1. Introduction

Nuclear experts and scientists represented by the International Atomic Energy Agency (IAEA) and United Nations Scientific Committee on the Effects of Atomic Radiation (UNSCEAR) reported that a severe accident consequences of the Fukushima accident do not significantly challenge the health of the general public or environmental contamination. IAEA ^[1] reported that “the release of radionuclides, and the corresponding doses to nonhuman biota occupying areas of high deposition in Japan, was much lower than in areas around Chernobyl.” United Nations Scientific Committee on the Effects of Atomic Radiation (UNSCEAR) ^[2] reported that “the doses to the general public, both those incurred during the first year and estimated for their lifetimes, are generally low or low. No discernible increased incidences of radiation-related health effects are expected among exposed members of the public or their descendants.”

However, certain countries became reluctant in building new nuclear power plants and pursued other alternatives such as regenerative energy soon after the Fukushima accident. Germany made a drastic decision to fade out nuclear power plants by year 2022 ^[3]. In the year 2017, new president in Korea announced that Korea would halt plans to build new nuclear power plants and would not extend the lifespans of existing ones ^[4]. Before the Fukushima accident, strict regulation was enforced only for a design basis accident, while regulation for a severe accident was minimal. The risk of severe accident was considered to be residual. There was little practice of enforcing back-fit on operating reactors to prevent or to mitigate severe accident. After the Fukushima accident, the safety of the nuclear power plant has been significantly improved in a global scale ranging from an establishment of formal regulation for a severe accident to strengthening the mitigation measures for unexpected accidents and natural hazards.

In Japan, strong regulation on nuclear reactor safety was enforced by the Nuclear Regulatory Agency (NRA) soon after the Fukushima accident. The NRA stated that a “so-called “safety myth” had critically impeded efforts for nuclear safety in Japan before the accident at Fukushima Daiichi nuclear accident; however, more stringent regulations have been developed with an underlying assumption that severe accidents could occur at any moment ^[5].” United State Nuclear Regulatory Commission (USNRC) strengthened the regulation for better safety by employing various measures: adding capabilities to maintain key plant safety functions following a large-scale natural disaster, updating evaluations on the potential impact from seismic and flooding events, new equipment to better handle potential reactor core damage events, and strengthening emergency preparedness capabilities ^[6].

New regulations for a severe accident adopted in global nuclear community allows only a fractional amount of radiological release at a low probability. As an example, Japanese Nuclear Regulatory Agency ^[5] requires that the frequency of an accident that causes discharging 

Cs over 100TBq should be reduced to not exceed one in one million reactor years. The amount of 100TBq of 

Cs is less than 1% of the release from the Fukushima accident, such that radiological impact would be minimal. Therefore, it is fair to say that the current fleet of nuclear power plants is at least better than before.

The probability argument element in the new regulation for a severe accident is based on a cost-benefit analysis ^[7]. An expert in the nuclear sector would easily accept that it will be extremely unlikely to experience another severe accident in his or her lifetime. However, nuclear experts and the public have different perceptions for nuclear risk. Recent studies shows that experts perceive radiological risks differently from the general public. Experts’ perception of medical X-rays and natural radiation is significantly higher than in the general population, while for nuclear waste and an accident at a nuclear installation, experts have lower risk perception than the general population ^[8]. The public needs to be convinced to understand the views from the nuclear expert.

The different views among nuclear experts and the public raised several issues, which called for international attention after the Fukushima accident. Prominent issues include uncertainties of severe accident analysis, more emphasis on nuclear safety, decommissioning of Fukushima nuclear power plant, discharge of radioactive water, and safety of fishery products. The author revisited these issues in a broader view where both the public and the expert engaged to provide perspectives on severe accident consequences.

