This spring, when the teams submitted their results to IARPA, evaluator teams graded how well each one did. In June, the teams learned who was moving on to SMART’s second phase, which will run for 18 months: AFS, BlackSky, Kitware, Systems & Technology Research, and Intelligent Automation, which is now part of the defense company Blue Halo.
This time, the teams will have to make their algorithms applicable across different use cases. After all, Cooper points out, “It is too slow and expensive to design new AI solutions from scratch for every activity that we may want to search for.” Can an algorithm built to find construction now find crop growth? That’s a big switch because it swaps slow-moving, human-made changes for natural, cyclical, environmental ones, he says. And in the third phase, which will begin around early 2024, the remaining competitors will try to make their work into what Cooper calls “a robust capability”—something that could detect and monitor both natural and human-made changes.
None of these phrases are strict “elimination” rounds—and there won’t necessarily be a single winner. As with similar DARPA programs, IARPA’s goal is to transition promising technology over to intelligence agencies that can use it in the real world. “IARPA makes phase decisions based on performance against our metrics, diversity of approaches, available funds, and the analysis of our independent test and evaluation,” says Cooper. “At the end of phase 3, there could be no teams or more than one team remaining—the best solution could even combine parts from multiple teams. Alternatively, there could be no teams that make it to phase 3.”
IARPA’s investments also often leak beyond the programs themselves, sometimes steering scientific and technological paths, since science goes where the money goes. “Whatever problem IARPA chooses to do is going to get a lot of attention from the research community,” says Hoogs. The SMART teams are allowed to go on to use the algorithms for civil and civilian purposes, and the datasets IARPA creates for its programs (like those labeled troves of satellite imagery) often become publicly available for other researchers to use.
Satellite technologies are often referred to as “dual-use” because they have military and civilian applications. In Hoogs’ mind, lessons from the software Kitware develops for SMART will be applicable to environmental science. His company already does environmental science work for organizations like the National Oceanic and Atmospheric Administration; his team has helped its Marine Fisheries Service detect seals and sea lions in satellite imagery, among other projects. He imagines applying Kitware’s SMART software to something that’s already a primary use of Landsat imagery: flagging deforestation. “How much of the rainforest in Brazil has been converted into man-made areas, cultivated areas?” Hoogs asks.
Auto-interpretation of landscape change has obvious implications for studying climate change, says Bosch Ruiz—seeing, for example, where ice is melting, coral is dying, vegetation is shifting, and land is desertifying. Spotting new construction can show where humans are impinging on areas of the natural landscape, forest is turning into farmland, or farmland is giving way to houses.
Those environmental applications, and their spinout into the scientific world, are among the reasons SMART sought the United States Geological Survey as a test and evaluation partner. But IARPA’s cohort is also interested in the findings for their own sake. “Some environmental issues are of great significance to the intelligence community, particularly with regard to climate change,” says Cooper. It’s one area where the second application of a dual-use technology is, pretty much, just the same as the first.
NASA engineers hope to have their massive moon-bound Space Launch System ready for liftoff in a couple of months, but so far they’ve encountered some bumps in the road. On March 17, NASA rolled the world’s most powerful rocket out onto the launchpad at Kennedy Space Center in Florida to ready it for the Artemis program’s inaugural lunar mission later this year. Since then, technicians have completed a raft of checks on the huge rocket’s systems, but after three tries they haven’t been able to make it through the final test, a practice countdown called the “wet dress rehearsal test.”
The key problems have been a faulty helium check valve and a liquid hydrogen leak, which led to several pushbacks of the test countdown. Finally, NASA officials decided over the weekend to disconnect the rocket and, starting next Tuesday, carefully roll the SLS and Orion crew capsule back to the Vehicle Assembly Building, a facility with the equipment needed for them to perform rocket surgery. They hope to have a quick turnaround, returning to the pad soon afterward to complete the countdown test, but the first Artemis mission around the moon—originally planned for early June—might be delayed.
“The mega moon rocket is still doing very well. The one check valve is literally the only real issue we’ve seen so far. We’re very proud of the rocket,” said Tom Whitmeyer, a deputy associate administrator at NASA headquarters in Washington, at a press conference this afternoon. “But we have a little bit more work in front of us.”
The precautions aren’t surprising; NASA doesn’t want to take a chance on its most expensive rocket or debut Artemis launch failing. “It comes down to what we consider to be the acceptable level of risk,” said Mike Sarafin, the Artemis mission manager, at an earlier press conference on April 15.
