Petra, the famed ancient city hidden deep in a craggy canyon in southern Jordan, is one of the world’s most renowned archaeological sites, filled with temple ruins and tombs carved into stone. Its grandest treasure, a mausoleum known as the Treasury, is an iconic artifact as well as an Instagram hot-spot, accessed by streams of tourists through a mile-long canyon walkway known as the Siq.
But if you walk past the Treasury, along the water conduit that helped the city prosper even through drought, and around the base of the canyon’s walls, you enter what might be called the “suburbs” of the original Petra, where locals still live in homes etched from rock by their ancestors. There, on one of many tall hills seemingly covered with nothing but scattered rocks, sits what could prove to be the newest major discovery in Petra: the ruins of an ancient temple.
Carved stone rectangles outline the foundation, remnants of walls line the perimeter, and shards of 2,000-year-old pottery litter the grounds. Hiding literally in plain sight, the spot is surrounded by scattered active dwellings, yet has been unknown to contemporary archaeologists.
That is, at least, until Sarah Parcak spotted it with satellites.
The renowned University of Alabama archaeologist, winner of the TED Prize in 2016, had been involved with a BBC-funded project earlier this decade to re-investigate areas of Roman occupation to see if anything had been missed. Satellite data led her and her colleagues to this hilltop site.
“As we were investigating, this interesting rectilinear feature showed up,” she recalled as she investigated the sight first-hand back in September. “Today you can see it pretty easily in Google, but at the time the best resolution satellite I could get was 2.5 meters. Now we can get 0.3 meters, which is an order of magnitude difference in the work that I do.”
This sort of ever-increasing sensing capability—from simple optical mapping to infrared and spectral imaging—has grown into one of modern archeology’s greatest assets, allowing scholars and field researchers to better detect undetected sites. Furthermore, this capability, according to Parcak, arrives at a time when it’s perhaps needed most, with population migration, farming, urbanization, and looting—and even political conflict—threatening untapped treasure troves of significant historic remains.
“In 2011, the Arab Spring created a big uptick in looting, and we’ve been tracking a lot of the looting through high resolution satellite imagery,” says Parcak, who’s primarily an Egyptologist, where much of the plundering has been located. “There are all these big questions—who’s doing the looting, how is it leaving Egypt, how is it ending up in Western markets, and how much is it worth in the black market?” After the Arab Spring, Parcak’s teams found over 200,000 “looting pits,” or holes dug in known sites to retrieve artifacts.
Archaeologists using satellites are able to conduct their research with available data, and can request time on satellites if they have specific requests. Drones equipped with radar and laser scanners can help flesh out a three-dimensional map of the location. In addition to using a wide variety of sensors, they compare locations from year to year and even season to season, such as when water saturation might change the color of material on the ground, helping differentiate between natural and man-made formations. Any remote sensing project could use between six and 10 different kinds of satellite images, which is critical to maintaining a scientific approach, Parcak says.
In Jordan this past September, Parcak was able to add a critical field approach to her understanding of the site. Though the ruins were barely perceptible in the initial satellite data, they were distinctive enough to prompt additional investigation, including her colleagues using drones to generate more detailed maps. Parcak’s visit, along with archaeologist Qais Tweissi from Petra’s Archaeological Park, was supported by the carmaker Infiniti as one of its recent technology and exploration outreach efforts, which included a drone-guided quest for dinosaur fossils in Mongolia’s Gobi desert.
After a miles-long, circuitous hike to reach the hilltop, Parcak’s eyes glowed with recognition, not just of the site as she observed from satellite images, but of the insights the ruins themselves generated immediately.
“We don’t know exactly what it is,” she says, “and we can’t call it a temple just yet, so we say it’s a ‘ritual feature’. There are several levels preserved from the foundation, column bases and bits of the columns themselves. This is not a house—this is much, much bigger, and it's the same size and shape as other ritual and or temple type features in the area. You also don’t typically find multiple enclosures and multiple walls in residences. Plus, it has a really amazing view of the central city. You can see all around it. So typically, especially Petra, they only put their most important sites on high points.”
The next step is to convince the Jordanian researchers to fund a full excavation of the site, Parcak says. Though it doesn’t have the impact of other, more intact sites in the area, such as the Treasury, there’s still much that can be learned from the remnants present.
“There’s this mentality of finding the golden goodies that is still very exciting to a lot of people,” the archaeologist explains. “But I like to say that in archaeology, we don’t look to find, we look to find out. That, I think, is the power of these applications of new technology.”
Fortunately, the site’s location amid well-known ruins will give it a degree of protection that other sites being discovered via satellite might not have. It’s that urgency that pushes archaeologists to learn as much as they can about what’s still out there. And there’s a lot out there: Parcak estimates that they’ve found 1/1000th of 1 percent of the archaeological remains in the Egyptian Delta alone, and likely a similar number globally.
The process of honing the methodologies includes bringing artificial intelligence into the process of analyzing satellite images. A recent experiment in which they fed systems with known looting sites—training it to distinguish between, say, a bush and a hole in the ground, which can look similar in overhead images—generated a promising 90 percent success rate in terms of the computer finding the sites itself.
But even that’s not enough to ensure that sites are identified and catalogued on a global level. For that, Parcak has turned to crowdsourcing the search. Her team created a not-for-profit, GlobalXplorer, in 2017 to enable data to be analyzed by private citizens. They’re able to look for potential sites—being guided in what clues might be present—but the sites specific locations are masked, in order to protect them from continued looting.
The system has had 90,000 volunteers from 120 countries scan 14 million satellite images, and so far they’ve mapped done things like map 20 percent of Peru and find 800 previously undiscovered tombs in Eqypt. There are gaps, of course, and they’re working to ensure that every area is actively searched.
“We’re missing Greenland, Chad, and North Korea,” Parcak says. “So I always tell people that if they know Dennis Rodman, ask him to give Kim Jong Il an iPad.”