implementation: the selection process for the ice wall was opaque, based on industrial lobbying and proceeded without response to serious concerns raised by many technical experts. Even now, the ice wall has served only to reduce – not stop – groundwater inflow, but limitations are not openly acknowledged and little effort has been made to consider alternatives, assess how recovery from perturbations will be managed or discuss how this will impact eventual site decommissioning.
In any case, the large volumes of low-level contaminated water certainly represent a hazard and, although releases to sea are now being discussed, progress is very slow. Associated R&D has also been poorly coordinated, appearing to reflect bottom-up proposals (eg, including studies of tritium removal from contaminated water) rather than a top-down structured process as adopted off-site.
Stabilisation of the damaged reactor buildings and, in particular, removal of fuel from the storage pools has progressed effectively, even if slowly due to the constraints of working on site. The most problematic pool for the 1F site was that of unit 4, which contained a full core load having being defueled at the time the accident took place. Unit 4 fuel pool has since been cleared of all stored fuel (completion December 2014) and more recently removal of fuel from the unit 3 pond commenced, in April 2019, with the aim of completion by March 2021. Removal of the rods from Units 1 and 2 is to begin in 2023. Additionally, extensive technology development has resulted in much more robust tele-operated or autonomous robotic systems for characterising the high dose rate areas within the spent fuel pools and reactor primary containment, with special emphasis on locating and characterising the fuel, from an initial total of 800 t, that melted through the reactor pressure vessels (often termed corium, but in Japanese reports grouped together with other damaged fuel and reactor internals as “fuel debris”). The resulting information on each of the damaged reactor cores can be captured in 3D using a virtual reality system, which facilitates planning future decommissioning actions (Figure 4).
Currently the government target is to initiate removal of fuel debris in 2021, which is both very ambitious and is in advance of establishing a clear plan for reactor decommissioning and management of associated waste (expected only in 2031). The cost, risk and environmental impact of decommissioning the 4 damaged units, especially 1-3, could be minimised by reducing the extent to which large components need to be cut down for packaging – although this is inherently linked to the final waste disposal approach adopted (as considered further below). Additionally, there are also benefits to be gained by minimising treatment / conditioning of the most problematic fuel debris.
Easing access and reducing on-site dose rates has resulted in huge volumes of contaminated waste, estimated to total ≈ 770 000m3, including ≈ 150 000m3 of rubble and ≈ 90 000m3 of felled trees in the more active waste category. This is in addition to 7000 “high integrity containers” containing more toxic secondary waste (slurry, ion exchangers, filters, etc.) from the cooling water decontamination units.
Although not assessed in detail as yet, the reference decommissioning plan calls for Fukushima Daiichi to be cleaned up to the level of a greenfield site within 40 years of the accident – an extremely ambitious goal without any international precedent. It is further assumed that waste will either be recycled or disposed of off-site – resulting in waste management costs that correspond to over 80% of the estimated budget (≈ US$ 90 billion). In addition to all surface contamination, this will also require an undefined extent of underground contamination from water leakage to be handled. From a purely technical viewpoint, there is a huge potential to reduce costs, risks to workers and environmental impact, if it is accepted that final site end state is brownfield and waste is disposed of locally to the extent possible. Indeed, benefits to the local community could be increased if the end state were to be defined as a national waste management facility, especially given the number of nuclear power plants that will eventually need to be decommissioned in Japan (Figure 5).
However, this concept is extremely politically sensitive and would require appropriate communication with local communities to gain acceptance.
Although rarely mentioned, it is worth noting that one of the three emergency diesel generators installed at Fukushima Daiichi 6 (both units 5 and 6 were offline at the time of the accident) remained available after the tsunami waters had receded, and operators were able to successfully connect it to both units 5 and 6, and providing sufficient power for accident management. This surviving generator was air-cooled, not water-cooled, and was installed at an elevation high enough to avoid the tsunami. This resulted in a defence-in-depth safety concept, providing resilience for a very unlikely event and emphasises the value of implementing protection against worst case perturbations on site rather than debating the likelihood of events that could lead up to them.
