More than one million people visit the famous temples around Siem Reap each year, but it took a remote sensing laser survey to discover traces of a vast urban network surrounding the Angkor and Koh Ker temple complexes and a previously unknown ancient city on nearby Phnom Kulen.
Using a laser scanner strapped to a helicopter, the researchers were able penetrate the vegetation that had long blocked a view of the ground.
The results of that April 2012 aerial survey and subsequent on-the-ground fieldwork, which are to be published this week in the Proceedings of the National Academy of Sciences of the USA, may well change history as we know it.
“Our jaws dropped,” said University of Sydney’s Damian Evans, who heads the partnership of eight organisations that conducted the research, including the Cambodian government’s APSARA Authority.
“It reveals quite clearly that the formalised, urban center of the city of Angkor extends over at least 35 sq km, rather than simply the 9 sq km conventionally recognised within the walls of Angkor Thom,” the paper notes.
The survey – the first of its kind in Asia – showed that mounds and depressions that appear pattern-less from the ground actually form the remnants of highly structured city grids, roads, dams and canals.
According to the paper, the data demonstrated that “the intensity of land-use and the extent of urban and agricultural space have both been dramatically underestimated in the Angkor region until now”.
These findings suggest that in the 12th century the area contained a “very large population” sustained by regular agricultural imports from the countryside.
Dependence on surplus agriculture and large water management systems, in turn, show how droughts contributed to the civilisation’s eventual collapse, researchers say.
Pointing to a mound outside of the crumbling walls of the Beng Mealea temple on a recent expedition with the Post, Evans explained that “this was once the foundation for a block of wooden structures”.
Those structures have long since disappeared into the jungle, leaving only slight mounds and dips that are easy to overlook amid the trees and underbrush, and are overshadowed by the mossy ruins of the temple. The mounds each span a few metres in diameter and look – to the untrained eye – like nothing more than natural undulations of the earth.
Archaeologists, however, have long known they meant more.
A different type of archaeology
Siem Reap is a naturally flat flood plain, so even such small variegations are evidence of human settlement, Evans said.
Still, with no way to see through the tree canopy, researchers were previously able to learn about the shape and extent of this settlement only by painstakingly hacking through the vegetation by hand.
The team’s choice of laser imaging detection and ranging technology, or “lidar”, is far more efficient – though also far more expensive.
Aerial lidar surveys fire millions of laser beams at the ground and measure the time they take to bounce back, using tiny differences in time to calculate variations in elevation.
The Angkor project, which is the most extensive archaeological lidar survey ever conducted, used a particularly high concentration of beams to ensure that some made their way through the trees to the earth below and then back to the machine.
After a computer screened out results reflected from vegetation, the data revealed – up to a resolution of a few centimetres – images of “every small dyke, every ancient road”, said Jean-Baptiste Chevance, program manager of the Archaeological and Development Foundation, another group involved in the project.
“We clicked a button, and – boom!” Evans said.
Dense urban networks
The results, according to archaeologist Michael Coe, are “absolutely mind-boggling”.
“This is the greatest advance in our knowledge of Angkor as a living city in the past century,” said Coe, who some 60 years ago, according to the team’s paper, first suggested using laser to map ancient forest civilisations.
Before now, Coe said, most scholars believed Angkor was urbanised in a “dispersed-urban pattern”, with populations loosely scattered around hubs like canals, roads and family-maintained ponds.
“What the lidar images reveal is something that we had not anticipated: a densely occupied city with streets and avenues laid out on a grid pattern aligned with the cardinal directions.”
Around Angkor, “the lidar imagery shows the area is very precisely organised into city blocks of a very specific size,” Evans said. “Each of the city blocks has four elevated mounds and pond, and there would have been wooden structures on each of those mounds.
“It would have been tightly packed and teeming with life – a very busy place.”
Cambodia historian David Chandler termed the discoveries “thrilling developments”.
“The lidar has shown streets and canals that obviously required central planning and coordination,” he said.
Additionally, Evans said, the lidar data showed that the networks of roads and settlements continued outside temple complexes’ moats and walls, which were previously thought to form the outer bounds of settlement.
“There are no gates where the roads cross,” Evans said. “So what’s happened is that there was some kind of pre-existing urban network in this area, and this temple and its enclosure have been stuck into the middle of it all.”
