Internal documents discarded at an abandoned workers' dormitory in Tonopah (Secrets of the Palo Verde Inn," June 1) indicate that Arizona Public Service Company concluded in 1985 that a dozen metal shelters used to house monitors could be skewing radiation detection downward.
Nine years later, Palo Verde managers continue to use the metal shelters rather than replace them--as APS environmental managers recommended in 1985--with wooden shelters. The wooden shelters, APS discovered, allowed radiation monitors to collect too many dust particles.
"Replacement of the metal shelters with wood shelters is the best option for assuring representative atmospheric sampling," Bob Estes, then director of APS' environmental department, concluded in a March 11, 1985, memo.
The faulty shelters raise serious questions about the validity of APS' monitoring program, says Robert Pollard, senior nuclear safety engineer for the Union of Concerned Scientists.
"If your samplers are not having adequate air flow through them, of course you're measuring a radiation level lower than that actually being released," Pollard says.
The off-site monitors dot the plant's perimeter and are posted in six outlying communities. Each week for the last 500 weeks, a crew from APS, which manages Palo Verde, has pulled air filters from the monitors and tested them for airborne radioactive elements.
With the exception of the few months following the 1986 Chernobyl nuclear disaster in the Soviet Union, the results have been consistent. They show low levels of airborne particles, reflecting nothing more than background radiation.
This isn't surprising, Palo Verde officials say, because the nation's largest nuclear power plant doesn't allow airborne releases that exceed Nuclear Regulatory Commission limits. Records submitted to the NRC by Palo Verde officials show the plant vents a steady stream of airborne radioactive materials in minute quantities. The lengthy list includes such obscure elements as xenon, vitrium and thorium.
The releases, Palo Verde operators insist, are well within the NRC standards designed to protect humans from harmful radiation.
Palo Verde spokesman Mark Fallon says APS is convinced that its monitoring program provides accurate readings. He says APS monitors detect nearly identical amounts of airborne radioactivity as those independently operated by the Arizona Radiation Regulatory Agency, which also has its monitors in metal housings.
Fallon says Estes' 1985 conclusion prompted APS to conduct tests that validated use of metal shelters. But Fallon declined to provide copies of those studies and indicated that APS has never formally overridden Estes' assessment.
"There is no singular document that will contradict what you have got," he says of Estes' report.
Fallon says APS found that the wood shelters allowed relatively large airborne dust particles to be detected by the monitoring equipment. He says the utility is more interested in detecting radiation in smaller particles, which are more likely to travel greater distances. Fallon says the metal shelters prevented the larger dust particles from entering, thus giving APS a more representative sample of airborne radiation likely to drift far distances.
That logic doesn't sit well with Alfred Schmidt, who built the radiation monitors and who began urging APS as early as 1980 to use wooden shelters.
"I tried to tell them 12 or 14 years ago they should use shelters that collected everything possible," says Schmidt, who was reached at his office in San Carlos, California.
Schmidt says APS should collect large and small dust particles for radioactive testing. He says his own experience has shown that the larger particles can be detected miles from their source. Furthermore, he says, a decades-long debate has yet to conclude what size of dust particles should be considered inhalable and ingestible by humans.
"I happen to be one of those who strongly believes in the necessity of being able to sample particles as large as 100 micrometers [about the thickness of a piece of paper]," says Schmidt.
While Fallon's assertion that wooden shelters collect larger dust particles is true, APS' own data also show that wood and metal shelters perform virtually the same in detecting particles less than ten micrometers in diameter. For particles larger than ten micrometers, the wood shelters were far better, a 1985 APS study shows.
"The wood shelter collected twice as large a sample total as the metal shelter," says Schmidt. "The metal shelter would show emissions were safer than they actually were."
Schmidt has spent 50 years working with radiation-related projects, including four years on the Manhattan Project, which produced the first nuclear weapon during World War II. He's also worked for Oak Ridge National Laboratory and the U.S. Navy to develop radiation monitors.
Schmidt will not suggest that anyone is in danger because of Palo Verde's air monitors, but he is concerned that the company is not using accurate testing equipment. Schmidt says he first urged APS to use wooden shelters because they wouldn't heat up as much as the metal variety, which, when overheated, can destroy monitoring equipment.
APS followed Schmidt's suggestion and conducted tests in 1981 to determine what type of shelter would stay cooler. As a side result of the test, APS also discovered the monitors housed in the wooden shelters collected far more dust. The wooden shelters were identical to ones used by the U.S. Weather Service to house its instruments.
"Just as I predicted," Schmidt says, "the insides of the metal doghouses ran considerably hotter than the insides of the weather bureau shelter."
But the unexpected radiation readings surprised and concerned Schmidt.
"The air samplers in the weather bureau shelters collected twice as much airborne dust even though the test was on the upper story of a four-story building in a place where you would expect to only find small dust particles," says Schmidt.
On April 28, 1982, Schmidt wrote to the NRC, informing the agency of APS' findings: that monitors in wooden shelters collected an average of 43 percent more dust.
This fact, Schmidt wrote, "can have a significant effect on measurement of airborne radioactivity, and may result in reported values that are considerably lower than actually exist [emphasis his]."
Schmidt urged the NRC to look into the matter, especially since most power plants were using metal shelters.
"These findings indicate an urgent need for additional tests," he wrote.
The NRC ignored his recommendations then, and still does.
"We have never found the metal housing to be a big enough issue to warrant NRC taking action or requiring the licensee [plant operators] to take action," says Susan Gagney, an NRC spokeswoman.