News FocusArchaeology

The End of Angkor

See allHide authors and affiliations

Science  10 Mar 2006:
Vol. 311, Issue 5766, pp. 1364-1368
DOI: 10.1126/science.311.5766.1364

The collapse of a great medieval city suggests that environmental miscalculations can spell doom for even the most highly engineered urban landscapes

SIEM REAP, CAMBODIA—Crouching in the bottom of a gully, Roland Fletcher traces with his finger the beveled edge of a pitted, grayish-red rock. The carved laterite block with a sloping face fits snugly in a groove in the block below. “It's a fancy piece of work,” says Fletcher, an archaeologist at the University of Sydney, Australia. Centuries ago, the people of Angkor built immense sandstone palaces and temples on foundations of laterite, a spongy, iron-laden soil that hardens when exposed to air. In excavations begun last year, Fletcher's team discovered that the half-meter-long block is just one piece of a dilapidated platform extending 20 meters underground in either direction. The platform appears to be the remnants of a massive spillway, possibly used to disperse floodwaters unleashed by monsoon rains. “Nobody had ever seen a structure of this kind here before,” Fletcher says.

The spillway helps resolve one debate, showing that the majestic waterworks of Angkor—a Khmer kingdom from the 9th to 15th centuries C.E. that at its height encompassed much of modern-day Cambodia, central Thailand, and southern Vietnam—were designed for practical purposes as well as religious rituals. But this singular piece of medieval engineering may also offer clues to a more profound riddle—not because the spillway exists, but because it was destroyed.

Ever since Portuguese traders in the late 16th century described the lotus-shaped towers of Angkor Wat rising from the forest canopy, people have wondered why the once-gilded temple devoted to Vishnu—humanity's largest religious monument—and the city connected with it were abandoned about 500 years ago. The list of suspects proposed so far includes marauding invaders, a religious change of heart, and geological uplift.

Beguiling.

A once-gilded Angkor Wat tower.

CREDIT: MUTSUMI STONE

Now Fletcher and his colleagues have new evidence that the very grandeur of Angkor's complex plumbing, the lifeblood of the city, left it vulnerable to collapse. In a provocative new interpretation of Angkor's demise, Fletcher, co-director of the Greater Angkor Project (GAP), a 5-year survey and excavation sponsored by the Australian Research Council, proposes that the trigger may have been a combination of rigid infrastructure, environmental degradation, and abrupt changes in monsoons. He and other scholars caution that the case is not closed. “It's hard to put a finger on any one reason for the collapse,” says Charles Higham, an anthropologist at the University of Otago in Dunedin, New Zealand, whose startling finds at earlier Thai sites are illuminating the origins of Angkor (see p. 1366).

If the GAP team is right, Angkor—the most extensive city of its kind in the preindustrial world, with a population numbering in the hundreds of thousands in its heyday—would not be the first civilization unraveled by environmental change. For example, many archaeologists now hold that a series of devastating droughts doomed the Maya and their sprawling city-states on the Yucatán Peninsula between 800 and 900 C.E. Angkor's downfall may be a cautionary tale for modern societies on the knife-edge of sustainability, such as Bangladesh. “The lesson to learn from all of this,” says Higham, “is don't abuse the environment.”

Waterworld.

(Left) Angkor's ponds, canals, and reservoirs sustained a vast city and its grand temples. (Left to right) The monuments still impress, including heavenly apsaras at Angkor Wat; a statue of a god outside Angkor Thom; and giant faces smiling placidly from the towers of the Bayon temple.

CREDIT: MUTSUMI STONE

Seeing the big picture

On a hazy January morning about 30 kilometers north of Siem Reap, the modern town near Angkor, Donald Cooney banks his ultralight plane hard right and heads toward a stand of trees a half-kilometer away. Soaring over forested lowland crisscrossed with waterways, rice paddies, and traditional wooden houses on stilts, some of the more stupendous Angkorian features are impossible to miss, including a gigantic 1000-year-old earthen reservoir called the West Baray, 2.2 kilometers wide and 8 kilometers long. Dozens of immense stone temples reflect Hindu cosmogony; the temples represent Mount Meru, the mythical home of Hindu gods, and the moats represent encircling oceans. Bas reliefs on sandstone facades depict everyday scenes—two men bent over a chessboard, for instance—as well as sublime visions such as the apsaras, alluring female dancers in elaborate headdresses who served as messengers between humans and the gods. Yet much of the kingdom remains inscrutable, like the giant faces that stare serenely from the towers of the Bayon in Angkor Thom, the walled heart of the kingdom.

