Exhumed bodies of the 1995 Srebrenica massacre, in which 8,000 Bosnian Muslims were killed and thrown into mass graves. Image: Wikimedia
By now, you have probably seen the images of hundreds of broken Syrian bodies lying in mass graves, victims of the alleged August 21 chemical weapons attack outside of Damascus. The images, from unconfirmed amateur videos, are horrific, but they are far from the first reports of mass graves in Syria. It’s impossible to know how many of these graves exist, and in all likelihood, we will probably never know just how many victims have been hastily abandoned in unmarked pits since the country’s civil war began two years ago.
It’s a problem that has plagued nearly all of history’s greatest tragedies. Mass grave sites are difficult to confirm, and often take years to find, if they are found at all.
In Iraq, families of Saddam Hussein’s victims rush to grave sites every time bodies are uncovered. In Cambodia, mass grave mapping teams have identified 20,002 sites in the Khmer Rouge’s “Killing Fields,” and it’s unclear how many more exist. And in Bosnia, where mass graves were routinely moved every time investigators got too close, it’s impossible to know just how many bodies were buried.
Now, a team of scientists from Oak Ridge National Laboratory and the University of Tennessee’s renowned forensic anthropology center are developing a suite of technologies to better detect mass graves. If successful, the project will make mass-grave mapping both safer and easier, allowing clandestine grave-hunters to identify possible burial sites from the safety of their own desks by using multispectral imagery from satellites and drones.
“The perpetrators of these crimes have historically done a really good job of covering up these crimes—generally these graves are very difficult to find, even if you have the right technology,” said Devin White, a senior research scientist for ORNL’s geographic IT division. “We're trying to find a more efficient and a more accurate way to do this, and in particular, in areas where we might not be able to get in there on the ground.”
The initial data for the project is being gathered by a team of forensic anthropologists at a mass grave experiment site in eastern Tennessee. Unlike most mass grave experiments, the University of Tennessee team is using actual dead bodies that have been donated to the forensics center’s “Body Farm,” a step that researchers hope will improve the accuracy of their data as they determine which geographical, environmental, and anthropological characteristics of a mass grave can be seen from space. While satellite images have been used since the 1990s to find large mass grave sites, the goal of this new project is to locate smaller sites—where six, 10, 20 bodies are buried—and test whether the technology can be that fine-grained.
"It's these types of graves that fall through the cracks, that are harder to find,” White said. "For the families that have missing people around the world, those people are still missing regardless of whether they are in a big grave or a small grave."
Once the forensic data is gathered, scientists at ORNL use it to inform a mechanical algorithm aimed at identifying other possible mass grave sites. The idea, White explained, is to use commercial satellite images to determine the probability that a mass grave exists at a given location, based on factors like vegetation growth, soil quality, and other environmental disturbances. Ideally, White said, researchers could combine these satellite pictures with lower-altitude aerial imagery, including thermal data and Light Detection and Ranging (LiDAR) technology, which measures subtle changes in ground elevation.
Researchers are also experimenting with other geospatial data, including information from crowdsourcing and social media, to more accurately pinpoint the site of mass graves.
“We just pull in anything we can get our hands on,” White said. “No one has ever really tried to do this before, so we're basically throwing the kitchen sink at it, looking at every possible sensor, every possible data set that we can collect, and trying to weed it down to the ones that we think are useful.”
White cautions that the project is still in its early stages, and the technology will probably never be able to identify grave sites with absolute certainty: The only way to prove that a mass grave exists is to actually dig up the bodies.
“This is not a predictive model,” White said. “It's a guide. We're by no means saying that we're going to come up with the Holy Grail here, and come up with one answer that is going to work everywhere. “
But the new technologies aim to focus grave-hunting methods, which currently rely on blind luck and unreliable eyewitness accounts. Moreover, in areas where investigators have been barred from entry, or where conflict makes it too dangerous to start digging up bodies, the aerial images could allow teams to track grave sites from afar, and give government intelligence agencies and international groups a bird's-eye view into opaque foreign conflicts.
“If you can come up with a way to look at these areas safely from a distance, then you can make a judgement call that can inform policy, what non-profits are doing, and enable the international community pressure on these groups that are killing people and putting people in the ground,” White said. “In the end, that's going to help so many people.”