Archaeology from Space

by Sarah Parcak

  Not all sites, however, can be turned into tourist attractions. It is costly and time-consuming, and there are not enough tourists to visit the countless places left to be discovered. Only the most die-hard tourists tend to visit small sites or those off the beaten path.

  Nonetheless, solutions can be found—new economic and educational opportunities for the people near archaeological sites can be created. These people can be given new skills to make local handicrafts based on ancient styles to sell through cooperatives in big cities or online, or in local bazaars. Archaeologists working on sites can engage the local communities. Many of my colleagues give site tours to local schools and community members during their dig seasons.

  Sometimes our work can have unintended positive consequences. At Tell Tebilla, a young villager named Abira worked with us as a basket-woman, removing excavation debris. Her interest was so impressive, and her teenage English put our Arabic to shame. The day her high school exam results came in, her uncle brought a crate of pop for the entire team: she’d passed at the top of her class. We were all proud of her. Abira ended up going to Cairo University to study Egyptology, and later told me seeing women working as Egyptologists on our dig inspired her.

  As for site protection, what if a worldwide database of sites existed, fully trackable at all times, with global hot spots appearing when looting, development, or climate change posed a threat? Think of a 21st-century version of the Monuments Men, the brave men and women who helped to save priceless treasures from the Nazis in Europe during World War II. Instead, we could have Monuments People, a force of millions, of all ages, across the globe, mapping high-resolution imagery, finding sites, identifying looting, and sharing that data with governments and archaeologists. Just imagine all it could achieve.

  Now, how can we possibly mobilize that force?

  12

  Space Archaeology for Everyone

  Given the devastating extent of site destruction, the odds seem stacked against us. Although the dire reality is clear, many archaeologists have become more optimistic because of the rapid advances in technology and science. Even with those advances, however, the harsh truth is that site destruction outpaces our capacity to protect traces of the past.

  We need to work faster and smarter, and we need more people. We need an archaeological revolution to upend traditional approaches and broaden participation. Not enough of us can work as archaeologists to begin to make a dent in the backlog. So many people have wanted to be archaeologists since the age of five, and of all times they should be able to realize that dream now, when we have such a massive workload ahead of us. Turning to the public for help might be one of the most intriguing new possibilities for archaeology.

  We also need to ask who the past is for, and who gets to add to our shared human story via new archaeological findings. The greatest age of discovery in archaeology may be ahead, not this moment, even though so many are now using new remote sensing and other technologies. Nonetheless, it is coming, and it will arrive when everyone can contribute. You probably don’t believe me, after reading so many stories of scientists’ discoveries. You think it’s impossible. Now those odds, I like.

  The Power of Many

  You have probably heard of crowdsourcing. Everyone does it and may not realize it. When you ask on Twitter or Facebook for advice about plumbers or restaurants or the best diapers, you engage the wisdom of the crowd. Some may think the crowd can’t be relied upon to help with anything scientific, but the answer is a resounding yes, it can.

  The first major crowdsourcing effort, called Galaxy Zoo,1 showed the world just how much the general public could help scientists. Based at Oxford University, the project began when scientists realized they had a million photographs of galaxies from the Sloan Digital Sky Survey but no way to classify all of them. As an experiment, they created an online classification platform that gave participants three choices: spirals, mergers, or ellipticals. The creators thought it would take the crowd years to get through the data, but 150,000 people made 50 million classifications in the project’s first year.

  With numerous people confirming the same images, the accuracy of the crowd equaled that of scientists. The experiment has now evolved into a website called Zooniverse, which hosts dozens of crowdsourcing projects, from identifying bird plumage to transcribing World War I journals. I had a try at that, finding this description: “Fine night, perfectly calm, troops crowded.” Resisting the urge to Google, I hoped that it remained fine and calm for the British 9th Battalion, King’s Royal Rifle Corps, fighting on the Continent 100 years ago.

  I can hear you saying, well, picking shapes and colors, reading messy writing … Welcome to kindergarten! It’s hardly brain surgery.

  Aha! Meet Eyewire,2 which is under the direction of super-brainy scientist Amy Robinson. Eyewire allows people worldwide to help map the neurons of the brain in 3-D through an online game. Fun, elegantly designed, and easy to use, the game includes competitions among the hundreds of thousands of regular users. Eyewire and Galaxy Zoo opened my eyes to the potential of getting netizens to pitch in on projects, and I’m not the only archaeologist seeing that. Crowdsourcing has taken off big-time in archaeology over recent years.

  The Levantine Ceramics Project3 crowdsources among scholars working across the Mediterranean, to share data on ceramic wares made over a 7,000-year span, from the Neolithic era through the Ottoman period. In early 2018, the project had some 250 professional contributors, who have uploaded information about more than 6,000 ceramics. That data can be sorted by shape, period, site name, country, and region, making it easy for specialists to find parallels or matches while in the field, and facilitating the interpretation of excavations as they happen. We can now imagine a future when every archaeological object type appears in a similar database, allowing for interpretation of finds within hours of unearthing objects. Heady stuff.