Cooney, a pilot based in Knoxville, Tennessee, takes a hand off the control bar and gestures toward a thicket of banyan trees. “Do you see the temple?” he asks, his voice crackling over the headset. Even from 300 meters up, the umber towers of a walled temple complex built by Yashovarman I in the late 800s emerge from the canopy only when the aircraft is nearly on top of it. “It's easy to see how so much of Angkor was hidden from view for so long,” he says.

Cooney's flights have helped the 30-person GAP team chart new Angkorian features, such as canal earthworks that are easy to overlook on foot. “It's ground-truthing from the air,” says Fletcher, who co-directs the $700,000 project with Sydney colleagues Michael Barbetti and Daniel Penny, as well as Ros Borath—a deputy director general of the APSARA Authority, the Cambodian agency that manages Angkor—and Christophe Pottier of the French Research School of the Far East (EFEO).

Although archaeologists have long marveled over Angkor's sculptures and temples, the aerial views have been particularly revealing of the extensive waterworks that sustained them. In 1994, a radar snapshot from the Space Shuttle Endeavor espied eroded segments of the Great North Canal, which shunted water from the Puok River to two reservoirs. Then a few years later, archaeologist Elizabeth Moore of the University of London used radar to spot undiscovered Bronze Age and Iron Age settlement mounds at Angkor.

Some epic legwork has also revealed the transformation the medieval Khmers wrought on the landscape. In remote sensing and ground surveys conducted on foot in the 1990s, Pottier, an architect and archaeologist, mapped hundreds of hitherto unknown house mounds and shrines clustered around artificial ponds, called water tanks.

For Pottier, the surveys were an epiphany. “The people of Angkor changed everything about the landscape,” he says. “It's very difficult to distinguish what is natural and what is not.”

Practical minded.

Roland Fletcher, on the bank of the West Baray, argues that Angkor's reservoirs were for irrigation and flood control, not just rituals.

CREDIT: R. STONE/SCIENCE

That transformation was extensive, as Fletcher, a specialist on the growth and decline of settlements, and University of Sydney graduate student Damian Evans learned by scrutinizing NASA radar images commissioned by GAP. They found Angkorian dwellings and water tanks scattered across roughly 1000 square kilometers and connected by a skein of roads and canals, many now barely discernible.

The surveys also revealed the outlines of the ingenious water-management system, centered on three great reservoirs, or barays. A labyrinth of channels north of the barays and of the complexes of Angkor Thom and Angkor Wat diverted water from the Puok, Roluos, and Siem Reap rivers to the reservoirs (see map on p. 1367). The system “brought large amounts of water to a halt and then bled the water off into other channels as required,” explains Fletcher. Canals leading south and eastward from the barays dispersed the water across the landscape, for irrigation and to blunt seasonal flooding. This allowed the growth of a vast urban complex: a low-density patchwork of homes, temples, and rice paddies.

Angkor's growth, and the king's power, depended on sustained rice yields. “If the king runs short of rice, he'd have to go cap in hand to other Khmer lords in the kingdom,” says Fletcher. Reliable yields required ample water at the right times of year. Angkor's water system, therefore, was the wellspring of power for its rulers.

Surprisingly—and frustratingly—the roughly 1200 inscriptions in Sanskrit and Khmer chiseled on Angkorian walls are mum on the water system. “They're full of references to boundary stones and land ownership, but virtually silent on water issues and water rights,” says Higham. Apart from the inscriptions, not a single written Angkorian word has been recovered. The oldest inscribed palm leaves, a likely medium for records, date from the early 18th century.

Angkor's inscriptions also betray nothing of the kingdom's decline. By the 14th century, “we really don't know what's going on in Angkor,” says Fletcher. Siamese annals recount how an army from a nearby kingdom seized Angkor in 1431. Why the city was ultimately abandoned is an enigma—although the consequences must have been devastating. “When a low-density city collapses,” Fletcher says, “it takes out the entire region.”

Angkor's past has remained mysterious in part because Cambodia's grim recent history deterred research here. Civil war, the brutal reign of the Khmer Rouge, and finally the invasion of Vietnamese forces turned Angkor into a no-go zone for nearly 20 years. Although the heritage park with the major monuments is safe for tourists, some terrain north of the barays still has landmines and unexploded ordnance. “You can't go bushwalking here,” Fletcher says. Fortunately, the warring sides left Angkor largely untouched.