2. Severe Accident Analysis Methodology

To investigate the accident progression and radiological releases in the Fukushima accident, there were research efforts on an international scale, such as the Nuclear Energy Agency of the Organization for Economic Development (OECD/NEA) project of Benchmark Study of the Accident at the Fukushima Daiichi Nuclear Power Plant (BSAF) ^[9]. The state-of-the-art computer codes including ASTEC ^[10], MELCOR ^[11], MAAP ^[12] were employed to predict and interpret the accident progression, including core damage, reactor vessel failures, and radiological releases to the environment in three units of the Fukushima Daiich nuclear power plant (FDNPP). Progression of core damage and resulting radiological releases were analyzed for unit 1, 2, and 3 of the FDNPP ^{[13][14][15][16][17][18]}. Yet, it had to be forensic in nature because initial and boundary conditions for the accident progression, such as responses of various components of the nuclear power plant at off-design conditions and net effects of operator actions cannot be confirmed ^[17][18]. The behavior of the components in the off-design conditions of high temperature, pressure, and radiation experienced during a severe accident were uncertain ^{[9][13][14][15][16][17][18]}.

Atmospheric dispersion calculations for the radiological releases to the environment during weeks after the accident were performed, but the results were not consistent with the measured terrestrial deposition pattern due to incomplete weather data and uncertain release histories from each unit ^[19].

Physical models for a severe accident phenomena had to accompany uncertainties because they handled the behavior of complex system, which consists of multi-phase (liquid, vapor, droplet, aerosol) and multi-component (water, molten fuel at various compositions of different materials, non-condensable gas) subsystems with complicated geometric configuration in a wide range of pressure and temperature. Therefore, the correlations employed for the fluid flow and heat transfer in a severe accident analysis code have model parameters to allow a range of predictions. As degradation of core fuels and neighboring materials at high temperature near 3000 K experiences phase change, it has to handle thermodynamics of multi-component system of UO

, Zr, ZrO

, stainless steel, and B

C ^[20]. In addition, phenomenon-specific model parameters such as flame speed for the hydrogen combustion calculation, range of the size of aerosols for fission product transport, size particulate debris bed for heat transfer, and melting temperature of composite materials are allowed to have a range ^[10][11][12].

Therefore, the accident progression and radiological releases were predicted with uncertainties. Recently, systematic approaches to handle these uncertainties in the phenomenological and physical modellings of a severe accident analysis have been developed ^[21][22], which enabled handling of the uncertainties in a rigorous statistical manner ^[23]. Typical uncertainty calculation included sensitivity studies on sequence related parameters such as primary safety valve stochastic number of cycles until failure-to-close, in-vessel accident progression related model parameters such as Zircaloy melt breakout temperature, molten clad drainage rate, and ex-vessel accident progression related model parameters such as hydrogen ignition criteria, chemical form of iodine, aerosol dynamics shape factor, containment convection heat transfer coefficient ^[21].

Using Monte–Carlo or Latin hypercube sampling methods, about 300–1000 calculations were performed for a typical scenario, such as station black out (SBO), whose accident progression is similar to that of Fukushima. Widespread radiological release was observed between calculated median, mean, and 95th percentile curves ^[21][22]. These analyses results are integrated with the MACCS (MELCOR Accident Consequence Code System) ^[24] analyses to predict latent cancer fatality (LCF) risk statistics. MACCS models atmospheric transport and deposition of radionuclides, emergency response, exposure pathways, early and latent health effects to the population.

The above methodology required a significant number of MELCOR and MACCS calculations up to 300–1000; however, it became feasible with recent advances in the computer hardware. Severe accident consequence analyses with rigorous uncertainty quantification would not only increase the confidence and accuracy of the results of radiation exposures to the public and health effects, but it also would help communication with the public or scientists in other fields.

3. Views on the Nuclear Reactor Safety

The continued investigations on the status of the damaged reactors of the FDNPP confirmed that molten fuels discharged out of the reactor vessel, and they were relocated on the floor of the primary containment vessel and interacted with the concrete structure of the floor in all three units. Cosmic muon radiography was utilized to visualize the presence of the reactor fuels in the reactor vessel ^[25]. The reactor vessel failed, and a mixture of steam, fission products, and hydrogen gas was released out of the primary containment vessel. This type of severe accident consequence is different from that of Three Mile Island (TMI) ^[26] with negligible radiological releases to the environment, where reactor vessel failure was prevented, although a significant amount of core was molten.