  Crowdsourcing that relies on the general public has the great advantage of bringing lots of fresh eyes to a project. My own work with satellite imagery is time-consuming, can be expensive, and takes enormous focus. Before I start a project, I have a pretty good idea of the types and range of things I will find on my screen, and I sometimes dismiss a lot, especially in places I know well, such as Egypt. But as we saw in earlier chapters, assumptions can hurt us in research when we’re dealing with the overly familiar.

  An Egyptology colleague told me about taking his friend’s eight-year-old daughter to the Valley of the Kings, where a complex scene of the underworld in the tomb of Ramses VI had perplexed him for months. The little girl pointed out a tiny detail he had missed, and the whole scene finally made sense to him. I did not appreciate this story until Greg and I brought our then-four-year-old son to Egypt in the summer of 2017. To explain things, we got down on his level, and for the first time looked up. We saw things we had never seen before, in our 50 combined years of work in Egypt.

  But if the untrained can help archaeologists with their research, it makes you ask why we need to learn to interpret and analyze imagery. There are chunks of academia that require years of intensive training, and gaining that expertise is essential for more nuanced analyses. But here’s why I want to include others in the fun: as a girl growing up in Maine, 5,000 miles away from pyramids and camels and everything exotic that I ever dreamed of, I could not have imagined the career I have now. That girl is still there, reminding me that everyone deserves a chance to explore and find amazing things.

  For everyone lacking access to the education, resources, or mobility to get into the field, we need a way to bring archaeology out of the field. There’s a very simple reason why: our race against time is perilous.

  Someone with major expertise could search a 100-kilometer-square area in a day, if that area did not require significant imagery processing. But with the entire Earth to explore—some 197 million square miles, minus the oceans—there are still 60 million square miles of surface left to analyze. That would take an imagery-processing specialist 4,5
66 years to finish. If I had been born in Egypt’s Old Kingdom, when Khufu was polishing his electrum-plated pyramidion to crown his tomb at Giza, I’d just be finishing today.

  The Story of GlobalXplorer

  My own plunge into the world of crowdsourcing began in 2015 when I was unexpectedly nominated for the annual $1 million TED Prize. It’s not something for which you can apply, and it’s not like the Oscars—it’s about what you’re going to do for the world, not what you’ve done. I had to invent a wish, one big, inspiring idea that would lead to global change. Easy if you say it quick. Oh, and you’re only allowed a 50-word pitch.

  I had a lot of help and did a lot of soul-searching while preparing the presentation. My good friend Raghava KK, a brilliant artist from India, told me that if the process didn’t nearly destroy me, then I wasn’t doing it right. And it nearly did. I got depressed and hopeless—what I wanted to do went against all my ingrained academic notions of archaeological ownership. I almost gave up. But Greg convinced me to stick with it, and so I plowed on … and resolved to give away everything I had worked on all my professional life. So here it is:

  “I wish for us to discover the millions of unknown archaeological sites across the globe. By building an online citizen-science platform and training a 21st-century army of global explorers, we’ll find and protect the world’s hidden heritage, which contains clues to humankind’s collective resilience and creativity.”

  We could turn archaeology upside down, empowering everyday people from all walks of life to participate in the process of exploration. It did not matter that I had only a small chance of winning, it was worth it, just for the chance for a world where all the kids who dream of seeing far-off places and making discoveries could do both.

  When Chris Anderson, the head of TED, called to let me know I had won, I was utterly unable to respond. A heavy mantle of responsibility settled on my shoulders. I would be championing a far larger vision than I could ever have imagined. I finally got to share this crazy idea at the yearly TED Conference in February 2016 in Vancouver, Canada.

  At the beginning of the process, in the fall of 2015, I knew nothing about creating an online platform design and even less about online gaming, aside from playing solitaire. If you had asked me what UX and UI were, I would have told you to go to the hospital to get your rash checked out. (They stand for User Experience and User Interface, in case you didn’t know, either.)

  Fortunately, I could draw on the expertise of the TED community and crowdsourcing veterans like Amy Robinson, who had gained invaluable insights during the creation of the Eyewire website and could not have been more generous with her time and advice. I slowly built up a team while exploring the Wild West of crowdsourcing. It was—and still is—a great age of experimentation.

  We wanted to create an online experience that would bring people back again and again, and engage a broad audience, reward them for their efforts, and, most important, make something that works. We had no idea what we were doing, and we would not know until we launched.

  We named the platform GlobalXplorer (GX). From the design stage, GX needed to be sophisticated but simple and easy to use, yet engaging enough to draw in computer-savvy aficionados. It took us six months. We did deep dives into user archetypes—that is, the kinds of people most likely to use GX.