According to Coe, the survey’s results at Angkor mean scholars are going to have to revise their population estimates for the city at its height.
Based on radar and ground surveys, scholars in recent years had offered “a ‘guesstimate’ of 750,000 inhabitants”, Coe said.
“Back in 1979,” he said, “the late Bernard-Philippe Groslier, based on his contention that Angkor was a ‘hydraulic city’ that used the great barays, or reservoirs, to irrigate the rice fields, proposed a figure of 1.9 million people. It now looks like Groslier was right.”
Evans was more cautious about the survey’s implications for population figures, stressing that the topic required much more study.
Researchers hope they can match the lidar data with fieldwork on the ground and written accounts from the time to suss out population figures. But the frequent rebuilding and relocation of Angkorian centres over the span of 1,000 years, with many sites built on top of others, makes determining a population at any given time more difficult, Evans said.
In the end, the team hopes they will gain insight into a key question: What happened to this civilisation?
Sceptical of explanations that attribute the decline of the great Khmer Empire to invasion by the Thais, the team suspects a more gradual source of decline – the civilisation’s unsustainable dependence on large-scale water-management systems.
Among the lidar data is evidence of great dams and canals that demonstrate the importance of water management not just for Angkor but for all Khmer cities at the time, said Evans.
The lidar shows, for example, that a five-kilometre-long elevated strip near Koh Ker, previously believed to have been a road, does not rise and fall with the land around it but remains constant in elevation, suggesting that it actually was a dam.
These findings give support to the theory that Angkorian residents “muddled through for a few centuries” with massive water-management systems but that the empire eventually collapsed due to major droughts that are visible in environmental evidence such as tree rings, Evans said.
“The new data lends further weight to an emerging consensus that the development of the vast engineered landscape of Angkor over several centuries was fundamentally unsustainable,” the researchers write.
“Increasingly sophisticated technologies of water management may have afforded a measure of resilience on an annual scale by ensuring food and water security for an ever-larger and increasingly urbanised population; paradoxically, however, those same systems would also have created a systemic vulnerability to longer-term climatic variation.”
A broader span
The team’s 20-hour lidar survey covered 370 square kilometres at a cost of about a quarter of a million dollars.
Generally archaeologists, finding such costs prohibitive, have used only whatever relevant data they could glean from lidar studies conducted for non-archaeological purposes – usually for less-detailed national topographical surveys.
The Angkor survey was only the second ever carried out over a large area specifically for archaeological research.
In 2009, the first such survey succeeded in penetrating vegetation to map the ancient Mayan city of Caracol in Belize. Plans are under way for an expanded follow-up study.
Finding adequate resources for the Angkor project required the cooperation and funding of eight different organisations, including the Cambodian government’s APSARA Authority and the French École Française d’Extrême Orient.
Archaeologists are now in the process of exploring on the ground the features detected from the air, downloading the lidar data into handheld GPS units and heading out into the field with maps of the aerial survey as guides.
Evans demonstrated the pro-cess to the Post at Beng Mealea last month, showing how the seemingly random mounds around the temple corresponded, on the lidar map, with a regular topographical grid.
Picking up a rust-coloured piece of ceramic lying on one mound, he identified it as part of an ancient roof from one of the wooden structures that once stood there.
Scattered bricks and ceramics tend to be the main traces of these structured settlements that remain for researchers on the ground, Evans said.
There have been some larger discoveries, however. A team led by Stéphane De Greef from the Archaeology and Development Foundation at the sacred mountain of Phnom Kulen to Angkor’s northwest has confirmed – through on-the-ground fieldwork – 28 new temples detected by the lidar and found dozens of carvings in a riverbed.
World Heritage potential
These discoveries have given some heritage experts hope the survey will help expand the international recognition and protection given the Angkor UNESCO World Heritage Site to include Phnom Kulen, which was long mine-riddled, relatively difficult to access, and lacking temples as iconic as Angkor Wat.
“It’s really been off of people’s radar,” said Evans.
Now, the survey’s findings of new temples, roads and a wholly new city there have buttressed heritage groups’ claims that the Kulen has “cultural and natural heritage of outstanding universal value” – the basis for world heritage site designation.
“This newly discovered urban landscape,” the study says, “corresponds to the 8th-9th [century] city named Mahendraparvata, one of the first capitals of the Khmer Empire, which was previously known only from written inscriptions but was commonly assumed to be located in the Phnom Kulen region.”
