THE WASHINGTON POST
THURSDAY, May 4, 1989

By Cass Peterson
Washington Post Staff Writer

Five to eight times more radioactive dust was released from a nuclear weapons plant in Ohio than the Energy Department estimated, according to a private analysis based on previously unreleased engineering documents and soil samples near the plant.

The analysis estimates that as much as 3.1 million pounds of uranium dust and other radioactive materials were released from the Feed Materials Production Plant near Fernald between 1951 and 1985. The Energy Department has previously put the figure at 395,000 to 555,000 pounds.

The Fernald plant produces uranium ingots and rods to be fabricated into fuel rods for reactors in producing material for nuclear weapons. The amount of uranium released into the environment is critical to health studies of the surrounding population that the government has agreed to conduct.

The report was prepared by the Institute for Energy and Environmental Research, using materials gathered as part of a lawsuit against Fernald's former operator, National Lead of Ohio. The lawsuit was filed by residents of the area who contend tend that radioactive releases from Fernald have lowered property values and damaged their health.

Arjun Makhijani, coauthor of the report, said that precise estimates of uranium losses from Fernald may be impossible. Of 430 potential sources of dust pollution, only a handful were equipped with gauges to measure releases, he said, and many of those were inoperative or inaccurate. "That plant was being operated like a talcum powder factory in the Third World," Makhijani said

The report faults DOE for preparing its release estimates without taking into account many of the unmonitored release points such as outdoor equipment, "despite evidence that this equipment may have released enormous discharges."

A DOE spokesman said the department had no comment.

Officials recently increased their estimate of uranium releases from Fernald, but the report said the new assessment "continues to use partly fabricated data and calculational methods which are incorrect."

As an example, Makhijani said, DOE officials assumed that stack scrubbers, which are supposed to capture dust particles in a liquid solution, were operating at top efficiency. According to internal documents, many of the scrubbers were corroded or missing parts.

Documents from National Lead of Ohio also acknowledged that its method of measuring uranium losses through the scrubbers was "inherently deceptive," relying on the amount of uranium dust captured in the liquid solution. If the liquid showed no trace of uranium dust, for example, the contractor assumed that no uranium was being lost through the stack even though the absence of dust could mean that the scrubber had failed entirely.

The report also cited a legal deposition by Michael W. Boback, former head of industrial hygiene at the Fernald plant, who helped prepare the official estimates of radioactive releases. According to Boback, stack filters sometimes went unanalyzed for years, during which plant operators routinely entered "zero" in forms used to record radioactive releases.

The report said that Boback's comments suggest that official data, even for monitored pollution sources, "must be deemed to be at best seriously suspect and at worst scientifically worthless."

The report's estimate of 3.1 million pounds of released uranium was derived from soil samples taken within a five-mile radius of the Fernald plant. Makhijani said the analysis was deliberately conservative, ignoring the possibility that some uranium may have been washed away in creeks or plowed deeply into the soil.

Nonetheless, he said, the average uranium concentration in the soil suggested releases that are five to eight times greater than DOE's official estimate range.