The test itself began on April 1, after the rocket had been ferried from the assembly building to Launch Complex 39B via an enormous crawler. Jeff Spaulding, the senior NASA test director, and his team began their process by hooking up the rocket’s electrical power and pressurization systems and filling the pair of white boosters on the side with propellants. Then they started loading the big orange fuel tank with more than 700,000 gallons of liquid hydrogen and liquid oxygen, supercooled to a frigid -423 and -297 degrees Fahrenheit, respectively. (That’s the “wet” in “wet dress rehearsal test.”) Their goal was to simulate the entire countdown process to just under T-10 seconds—the closest thing to a real launch without firing up the core stage’s RS-25 engines.
Throughout the test, Spaulding and his colleagues monitored instruments, pressures, temperatures, and valves to check that all the systems were working within acceptable parameters. (“If it turns out they’re a little outside of the limits, that’s what we want to know now—if there’s something we need to fix or adjust,” he had said in the days leading up to the rehearsal.)
The test revealed the need for several adjustments. The process was delayed the first time on April 2 by lightning bolts, which hit the towers around the rocket. Then the following day, NASA officials encountered problems with launch tower fans and their backups, according to Charlie Blackwell-Thompson, the Artemis launch director. These fans provide pressure in the mobile launcher, the tall structure next to the rocket, to keep out hazardous gasses. That led to a delay while the fan malfunction was resolved.
When hydrogen atoms first formed, they absorbed and then emitted ambient 21-centimeter radiation at equal rates, which made the clouds of hydrogen that filled the primordial universe effectively invisible.
Then came cosmic dawn. Ultraviolet radiation from the first stars excited atomic transitions that enabled hydrogen atoms to absorb more 21-centimeter waves than they emitted. Viewed from Earth, this excess absorption would appear as a drop in brightness at a specific radio wavelength marking the moment the stars turned on.
In time, the first stars collapsed into black holes. The hot gases swirling around these black holes generated x-rays that heated hydrogen clouds throughout the universe, increasing the rate of 21-centimeter emissions. We would observe this as an uptick in brightness at a slightly shorter radio wavelength than that of the older light. The net result would be a dip in brightness over a narrow radio wavelength range, like the one detected by EDGES.
But the observed dip, which occurred around a wavelength of 4 meters, was not what theoretical cosmologists had expected: The timing and shape of the trough were off, indicating that the first stars turned on surprisingly early, and that x-rays flooded the universe soon afterward. Stranger still, the dip was very pronounced, suggesting that hydrogen in the early universe was colder than theoretical models predicted, possibly because of exotic interactions with the dark matter that fills the cosmos.
Or perhaps the EDGES dip had a more mundane origin.
Hydrogen’s 21-centimeter emissions from the cosmic dawn era reach Earth with wavelengths of several meters, in the range used for FM radio and television broadcasts; that’s why EDGES operated in such a remote location. What’s more, the signal is overwhelmed by radio emissions thousands of times brighter from our own galaxy, and it’s distorted by its passage through the upper layers of the Earth’s atmosphere.
No less important are subtle effects from the antenna itself. A radio antenna’s environment can slightly change the area of the night sky to which it is sensitive. In such a precise experiment, even faint reflections off surfaces tens of meters away can matter. The effect of such reflections would be enhanced at certain radio wavelengths, resulting in a small variation in the antenna’s observing area—and thus potentially in the measured brightness—at different wavelengths.
The EDGES team did see this kind of ripple in their data, and the prime culprits, perhaps fittingly, were the edges of a 30-meter-wide metal screen placed on the ground surrounding the antenna to block radio emissions from the ground itself. The team corrected for possible reflections off of these edges in their analysis, but as some astronomers noted at the time, if the correction was even slightly off, the result could be a dip in background brightness over a narrow wavelength range indistinguishable from a real cosmic dawn signal.
The SARAS team took a different approach to antenna design in pursuit of more uniform sensitivity across all wavelengths. “The entire design philosophy is to preserve that spectral smoothness,” said Saurabh Singh, the lead author on the SARAS paper. The antenna—an aluminum cone propped on a Styrofoam raft—was floated in the middle of a placid lake to ensure there would be no reflections for more than 100 meters in any horizontal direction, which Parsons called “a really cool and innovative approach.” Moreover, the slow speed of light in water reduced the effect of reflections from the lake bottom, and the uniform density of the water made the environment much easier to model.
Lia Siegelman had just been studying the swirling waters of the Southern Ocean, which surrounds Antarctica, when she happened to come across a poster image of cyclones around Jupiter’s north pole, taken by NASA’s Juno spacecraft. “I looked at it, and I was just struck: ‘Whoa, this looks just like turbulence in the ocean,’” she says.
So Siegelman, a researcher at UC San Diego’s Scripps Institution of Oceanography, turned her eye to the latest detailed images of the outer planet. She and her team proved for the first time that a kind of convection seen on Earth explains the physical forces and energy sources that create cyclones on Jupiter. (Since air and water are both “fluids,” from a physics perspective, the same principles apply to the gas giant’s atmosphere and our oceans.) They published their findings today in the journal Nature Physics.