Chandler called the evidence of the city “mind -bending, suggesting that the very first years of Angkor – early to mid-9th century, perhaps earlier – so poorly documented in inscriptions, might already have involved relatively dense populations and a level of fairly sophisticated urban planning.”
Anne LeMaistre, country director for UNESCO in Cambodia said: “Phnom Kulen is really the symbolic source of Angkor.”
“At Phnom Kulen, King Jayavarman in 802 declared himself king of kings, unified all the little kingdoms and integrated them into the Angkorian Empire. So it makes sense to include the Kulen in the Angkor perimeter.”
But when Angkor initially was inscribed as a World Heritage Site in 1992, the Khmer Rouge still occupied Phnom Kulen, so including it in the World Heritage Site was not an option, LeMaistre said.
Since then, incorporating Phnom Kulen “has been in APSARA’s mind and in UNESCO’s mind for a long time,” she said. Now the lidar study’s remarkable results have spurred the organisations to start the application process, which is “just at the beginning”.
UNESCO received support for the scheme in December from the government’s APSARA Authority and from Deputy Prime Minister Sok An, chairman of the National Commission from UNESCO.
Going forward, UNESCO and the government will need to work to restore and renovate the site and write the nomination, which then would be evaluated by the World Heritage Committee. A final decision would be about three years away, LeMaistre said.
Evans noted that Phnom Kulen was one of a handful of Cambodian sites under consideration for World Heritage Site status.
It is not yet decided whether the proposed extension would coincide with the current Phnom Kulen National Park or be smaller or larger – a matter up to APSARA.
Bun Narith, director-general of APSARA, said his organisation was “in the process of preparing” for the project.
“The site must have an exceptional value,” he said, joining LeMaistre in noting that reforestation was necessary to improve the site.
Chevance, too, said he hoped the extension of the World Heritage site would encourage both archaeological and environmental protection.
But he noted that despite international support, the Cambodian government, as sovereign manager of the current Heritage Site around Angkor, still struggled to provide adequate protection and monitoring of the site against loggers and careless tourists.
Still, he said, “certification and registration of Phnom Kulen as a World Heritage Site should bring a better level of protection. Because you can’t do whatever you want in a World Heritage Site.”
Not just at Phnom Kulen but across the former Khmer Empire, the lidar results offer important information for further research and management of the sites, Im Sokrithy, one of the researchers and a spokesman for APSARA, said.
“APSARA will use the data for, firstly, managing the land use in the region. It will be helpful to make an accurate land use map,” he said.
The data would also help the government prioritise which areas to preserve, and to “develop public and tourist infrastructures in the park without impact to the historic sites,” he said.
Balancing these interests could be a delicate matter, said Evans.
“You can’t just say that you can’t build anywhere that there is an ancient feature or something like that, because pretty much everywhere in the greater Angkor area, including in Siem Reap town, was a part of the ancient urban landscape,” he said.
The prevalence of ancient features, however, “highlights the need for us to do as much survey work as quickly as possible in order to more effectively plan things like development and more effectively manage the site.”
This is one of the reasons that Evans’s team hopes in 2014 to conduct a second lidar survey, which would map a larger swath of sites not covered last April. The project would include more of Phnom Kulen and the ancient “industrial city” of Preah Khan (not to be confused with the Angkor temple of the same name).
“In the Phnom Kulen area, all we’ve done is just cover a small subsection. The same is really true in Angkor,” Evans said. “But we now know where the sweet spots are.”
In some areas, traces of urban networks extend all the way to the end of the lidar maps, and researchers are eager to learn how far beyond they continue.
Evans is currently in the process of securing funds for the second survey. At roughly $500,000, it would cost twice as much as the first one, but Evans said the groundwork for the first study and its success has helped clear the way.
Meanwhile, some features the lidar picked up are still mysteries, including “a series of rectilinear coil-shaped embankments of indeterminate function” just south of Angkor Wat’s moat.
“They don’t fit into any conventional idea of what the Angkorian landscape looks like,” said Evans. “They’re not in any Angkorian art, and there are no parallels to things in other parts of the world. They don’t really make sense in terms of agriculture or water management.”
But further research would eventually reveal their purpose, he said.
“There always turns out to be a rational explanation.”