After a devastating tsunami left 18,000 people dead in 2011, Japan was about to face a potentially more significant disaster as several reactors at the Fukushima Daiichi nuclear power plant started to melt down. More than 300,000 people were evacuated.

Since then, the Japanese government has tried to defend the Tokyo Electric Power Co. (TEPCO), which operated the plant, and offer reassurances that the country’s nuclear reactors are secure. A study released Monday in the science journal Philosophical Transactions reaches a different conclusion, however: “The Fukushima accident was preventable, if international best practices and standards had been followed, if there had been international reviews, and had common sense prevailed in the interpretation of preexisting geological and hydrodynamic findings.”

The study’s authors paint a bleak picture of TEPCO’s failures before the disaster, as well as the company’s handling of the crisis. Using documents provided by the U.S. National Research Council, the U.S. Nuclear Regulatory Commission, Japan’s bicameral legislature and TEPCO itself, they conclude: “Had the TEPCO modellers had any experience with tsunamis, they would have had immediately recognized that their ‘high’ resolution predictions were underestimating the hazard.”

Researchers Costas Synolakis and Utku Kanoglu allege that the accident revealed striking inaccuracies in TEPCO’s internal risk analysis, as well as a “cascade of engineering and regulatory failures.”

“The entire experience with TEPCO’s pre-event internal studies not to mention the entire methodology that has been used in Japan to assess tsunami hazards points to the perils of insularity,” the study’s authors wrote.

According to Synolakis and Kanoglu, TEPCO researchers had long known that earthquakes could threaten power plants in the region. Despite that, necessary safety measures were not implemented.

The researchers also question the designs of some Japanese nuclear power plants, which could leave certain structures more vulnerable than others. “Interestingly, while the Onagawa nuclear power plant was also hit by a tsunami of approximately the same height as Dai-ichi, it survived the event ‘remarkably undamaged.’” The differences in vulnerability could partially be due to methodological mistakes “which almost nobody experienced in tsunami engineering would have made,” according to Synolakis and Kanoglu, who warn that similar flaws could lead to more accidents in the future.

“When it comes to studying hazards or designing structures whose catastrophic failure will transcend national boundaries, even countries with sophisticated technologies need to take note of Godel’s incompleteness theorem,” they wrote, referring to a mathematical concept most commonly interpreted as signifying the limits of provability.

Regulatory measures could have prevented the 2011 accident, but the researchers found “substantial inadequacies” there, as well.

The allegations are among the most extensive — but they are far from being the first. The Japanese government under the leadership of Shinzo Abe has refuted allegations of structural failures in the past and said that its response to the disaster had been adequate. Shortly afterward, the government decided to restart many of the country’s nuclear power plants, despite protests and safety concerns.

Meanwhile in Fukushima, the nuclear power plant is far from being secured: Hundreds of fuel rods stored nearby have not been removed.