The U.S. stopped testing a quarter-century ago, but it continues to refurbish aging devices, and needs accurate predictions of how those weapons would perform. At the Lawrence Livermore National Laboratory (LLNL) in Livermore, Calif., nuclear weapons physicist Greg Spriggs has been tasked for more than a decade with writing and updating computer code to predict fallout.
He’s long had difficulties in using test films created more than half a century ago under extreme circumstances on sometimes unstable stock, then stored under varying conditions. The answer, Spriggs realized more than a decade ago, was to move from analyzing copies of tests to re-examining at the original tests themselves.
Trouble was, his models wouldn’t agree with the data — so as a “sanity check,” he convinced his supervisor to let him scan about a dozen of the estimated 10,000 such films that exist to check cloud heights and dimensions.
When data and model still didn’t align, something not entirely unexpected occurred: His boss gave him the job of reanalyzing all the films to obtain usable data.
“What we’re hoping to get out of this project is to generate a bunch of benchmark data, particularly concerning the yield of a nuclear device,” Spriggs told Government Technology. “Then, we can run our codes and try to recalculate some of these old weapons; and if we’re able to actually reproduce the yield that we measure, then we can go back to the [National Nuclear Security Administration] and say, ‘Look, we’re doing a good job. Our codes are validating what we did and we think that our new predictions are going to be correct.'"
Joined by film technician and expert Jim Moye and supported by a staff that includes programmer Jason Bender, Spriggs began the painstaking process of scanning and scrutinizing scientific-grade films done by the company EG&G, now part of Aecom but still renowned for its contributions to high-speed photography during the past eight decades.
The group has only located about 6,500 films, has scanned more than 4,200 of them, and analyzed about 500.
But thanks to hours of laborious work along with some recent technological developments by Bender, Livermore hopes to go live around the end of the year with a website documenting its work, and to continue releasing scanned images from declassified films. Examples can currently be found in the LLNL Atmospheric Nuclear Tests playlist on its YouTube channel.
A final report on the project, Spriggs told Government Technology, likely won’t be complete for as long as two years.
Nolan O’Brien, LLNL’s public information officer and communication manager, said the project has attracted a lot of positive attention even from people “that we traditionally might expect to be Internet trolls or anti-nuclear deterrent in general.”
“They’re praising the transparency that’s kind of involved in this project. In an essence, what Greg and his team are doing is they’re finding ways to apply new techniques and technologies to get better answers without having to go back to testing,” O’Brien told Government Technology.
One of the inherent issues that prompted the work Spriggs, Moye and others are doing may not always be considered in scientific circles: shrinkage.
EG&G, incorporated in 1947 by Massachusetts Institute of Technology professor Harold Edgerton and grad students Kenneth Germeshausen and Herbert Grier, did fine work on the tests, those interviewed agreed.
It used at least 50 cameras in every shoot; and while Livermore hasn’t yet been able to locate more than 30 films of any one test, EG&G’s comprehensive data sheets — which documented camera locations, lens focal lengths, filters, emulsion types and other particulars — proved crucial to reanalyzing the films.
But during the past 50 years, the film stock used has changed, shrinking by anywhere from 1.5 percent to 2.5 percent, Spriggs said. So long as the films shrink equally throughout in a proportional sense, they can be reanalyzed accurately.
Changes must be reflected and measured, for a weapon’s yield is calculated by first measuring the shockwave position then the yield — and so if a shockwave is incorrectly calculated by 1 percent, the yield will be off by 5 percent.
Despite the changes to the film stock, Moye said in an LLNL video on the lab’s YouTube channel that he is “very surprised” by their condition, which is generally good.
Scanning the films for digital capture enabled their measurement by computer, and Bender created a tool using Python software that automates the reading of the images’ optical density to more quickly find a fireball’s edge.
The tool, which the Spriggs said brings the team “great hope,” speeds analysis from about four hours per film to three to four minutes. It’s currently testing and could be out of beta mode anytime between the end of summer and the end of the calendar year.
Concurrently, Spriggs and other personnel are watching anywhere from 50 to 200 test films per month, and subsequently declassifying them.
It’s a combined process that he and Moye said could have significant value for state, county and local governments despite the fact that the threat of nuclear war has diminished since the height of the Cold War, when these films were produced.
“I think one thing that [Hurricane] Katrina taught us was that you need to be prepared and just put something down on paper. Sometimes it doesn’t always work the way you think it’s going to work. A lot of emergency response personnel try to prepare for a nuclear detonation. There’s a usefulness to this, that’s basically a training aid for our civilian defense personnel,” Spriggs said, noting staffers hope, naturally, that such an attack never comes.
Moye said the process has been “real interesting and rewarding” from a scientific and preservation standpoint.
“It’s exciting when I hear that lot of good data is coming from these films," he said, "and then at some point in the future, now that it’s being scanned and preserved digitally, scientists in the future can look at this and who knows what else they can glean from it."