  In an ideal world, we would create something that everyone would want to use. But nothing can be everyone’s thing. After much debate and narrowing down, we picked four archetypes we thought would bring in the broadest audience possible: a master’s degree archaeology student who wants to make a difference in the field; a tech-savvy early thirtysomething who loves gadgets and exploration but does not have a lot of time; a retired professional who loves traveling and has a little more time, but perhaps needs help with technology; and a disabled stay-at-home grandparent who is outright scared of technology, but has hours to dedicate to a new interest.

  Our team worked with Mondo Robot, a wonderful platform-design team based in Colorado. We had our dream laundry list of the features we wanted to include, and then we had to pare it back. One point of no compromise was giving our participants an immediate sense of community on GX, as if they were on an archaeological excavation. Another important focus was creating a game that rewarded participants for their efforts with interesting content about the country where we would start.

  Discovery image from the globalxplorer.org platform [IMAGE COURTESY GLOBALXPLORER]

  Looting on a satellite image of a site in Peru [IMAGE COURTESY GLOBALXPLORER AND DIGITALGLOBE]

  After a year of hard work, my team and I launched GX, an online, citizen-science, crowdsourcing, satellite-imagery platform that allows anyone in the world, whether 5 or 105 years old, to help in the process of locating and protecting ancient ruins. We focused on Peru for several reasons: first, Peru is world famous because of Machu Picchu; the archaeological sites are not difficult to spot in satellite imagery, since they mainly comprise stone or mud brick; and the Peruvian government has a strong tradition of supporting innovative archaeological work. A drone-based mapping program is already in place at the Ministry of Culture, streets ahead of other countries.

  We launched with a fantastic on-the-ground partner in the Sustainable Preservation Initiative (SPI), a group whose mission includes empowering local women living near ancient sites to become economically independent through selling their handicrafts.

  Ultimately, our goal with GlobalXplorer was to empower the world to become archaeologists and see the world as we do. We wanted to give users the tools they need to imagine the past and the people who lived in it, and to make them stakeholders in how history gets written.

  Playing the Game

  At the outset, users view a short tutorial for the first level: looting detection. That includes examples of what looting looks like in Peru, both close up and farther away, and in different landscapes. On completion of the tutorial, people begin their expedition. Each image is a snippet of a larger satellite image measuring 300 square meters, offered at random like a card dealt from a deck.

  When we process satellite imagery, we typically zoom in no more than that, so this is an authentic experience. People have two categories to choose from with a click: “looting” or “no looting,” depending on how they assessed the image. And then onward, to another tile.

  And no, we do not include any GPS points or map information on the expedition page, before you get concerned that these images potentially give a leg up to looters. Unrecognizable and untraceable, the image could come from anywhere in Peru except the impenetrable rainforest.

  After viewing 1,000 images, participants progressed to the next level, identifying illegal construction on archaeological sites. This seemed to confuse people, so we eventually took it down. I’ve learned that making any platform responsive to user feedback is so important.

  The last level, site discovery, proved to be the most difficult level for our citizen scientists. Archaeological sites, as it turns out, are hard to spot from space … and now I see what the PhD was for! We show many examples in the discovery-phase tutorial, and people learn that it just takes time and practice to discern what might be an ancient feature: another authentic experience.

  To make it all fun, we gamified everything into 10 distinct levels, each with a badge bearing the image of an ancient Peruvian artifact. Users level up as they view more images. They begin as a Pathfinder, advance to Wayfinder, and continue all the way up to level 10, Space Archaeologist. Quelle surprise!

  Every week during the opening campaign, which officially lasted three months, users unlocked new content about Peru’s archaeology and history that was provided by the National Geographic Society. The more levels people completed, the more rewards they unlocked, like Google Hangouts, Facebook Live sessions, and personal messages from our team. The platform is still live on the GlobalXplorer website, and all the National Geographic content is still available—please do have a look, my bet is that you’ll enjoy it. As user
s get better at finding Peru’s sites, they learn more, see more known sites, and find even more sites.

  On the back end, we set up data delivery to show us pins where users had found features. A minimum of six users had to agree that a feature appeared in an image before the tile came to us to assess. Also, to give users a sense of their performance, everyone received a “consensus score” showing how much they agreed with other users. Everyone started off at only 50 percent, by the way, even me. Fair is fair.

  We could not guess what would happen when we opened the platform to the public, on January 30, 2017. It might crash. Users might expect far more than we could do on our budget; a million dollars might sound like a lot, but with complex technical work, it shrinks quickly. The response exceeded our wildest imaginations. Within a week, users had examined more than a million tiles, and we started getting emails from our participant community. The idea that anyone could contribute to finding archaeological features from anywhere, as long as they had a computer or mobile phone, blew people’s minds.

  As I write this, over a year after the platform’s launch, we’ve had over 80,000 users from more than a hundred countries, including Afghanistan, Yemen, and American Samoa. No Greenland yet, but if you read this and live in Greenland, please give it a try. Our users have examined over 100,000 square kilometers of satellite data—that’s over 15 million tiles to date.