If Country X tested a nuclear weapon by exploding it deep below the Earth’s surface, would we know? Would we care?
“We would absolutely care,” says Neill Symons, a seismologist with the Laboratory’s Seismoacoustics Team. “Our job is to do research and development to advance the U.S. capability to monitor the globe for underground nuclear tests. We do that in direct support of the Department of Energy’s nuclear nonproliferation mission.”
The “would we know?” question is a bit more complicated to answer. That’s because, at first glance, an underground nuclear test and an earthquake are similar in one very important way: they both release energy in the form of seismic waves.
During an underground nuclear explosion, seismic waves are created as rock is pushed away from the explosion’s source and compressed. These compression-caused seismic waves are mostly primary waves (P-waves), which shimmy forward and back like a slinky, along the direction the wave is moving. As they travel through the earth, some P-waves get converted to shear waves (S-waves), which shake up and down perpendicular to the direction the wave is moving (imagine a kid on a pogo stick bouncing down the street).
"By looking at thousands of waves, I can get the best guess of what an area looks like geologically." —Scott Phillips, seismologist
During an earthquake, on the other hand, seismic waves are created as two slabs of rock suddenly slide past one another. This motion creates both P-waves and S-waves.
In comparing an earthquake and an explosion with similar P-wave energy, the explosion will generally have less S-wave energy. The technical challenge, however, is doing this comparison correctly, which can be tricky if the geology around and depth of the explosion are not well known.
If they’re big enough, waves from seismic events are recorded at seismic stations around the world. With that recorded seismic data in hand, scientists on the Seismoacoustics Team have developed computer codes—complex algorithms that take into consideration things like source, seconds of time, kilometers of distance, and rock properties. They use those codes to simulate the recorded waves. These simulations can reveal whether a seismic event is an earthquake or an explosion. If it’s an explosion, the simulation also sheds light on the explosion’s depth and yield (explosive power). Yield is important because an explosion surpassing a certain yield is likely to be nuclear.
One challenge, however, is that waves travel differently through different types of rock. For example, waves travel quickly through a hard rock such as granite and tend to move more slowly (but with greater motion and higher energy loss) through soft rock such as tuff.
Laboratory seismologists know a lot about explosions in soft rock because the underground tests the Lab conducted at the Nevada Test Site occurred mostly in soft rock. Lots of seismic data— relating to depth, containment, and yield—were collected from these tests and have been used to develop the codes used for current monitoring techniques.
Scientists also have a pretty good understanding of both the geology and seismic activity in places like the Soviet Union’s old Semipalatinsk Test Site and China’s Lop Nur Nuclear Test Base. That’s because much of the data from the hundreds of nuclear tests conducted in these locations is now public.
But what about the geology of places that don’t have a lot of seismic activity, natural or otherwise? Much of the rock below Earth’s surface remains a mystery, especially in terms of how seismic waves might propagate through it.
Take North Korea, for example. “By looking at topography and old geology maps, we know the Korean test site is in granite,” says Laboratory seismologist Scott Phillips. “But beyond that, what’s the structure of the granite? Are there layers? Rock is rarely uniform for miles in all directions.”
To answer these questions, Phillips looks at every seismic wave that travels through a specific area. The process is rather time-intensive, considering that everything from traffic and construction to landslides creates seismic waves. “Then we feed all of our wave info into a program that’s similar to a CAT scan,” he says. “Hopefully, we can get a picture of all the ‘bones and guts’—where waves do or don’t travel well and how fast they are. By looking at thousands of waves, I can get the best guess of what an area looks like geophysically.”
Traditionally, the Seismoacoustics Team has been concerned with big seismic events that create seismic waves traveling thousands of miles around the globe. However, as underground nuclear tests, such as the recent rogue tests by North Korea, have smaller yields than many of the well-studied tests of the 20th century, they produce fewer long-distance seismic waves and are harder to distinguish from earthquakes.
For example, on October 9, 2006, North Korea experienced a magnitude 4.3 seismic event, which is not particularly big except for the fact that seismic events of this size aren’t common on the Korean peninsula. The North Korean government claimed the event was caused by a nuclear test— but was it really?
Using their traditional methods, the Seismoacoustics Team could use codes to simulate waves to characterize this alleged test. “But it would be kind of like looking at a car from 100 yards away,” explains Seismoacoustics Team leader Mike Cleveland. “We could see it’s a white car, but that’s it.” The team wanted the 10-yard view. “We want to be able to tell that it is a 2006 white Subaru Outback with tinted windows and a roof rack,” Cleveland continues. “In other words, we realized our understanding of an underground nuclear explosion would be much better if we could simulate not just the seismic waves but also the nuclear explosion and other geologic effects of the explosion.”
