TOKYO -- Three major atomic accidents in 35 years are forcing the world's nuclear industry to stop imagining it can prevent more catastrophes and to focus instead on how to contain them.
Of the 176 new reactors planned across the globe, half will be in nations that had no nuclear plants when disaster crippled the U.S. Three Mile Island reactor in 1979 and the Chernobyl reactor blew up in present-day Ukraine in 1986.
As countries including China and India embrace atomic power even after the Fukushima reactor meltdowns in 2011 caused mass evacuations because of radiation fallout, scientists warn the next nuclear accident is waiting to happen and could be in a country with little experience to deal with it.
"The cold truth is that, no matter what you do on the technological improvements side, accidents will occur - somewhere, someplace," said Joonhong Ahn, a professor at the Department of Nuclear Engineering of University of California, Berkeley. The consequences of radiation release, contamination and evacuation of people is "clear and obvious," Ahn said. That means governments and citizens should be prepared, not just nuclear utilities, he said.
While atomic power has fallen from favor in some western European countries since the Fukushima accident - Germany, for example, is shutting all of its nuclear plants - it's gaining more traction in Asia as an alternative to coal.
Still, the association defends the global safety record of nuclear power, noting that the three high-profile disasters "are the only major accidents to have occurred in over 14,500 cumulative reactor-years of commercial nuclear power operation in 33 countries."
The problem is that the causes of the three events followed no pattern, and the inability to immediately contain them escalated the episodes into global disasters with huge economic, environmental and political consequences. Even if no deaths have yet been officially linked to Fukushima radiation, for example, cleanup costs have soared to an estimated $196 billion and could take more than four decades to complete.
If nuclear is to remain a part of the world's energy supply, the industry must come up with solutions to make sure contamination - and all other consequences - do not spread beyond station grounds, Gregory Jaczko, ex-chairman of the U.S. Nuclear Regulatory Commission, said in an interview in Tokyo.
"We have this accident and people will say, you know, it was caused by this and that," Jaczko said. "But the next accident is going to be something different. Nobody can tell you where or when or what exactly it is going to be. You really need to do more on the consequence side."
Since the introduction of nuclear stations in the 1950s, the industry has focused safety efforts on design and planning. Research and innovation has looked at back-up systems, passive technology that would react even if no human operator did, and strengthened materials used in construction of atomic stations.
The mostly engineering solutions were based on calculating the risk of radiation fallout on populations outside of the plant. In the U.S., a typical nuclear station has cut its probability risk of a fuel meltdown to once in 20,000 years of operation, according to the University of Michigan.
The analysis states that in two out of three fuel meltdowns there'd be no deaths, Michigan's Health Physics Society said in a report on its website. When the probability of 100,000 meltdowns are considered, the math says one of these would cause 50,000 deaths, the report said.
The usefulness of the math after the world's three major civilian accidents is academic, according to Jaczko.
"Once you have an accident, a low-probability and high consequence event, you can no longer call it a low probability event," Jaczko said. "It is an event that's happened and you cannot ignore the consequences simply because it was never supposed to happen. The consequences are real. Probabilities are always hypothetical."
It was not until the late 1970s that analysis and large-scale testing began of what could happen in case of an atomic accident, according to the London-based WNA. The WNA says that testing - and the Fukushima disaster in March 2011 - show that "even the worst possible accident in a conventional western nuclear power plant or its fuel would not likely cause dramatic public harm."
The official toll from the reactor explosion at Chernobyl was put at 31 deaths. Radiation clean-up work, however, involved about 600,000 people, while 200,000 locals had to be relocated.
The accident contaminated 150,000 kilometers of land and according to the last Soviet leader Mikhail Gorbachev it was a factor in bringing about the collapse of the Soviet Union in 1991.
In Japan, the meltdown of three Fukushima reactors helped unseat premier Naoto Kan and forced the evacuation of about 160,000 people, destroying local fishing, farming and tourism industries along the way. It also brought tens of thousands of anti-nuclear protesters out onto the streets in the country's biggest demonstrations since the 1960s. Tokyo Electric Power Co., the plant operator and once the world's biggest non-state power producer, would have been bankrupted by the Fukushima accident but for billions of dollars in government aid.
The $196 billion cleanup bill was an estimate in a March 2012 report by the Tokyo-based Japan Center for Economic Research. The cost of cleaning up Fukushima may be more than the total cost of building all the world's nuclear plants to date, Jaczko said.
On the back of Three Mile Island and Chernobyl, the International Atomic Energy Agency introduced a five-part, defense-in-depth concept in 1996. This made the management of an accident and dealing with radiation fallout part of the skills station operators needed.
Even so, the industry kept its focus on accident prevention rather than on what to do once a disaster hits, said Gennady Pshakin, a nuclear physicist at a research institute in Obninsk, site of Russia's first nuclear plant.
Building a plant that would contain an accident within the facility boils down to cold cash, he said.
"The demand may well be a fair one," Pshakin said. "But we could take it so far that it'd cost a phenomenal amount of money and everyone will say: 'So, why do we need this?' It's basically a debate between safety and economics."
For Jaczko, the industry's inability to resolve this issue could mean the end of nuclear generation.
"If we look at this technology and we challenge ourselves to make technology that meets this standard then we'll see that there are ways to do it," Jaczko said. "But if there aren't ways to do it - economically viable ways to do it - then we have to face the consequences of that decision. That means that this is perhaps then not a technology that we want to rely on well into the future."
Last year, spurred on by public reaction after Fukushima, a unit of the IAEA began a review of reactors to look at how the consequences of an accident can be mitigated, according to the Vienna-based organization's website.
The review calls for new reactor designs to make a major release of radioactive fallout outside the station site "practically impossible," the IAEA said. The standard would be "crucial for public acceptance and for the sustainability of nuclear energy."
Specialists on the review met for the first time in March and no conclusions are yet available, IAEA spokesman Greg Webb said by e-mail.
The problem with an engineering solution - an ever better reactor design or grander safety systems - is that based on the premise that all technology is fallible, those defense systems can also fail, Berkley's Ahn said.
"This is an endless cycle," Ahn said. "Whatever is your technology, however it is developed, we always have residual risk."
When the next nuclear accident occurs the world needs to have better knowledge of how to limit the spread of radiation and do the clean-up, including removing radiation from the soil, water and having an efficient evacuation drill for the population in danger zones, Ahn said. We also need more understanding of the impact of low-dose radiation on organisms, he said.
"This is about recovery from an accident, not preventing an accident," Ahn said. "It's completely different. And I think this concept is very necessary for the future of nuclear utilization."