Even though Japan is recognised as having one of the best natural hazard defence systems in the world, the combination of such a large magnitude earthquake and associated large tsunami resulting in a nuclear accident involving three reactors had never been considered credible. If catastrophe response plans had allowed the seriousness of the situation to be recognised sooner and external support mobilised more quickly (eg, supply of emergency generators by helicopter, perhaps by the military), it is likely that the accident could have been prevented.
Communication failed at every level. Before the accident, known tsunami risks were not communicated to those charged with assessing or regulating hazards to nuclear facilities. The developing situation in the damaged units after the tsunami was poorly communicated (including down-playing) to both the government and the general public, delaying decisions that could have limited the consequences and helped avoid unnecessary panic. Tools that could have been used to help plan the response to releases of radioactivity were introduced too late, and their output was presented in a way that only increased confusion and public concern. Both evacuation and much of the decontamination was not strictly required from a radiation protection perspective. A rich country like Japan can justify such actions, even if the health benefit is marginal compared to the effort invested, but such an objective perspective on the risks involved should be clearly explained in order to reassure local residents and involve them in decisions about the actions that need to be taken.
The consequences of this accident have been consistently exaggerated, especially by sensationalist media and, rather unashamedly, by anti-nuclear groups to further their political agendas. From a purely radiological perspective, the environmental consequences of the core meltdowns were significantly less than a worst-case scenario and, media reports to the contrary, nothing at all like Chernobyl. Nevertheless, better communication, both internally (eg, accident response team) and externally (eg, to the Japanese and international communities) is a recognised need – requiring a change in culture to facilitate collaboration and presentation of results to key non-technical audiences. It should be emphasised that this was not simply a result of the chaos immediately after a major disaster nor was it limited to Japan - 10 years on, newspaper, TV and Internet articles continue to misrepresent the accident.
As repeatedly noted, waste management off-site was the weakest part of recovery actions and clear, top-down concepts for effective management of the huge volumes of waste produced have yet to be developed. For example, major efforts are being invested to manage lightly contaminated soil, but complete decoupling of on- and off-site work prevents consideration that this soil could have been better utilised as ground cover on the more highly contaminated reactor site, providing radiation shielding and reducing doses to workers. Many problems experienced resulted from the absence of required regulatory infrastructure and hence pre-emptive legislation to facilitate accident response would have been invaluable.
On-site, waste is even more problematic and is closely coupled to the short-term management of contaminated water and later decommissioning of the damaged reactors. Here decisions on approaches adopted and plans for the future are still based on compartmentalised thinking and constrained by assumed socio-political boundary conditions which have not been developed or assessed in a logical structured manner. There is a huge potential to minimise risks, costs and environmental impacts while maximising benefits to local communities, but this will require introducing more modern approaches to managing complex, multidisciplinary projects (eg, based on advanced knowledge management tools) – the development of which could be a priority for international collaboration.
The communication problem extends to wider, international issues resulting from this incident. In several cases, knee-jerk reactions from poorly informed politicians (eg, in Germany) have resulted in moves away from nuclear power, without any balanced consideration of the potentially larger environmental and public health hazards of the fossil fuel alternatives that would be introduced in its place. More seriously, the fact that two recent devastating tsunamis have resulted from megathrust earthquakes has focused risk assessment on this particular hazard combination - without considering that the historical record shows much larger tsunamis from other sources, such as volcanoes, landslides and sea mount collapse. Indeed, much of the infrastructure that supports our high population densities is located in areas where a natural catastrophe in coming decades is not just possible, it is inevitable. The next major natural catastrophe to hit an industrialised country might look quite different, but such a disaster is certain to happen and its consequences will be even greater if we don’t learn from the painful and costly experience gained in Japan.