This blog post draws on data and research discussed in our entry on Energy Production & Changing Energy Sources.
This post has been extended by the author on 26th July 2017.
When it comes to the safety of nuclear energy, discussion often quickly turns towards the nuclear accidents at Chernobyl in Ukraine (1986) and Fukushima in Japan (2011). These two events were by far the largest nuclear incidents in history; the only disasters to receive a level 7 (the maximum classification) on the International Nuclear Event Scale.
How many deaths did each of these events cause?
When it comes to nuclear accidents there are really two fatal impacts to consider: the first being the number of direct deaths which occurred either at the time of incident, or in the days which followed (i.e. the acute impacts); the second being the long-term (chronic) impacts of radiation exposure, which has known links to the incidence of several forms of cancer.
# Deaths from Chernobyl
In the case of Chernobyl, 31 people died as a direct result of the accident; two died from blast effects and a further 29 firemen died as a result of acute radiation exposure (where acute refers to infrequent exposure over a short period of time) in the days which followed.1
The number of people who were impacted over long-term radiation exposure is more difficult to discern and remains highly contested. Part of this difficulty lies in the methodology used to estimate long-term deaths from low-level radiation exposure. In the published estimates below, studies have utilised a methodology termed the ‘linear no-threshold model’ (LNT); this model is typically applied in assessments of radiation risk and in setting regulatory limits for environmental protection. However, the LNT method remains strongly contested, and is assumed to provide a conservative estimate of potential mortality [we have provided a short discussion on the LNT model and its implications in the technical notes at the end of this post]. As such, we may expect that the numbers quoted below to be interpreted as the upper limit of a given source’s estimate.
The chart below reflects a range of published estimates on the number of deaths resultant from the Chernobyl disaster. In its 2005/06 assessment ‘Chernobyl’s Legacy: Health, Environmental and Socio-Economic Impacts’ the World Health Organisation (WHO) estimated that the total number of long-term deaths will be around 4,000.2 However, this figure is related only to the proximate populations of Ukraine, Russia and Belarus which were exposed to high radiation levels; if extended to estimates of those exposed to low-level radiation across the region, this number rises to 9000.3 Other studies have suggested higher figures. A study in the International Journal of Cancer by Cardis et al. (2006) estimates a total of 16,000 deaths across Europe.4 Radiation scientists Fairlie and Sumner provide some of highest estimates, predicting between 30,000-60,000 deaths.5
The challenge of cancer risk attribution- especially at low doses of exposure in further geographic regions- makes this process of estimation particularly difficult. In its 2008 report, the United Nations Scientific Committee on the Effects of Atomic Radiation (UNSCEAR) refrained from quoting a figure for the absolute number of deaths within populations exposed to low radiation doses from Chernobyl because of uncertainty in the limit no-threshold model and “unacceptable uncertainties in the predictions”.6 As shown in the chart below, most published estimates lie in the range between thousands and tens of thousands.
# Deaths from Fukushima
In the case of Fukushima, although 40-50 people experienced physical injury or radiation burns at the nuclear facility, the number of direct deaths from the incident are quoted to be zero. However, mortality from radiation exposure was not the only threat to human health: it’s estimated that around 1,600 people died as a result of evacuation procedures and stress-induced factors. This figure ranges between 1,000-1,600 deaths from evacuation (the evacuation of populations affected by the earthquake and tsunami at the time can make sole attribution to the nuclear disaster challenging). Stress-induced deaths affected mostly older people; more than 90 percent of mortality occurred in individuals over the age of 66.
How many are projected to suffer in the long-term from low-level radiation exposure? In its Health Risk Assessment of the nuclear disaster, the World Health Organization (WHO) note exposure levels too low to affect human health for the national population, with exception to a few communities in closest proximity.7 In these localities, it is those who were infants at the time of exposure who are at greatest risk of cancer—at the two closest sites, the incidence of cancer in this demographic is projected to be between 4-7 percent higher than baseline cancer rates for both males and females (with the exception of thyroid cancer in females, which is 70 percent higher). The WHO project the number of deaths from low-level exposure to be close to zero, and up to 400 in upper estimates.
In the chart below we have shown the estimated number of total deaths by attributed cause.
In more recent evaluations of rates of perinatal mortality (that is, stillbirths or deaths within the first week of life) in areas closest to the Fukushima site, there were no statistical indications of increased incidence.8 In fact, rates of perinatal mortality showed an overall decline with time—the general trend we see through improved healthcare and healthier lifestyles.
The death toll of the Fukushima nuclear accident dominated headlines for weeks after the event and overshadowed the much larger tragedy that happened at the same time and place: the Tsunami killed 15,893 people, around eight times (if we assume the upper estimate of the long-term death toll from the nuclear incident) the number of the nuclear accident.9
# Why was the death toll from Chernobyl so much higher than Fukushima?
Chernobyl and Fukushima are the only two disasters to receive a level 7 (the maximum classification) on the International Nuclear Event Scale. But why is the upper estimate of deaths from Chernobyl almost fifty times higher than that of Fukushima?
