BYU professor looks at numbers behind safety of aging nukes

In a new study for the journal Technometrics, BYU statistics professor Shane Reese and his former colleagues at Los Alamos National Laboratories calculated the certainty of AmericaÕs nuclear weapons working if theyÕre supposed to -- and not working when theyÕre not supposed to. Photo by Jaren Wilkey/BYU

PROVO -- You can taste a bite of frozen yogurt, dab on a little perfume or try on a pair of pants to make sure they fit.

Trying out nuclear weapons is a more critical and far more complex test, but one BYU professor is working on ways to make those tests both safer and more targeted.

Shane Reese, a statistics professor, is a co-author on research that analyzes past tests of nuclear weapons to look at the reliability of the United States' nuclear weapons stockpile at Los Alamos National Laboratories in New Mexico. The weapons degrade over time, so scientists have to test the weapons to determine their safety and effectiveness, meaning they won't go off when they're not supposed to and they will go off when they are. Up until a couple of decades ago, this took the form of detonating nuclear bombs. More recently, it has involved dismantling the weapons and testing each part.

As defense spending increases, it also has sparked a debate on the need to regularly test the weapons.

"Conventional wisdom says your estimate on the probability that things continue to work as they ought will remain unchanged if you don't continue to look at them, but of course conventional wisdom is often wrong," Reese said.

He said their approach "was to consider the possibility that the stockpile 'might' degrade over time, due to aging, and, using complex mathematical models and modern computing algorithms, compute the probability that the stockpile would be 'ready' if needed. We also computed the confidence we had in that number."

Nuclear weapons that were reliable last year may be OK this year, but they also might not be. His research looks at how reliable the testing information would be if it was done less frequently and found that if the government monitored the stockpile half as much as it currently does, confidence in the stockpile's reliability would decrease from 95 percent to 80 percent in seven years. If the monitoring stopped, reliability would basically be a coin toss.

Knowing that is important, Reese said, because the United States has based a significant amount of its national defense strategy on the nuclear weapons stockpile. Also, leaders need to know that there's some muscle behind those nuclear arms. The country may never use those weapons again, he said, but those weapons need to work if they're a factor in any strategic decision.

"Those are hugely impacted by the probability that you think it's going to work," he said.

This also increases nuclear safety because it gives scientists a better idea of what weapons are most likely to be less reliable. Reese is planning to do more research that looks at specific components, which should show that weapons built during specific time periods or using specific metals or processes are more likely to be dysfunctional, meaning scientists will have a better idea of which weapons need additional surveillance.