Jupiter, the 4-octillion-pound elephant in our solar system, makes gigantic cyclones, big storms that rotate around areas of low pressure. Some are thousands of miles wide—as large as the continental United States—with gusts of wind up to 250 miles per hour. Eight of the largest have been spotted at the planet’s north pole and five at the southern one. Scientists have speculated for years about their origins, but by mapping out these storms and measuring their wind speeds and temperatures, Siegelman and her colleagues showed how they actually form. Little spinning vortices pop up here and there among the turbulent clouds—not so different from the ocean eddies Siegelman’s familiar with—and then they start merging with each other. The cyclones grow by continually gobbling up smaller clouds and gaining energy from them, so that they keep on spinning, she says.
It’s a clever way to study extreme weather on a planet that’s more than 500 million miles away. “The authors are clearly drawing from meteorology and oceanography disciplines. These folks are taking this rich literature and applying it in sophisticated ways to a planet we can barely touch,” says Morgan O’Neill, a Stanford atmospheric scientist who models the physics of hurricanes and tornadoes on Earth and has applied her work to Saturn.
In particular, O’Neill says, the team of scientists demonstrate how, like thunderstorms on Earth, Jupiter’s cyclones build up through a process with a gross-sounding name: “moist convection.” Warm, less-dense air, deep down in the planet’s atmosphere, gradually rises, while cooler and denser air, near the frigid vacuum of space, drifts down. This creates turbulence, which can be seen in Jupiter’s swirling, moisture-packed clouds of ammonia.
As the Perseverance rover drilled into a rock on Wednesday to collect a sample from Jezero Crater on Mars, Justin Simon, a planetary scientist at NASA’s Johnson Space Center in Houston, felt both nervous and excited. He has the honor of serving as the “sample shepherd,” leading the effort from millions of miles away, but the pressure’s on. “These samples not only will allow us to understand the geology of the crater but also minerals likely related to the history of water there,” he said yesterday.
But first, the rover had to actually capture a chunk of rock in a test-tube-sized container. An initial attempt in early August had come up empty. That first rock, nicknamed “Roubion,” simply crumbled to dust as the drill bored into it, and none of those bits made it into the container.
Simon can now breathe a sigh of relief. Perseverance’s second try, with a different rock, appears to have successfully extracted a Martian core slightly thicker than a pencil.
“We got that image of just a spectacular-looking core, a fantastic-looking cylinder, broken off cleanly. It looks geologically very interesting, something scientists of the future will enjoy working on,” says Ken Farley, a Caltech geochemist and project scientist of the Perseverance mission, which is led by NASA’s Jet Propulsion Laboratory in Pasadena, California.
But the analysis of the new sample is going to take awhile, because NASA scientists haven’t been able to get clear photographs due to low lighting conditions, which makes the images tough to interpret. To add more drama for the scientists, when Perseverance did a “percuss-to-ingest” procedure—shaking the sample to make sure the tube wasn’t overfilled, which would make the system jam when it’s stored—one image appeared to show an empty sample tube. (They’re pretty sure they got the sample, but they’re going to try taking more images in better light over the next couple of days.)
Perseverance’s first drill attempt, which essentially pulverized the sample, wasn’t a complete failure, as it yielded evidence suggesting the rock had been weathered, worn down by a river flowing into the lake crater billions of years ago. “It always had been possible that this lake was a transient event, like maybe a comet, rich in water, hit Mars and made lakes, and then it boiled away or froze within tens of years. But that would not produce the weathering we see,” said Farley in an interview earlier this week.
Since that rock was too powdery, the scientists then piloted the rover to a new area, looking for a different kind of rock to sample, using the Ingenuity copter to scout ahead. Perseverance trundled slightly to the west, where on a ridgeline the researchers found a larger, boulder-like rock, which they nicknamed “Rochette” and which seemed less likely to disintegrate when the rover deployed its tools on it. “It looks like a rock that, if you could throw it, would clank down on the ground. A good, healthy rock,” Farley said.
Before each sampling attempt, Perseverance performs reconnaissance by snapping a bunch of photos of a candidate rock. Last weekend, it also performed an abrasion test to see if Rochette was durable enough to sample. The rover is equipped with a rotary percussive drill (with extra drill bits) that both spins and hammers into the rock. This tool helps clear away dust and chip through the weathered outer layer. The abrasion was spectacularly successful, according to Farley, so the scientists decided to go ahead with grabbing a sample. Perseverance extended its 7-foot-long robotic arm, fired up the drill, and carefully extracted a core sample. Then it rotated the arm’s “hand” so that the sample tube could be inspected.