There are a couple of factors which are likely to have played a key role here. The first of these concerns the technical functionality and safety measures of the respective nuclear facilities. Chernobyl occurred 25 years prior to Fukushima; it was the first instance of a nuclear accident at this scale. From a technical perspective, the nuclear reactors at Chernobyl were poorly designed to deal with such a scenario. Its fatal RBMK reactor had no containment structure, allowing radioactive material to spill into the atmosphere (in contrast, Fukushima’s reactors had steel-and-concrete containment structures, although it’s likely that at least one of these were also breached). Crucially, the cooling systems of both plants worked very differently; at Chernobyl, the loss of cooling water as steam actually served to accelerate reactivity levels in the reactor core, creating a positive feedback loop towards fatal explosion (the opposite is true of Fukushima, where the reactivity reduces as temperatures rise, effectively operating as a self-shutdown measure).
These technical differences undoubtedly played a role in the relative levels of exposure from both events. However, the governmental response to both events is also likely to have played a crucial role in the number of people who were exposed to high levels of radiation in the days which followed. In the case of Fukushima, the Japanese government responded quickly to the crisis with evacuation efforts extending rapidly from a three kilometre (km), to 10km, to 20km radius whilst the incident at the site continued to unfold. In comparison, the response in the former Soviet Union was one of denial and secrecy.
It’s reported that in the days which followed the Chernobyl disasters, residents in surrounding areas were uninformed of the radioactive material in the air around them. In fact, it took at least three days for the Soviet Union to admit an accident had taken place, and did so after radioactive sensors at a Swedish plant were triggered from dispersing radionuclides. It’s estimated that the delayed reaction from the Soviet government and poor precautionary steps taken (people continued to drink locally-produced, contaminated milk, for example) led to at least 6,000 thyroid cancer cases in exposed children.
Whilst prevention, and ultimately containment (which are predominantly technical issues), are crucial to the safety of nuclear energy production, these two events also highlight the importance of political governance and response in the aftermath of such disasters.
# Risk of nuclear in context
The potential risks of nuclear energy are real: in both Chernobyl and Fukushima, deaths occurred as a result of direct nuclear impacts, radiation exposure and psychological stress. Nonetheless, of the two largest nuclear disasters, the death toll was of the order of thousands to tens of thousands in one, and thousands in the latest. Arguably still too many, but far fewer than the millions who die every year from impacts of other conventional energy sources.
As covered in a separate blog post on the relative safety of energy sources, the comparatively low death toll from nuclear energy (resulting in 442 times fewer deaths relative to brown coal per unit of energy, even with radioactive exposure deaths included) is largely at-odds with public perceptions, where public support for nuclear energy is often low as a result of high safety concerns. The key distinction here is that nuclear risk is generally focused within low-probability, high-impact single events in contrast to air pollution impacts which provide a persistent background health risk.
The process of estimating the number of potential deaths attributable to radiation exposure is a complex and contested process. The selection of particular methodologies used to carry out such assessments are strongly contested.
The most common approach applied (and which has been utilised in the published estimates we reference in this post) is that of the linear no-threshold (LNT) model. The LNT model assumes that cancer risk holds a linear relationship with radiation dose (e.g. a doubling of dosage would double cancer risk). This model also assumes that there is no lower limit to this relationship.
The LNT model is typically applied in the context of radiation protection, and typically adopted by governmental organisations in nuclear risk assessments. However, the application of the LNT model is widely contested: since it has no lower threshold, this model suggests that even very low dosages of radiation increase cancer risk. As a result, it is suggested that models which estimate the number of deaths using the LNT methodology may provide an overestimation, especially within populations which experience only low radiation exposure. We may therefore expect the figures quoted above to provide a conservative (upper) estimate of long-term deaths from radiation exposure.
No accepted consensus on the LNT methodology has been reached amongst governmental, scientific and regulatory bodies. In its 2000 report, the United Nations Scientific Committee on the Effects of Atomic Radiation (UNSCEAR) remained cautiously in favour of the LNT method, stating: Until the […] uncertainties on low-dose response are resolved, the Committee believes that an increase in the risk of tumour induction proportionate to the radiation dose is consistent with developing knowledge and that it remains, accordingly, the most scientifically defensible approximation of low-dose response. However, a strictly linear dose response should not be expected in all circumstances.”
However, in its 2008 report UNSCEAR refrained from quoting a figure for the absolute number of deaths within populations exposed to low radiation doses from Chernobyl because of “unacceptable uncertainties in the predictions. It should be stressed that the approach outlined in no way contradicts the application of the LNT model for the purposes of radiation protection, where a cautious approach is conventionally and consciously applied. 10