Fortuitously, Los Alamos is one of two national labs that have access to America’s nuclear weapons codes, which are the highly specialized and validated codes that are used to simulate the performance of the nuclear weapons in the U.S. stockpile.
And the Laboratory employs a plethora of scientists who develop infrasound, electromagnetic, and radionuclide codes that simulate low-frequency sound waves, electromagnetic energy, and gas transport, respectively.
In the end, all these different scientists were willing to come together in an effort to, for the first time ever, link their respective codes to more accurately simulate an underground nuclear explosion—from the detonation to the resulting geologic effects. (It helped that the NNSA’s Office of Defense Nuclear Nonproliferation Research and Development agreed to fund the collaboration.) By being able to simulate, for example, a specific hydrogen bomb explosion in a specific type of granite, scientists will be better able to verify such an event should it ever happen in real life.
Linking codes is harder than it sounds
“We already have the many individual codes that have been written for different pieces of the puzzle,” explains weapons designer Amy Bauer. “The problem is that these codes have never spoken to each other. They haven’t even been introduced.”
"These codes have never spoken to each other. They haven’t even been introduced.”
—Amy Bauer, Weapons Designer
It’s almost like playing a game of telephone with a bunch of people who all speak different languages. “We want them to speak each other’s language so the output from one code can be the input for the next code,” Bauer continues. “It becomes a computer science problem—writing scripts so the codes can talk to each other.”
But it’s also more complicated than that because different codes are concerned with different time and length scales. “Nuclear weapons codes are concerned with shakes, each happening in 10 billionths of a second, while other codes are concerned with things happening over minutes or even months,” Symons explains. “It’s hard to make the handoff from code to code because they’re working on different time scales.”
For example, a nuclear test detonates in the blink of an eye and causes the rock around it to fracture. The size of the fracture depends on the type of explosion and the type of rock. Very large fractures will reach Earth’s surface, which allows gas from the explosion to also reach the surface, but over the course of weeks or months.
“One of the biggest challenges in computational physics is the coupling of physics across disparate time and length scales,” Bauer says. For the collaboration to work, many factors—the type of explosion, the type of rock, the size of the fracture, the amount of gas—along with their time frames must be accounted for in the codes. And the codes must be compatible with one another to ensure and the resulting geologic effects.
If that sounds complicated, it is. But Bauer stresses that the work is also rewarding. “The physics is fun—there are so many different scientific problems,” she says. “And I get to work with worldclass, engaged scientists.”
Up until 1992, if scientists wanted to better understand a certain type of nuclear explosion and the resulting seismic behavior, they would simply attach sensors to a test device and blow it up. Testing has since been replaced with the Stockpile Stewardship Program—nonnuclear experiments that provide data for the codes and the resulting computer simulations that give scientists the confidence they need to ensure the safety, security, and effectiveness of the nuclear weapons in the U.S. stockpile.
“We rely on the work in the Stockpile Stewardship Program to give us confidence that the weapons codes are extremely accurate,” Symons says.
To gain confidence in the rest of the codes, particularly those that involve geology, the Lab is conducting several experiments.
The Source Physics Experiments (SPE) series is currently underway at the Nevada National Security Site (NNSS). Ten different-sized tests will be detonated at different depths in different rock types. These tests, which are chemical (nonnuclear), can be used to validate codes in places with known geology.
“The data collected from the SPE trials strengthen our national security by advancing technical solutions for treaty monitoring and improving computer simulation methods used to evaluate potential explosions anywhere in the world,” says Brent Park, the NNSA’s deputy administrator for the Office of Defense Nuclear Nonproliferation.
In 2022, Physics Experiment 1 (PE1) will take place at NNSS, generating data on how rock changes as energy is released during an explosion. The $70-million experiment comprises three chemical explosions underground in a tunnel and will validate weapons and geologic code calculations.
In the meantime, scientists can create simulations of known (historical) tests to see how their simulations compare with real data. If the simulations and the data match, the new monitoring technique is working.
The end product
Let’s circle back to a seismic event occurring in Country X. “We want to be ready,” Phillips says. “We want to be able to tell if it is a nuclear explosion.”
Being ready involves having a solid idea of what that explosion would look like, despite possibly never having seen any seismic activity in that location before.
“It’s a totally new way of doing monitoring,” Symons says. “It lets us go beyond just looking at places we know about to imagining what the signals are going to look like in places we can think about; it lets us overcome data limitations.”
Bauer agrees. “This is a great example of a program that only succeeds by doing top science to solve big problems,” she says. “Having passionate scientists available to work on something like this is why Los Alamos is one of our nation’s gems—nobody does it better than we do.” ★
"It's hard to make the handoff from code to code because they're working on different time scales."
—Neill Symons, seismologist