The radiological releases in the Fukushima accident consist of atmospheric release and direct release of contaminated liquid into the ocean. Atmospheric releases are in the form of gas and aerosols transported to the land and sea by the wind and rainfall.

The numbers reported by UNSCEAR ^[2] can be considered as representative values. The total releases to the atmosphere of 

I and 

Cs ranged generally between 100–500 PBq and 6–20 PBq, respectively. The releases of less volatile radionuclides (e.g., 

Sr and 

Pu) were negligible. The total releases of 

I and 

Cs to the atmosphere are 120 and 10 PBq, respectively. Releases of radionuclides into the Pacific Ocean occurred directly and indirectly. They comprise: (i) direct releases in the first three months amounting to about 10 to 20 PBq of 

I and about 3 to 6 PBq of 

Cs; (ii) deposition of radionuclides on to the ocean surface following their release to, and dispersion in, the atmosphere, amounting to about 60 to 100 PBq of 

I and 5 to 8 PBq of 

Cs.

While the amount of atmospheric release was about 1/10 of the one in the Chernobyl accident, the amount of direct release to the ocean was in the same order as that of Chernobyl. Atmospheric release consisted of noble gases such as Xe, Kr, cesium and iodine, mainly in the form of aerosols. As half-life of 

Cs is about 30 years and transported by the wind and finally deposited on the terrestrial land and ocean, we are highly concerned about this species. Although noble gases such as Xe and Kr were released more than in the Chernobyl accident, they would be fully dispersed into the atmosphere in the form of gas. To protect the public in the contaminated area, decontamination efforts for the soil and plant were carried out in a massive scale. In addition, foods are screened with proper regulatory limits to prevent consumption by the public.

Fortunately, no casualty due to radiation was reported. However, there were large numbers of casualties due to relocation-related stress in the Fukushima accident. The Fukushima mental health and lifestyle survey disclosed that the Fukushima accident caused severe psychological distress in the residents from evacuation zones ^[27]. The effects of a major nuclear accident on societies are diverse and enduring. One of important issue for the future would be the establishment of counter measures for the potential relocation due to major nuclear accidents. The countermeasures should include disaster management, long-term general public health services, mental and psychological care, behavioral and societal support, in addition to efforts to mitigate the health effects attributable to radiation ^[27].

After the Fukushima accident, IAEA compiled an effort to assess the safety of the FDNPP in various aspect including relation to external events, failure to maintain fundamental safety function, treatment of beyond design basis accidents, accident management provisions and implementation, regulatory programs, human and organizational factor ^[28]. Song and Kim ^[29] also discussed the issues on the nuclear reactor safety raised by the Fukushima accident.

The safety of nuclear power plants has been significantly improved after the Fukushima accident on a global scale due to efforts such as the EU-stress test ^[30], new regulations adopted in Japan ^[5], strengthening of the mitigation features in the United States ^[4], such that that we can claim that the current fleet of nuclear power plants is at least better than before. However, the comment from the director general of IAEA on the nature of Fukushima accident ^[31] below gives us a warning. “A major factor that contributed to the accident was the widespread assumption in Japan that its nuclear power plants were safe that an accident of this magnitude was simply unthinkable. This assumption was accepted by nuclear power plant operators and was not challenged by regulators or by the Government. As a result, Japan was not sufficiently prepared for a severe nuclear accident in March 2011.”

The chance of a severe accident accompanying large release of radionuclides cannot be denied, although the probability of occurrence is low. However, the consequences would be limited by new engineering safety features imposed after the Fukushima accident in many countries. It is fair to say that there is a possibility of nuclear accident with large radiological releases but with limited casualties due to radiation. New regulations for a severe accident adopted in global nuclear community allows only a fractional amount of radiological releases at low probability. However, as there are uncertainties in the modeling of severe accident phenomena, the limits of 100 TBq of 

Cs and the frequency ^[5] had to be properly estimated with proper confidence levels, considering the uncertainties. As there has been little practice in the new regulatory frame yet, extensive efforts need to be pushed to guarantee the safety of the nuclear power plant.