Then when Pottier reopened EFEO in 1992, the emphasis was on restoring the temples. Centuries of neglect had turned some complexes into tumbled ruins, whereas others required urgent measures to stabilize them or restore sandstone facades. More than 20 teams from around the world are working here, says APSARA archaeologist So Peang. He points to masons repairing a 12th century causeway that bisects one of Angkor Wat's moats. “Many blocks have decayed. Rains wash away the sands,” he says.

The vital repairs, not to mention efforts to interpret the structures and inscriptions, have conspired to keep attention riveted on the monuments. “Archaeologists here have tended to focus on what they can see,” Fletcher says. “Imagine trying to learn about life in New York City by only examining its churches.” As a result, says Pottier, “Huge parts of the site remain complete blanks.”

For that reason, one can stumble upon hidden treasures. Walking on the bank of the East Baray, Fletcher spots a triangular, darkgray object in the sandy grass. He picks up the palm-sized stone fragment and points to some squiggly lines. “It looks like old writing,” he says. An EFEO expert later confirms that the inscription is from the reign of Yashovarman I. “So much here is just waiting to be discovered,” Fletcher says.

Going with the flow

On the western edge of the Mebon, an artificial island in the middle of the West Baray, a tangle of grass and vines hides some laterite foundations. Seventy years ago, during the dry season, a villager looting the Mebon stumbled upon part of a gigantic bronze torso jutting from the muck near the foundations. He alerted an EFEO curator, explaining that the Buddha told him in a dream that he was buried in the Mebon and couldn't breathe. French excavations later unearthed the 2-meter-tall head and shoulders of a statue of Vishnu.

Beyond the baray, a swelling red sun is just meeting the horizon. “Imagine the rays of the setting sun glinting off Vishnu,” says Fletcher, standing beside the remnants of the temple wall. Pressure from the reservoir's water column would have forced water through the base of the statue and out of Vishnu's navel. Pollen grains preserved in mud inside the temple show that lotus plants flourished in the pond gracing either side of a causeway leading from the temple to the statue.

A controversy has simmered over whether the magnificent Mebon and West Baray were brought into being solely to inspire awe. Some advocates of the purely ritual argument long claimed that no canals exited the massive reservoir—so how could it have been used for irrigation?

In the last few years, the GAP team has amassed new details on the vestiges of outlets, some lined with laterite blocks, from both the West and East Barays. To many experts, it's now beyond dispute that the reservoirs were used to store water for irrigation during the dry season and possibly to prevent flooding of houses and fields. “I'm really impressed with what the GAP team is doing,” says Yale University anthropologist Michael Coe, an expert on the Maya and Angkor civilizations. They “have pretty much laid this debate to rest.”

But the question remains as to how Angkor's endgame played out, as scholars over the years have engaged in a lively Angkorian version of whodunit. Some experts hold that shifting economic, political, or religious winds trimmed Angkor's sails. A “building orgy” of Jayavarman VII, who in the late 12th century ordered the construction of numerous temples and Angkor Thom with its then-gilded Bayon, might have emptied the kingdom's coffers and thus reduced its influence. Another possibility is the growth of maritime trade, with centers of wealth shifting south, closer to the sea. Or perhaps Angkor's star dimmed when Theravada Buddhism and its tenet of social equality began to eclipse Hinduism in the kingdom in the 13th and 14th centuries.

A prevailing view is that crop yields declined precipitously. Nearly 30 years ago, EFEO researchers Bernard-Philippe Groslier and Jacques Dumarçay speculated that the waterways and barays filled with silt, choking off irrigation during the dry winter months. The mechanism might have been deforestation, as fields were cleared for planting to feed a booming population—making Angkor a victim of its own success.

To test these ideas, Fletcher and others set out to unravel the details of Angkor's canals and barays. They found a number of clues hinting at problems with the grand waterworks. Engineering flaws may have doomed the East Baray, at least. “By modern engineering standards, it was a total failure,” says Heng Thung of the Regional Centre for Archaeology and Fine Arts in Bangkok, a specialist in satellite data. To construct the baray, workers simply piled up earthen dikes on the sides. A canal diverted water from the Siem Reap River to the baray, but because the reservoir's bed was no deeper than that canal, the baray was perpetually shallow during the dry season. Water would have evaporated rapidly when the land is thirstiest.

The engineers got smarter when they built the West Baray a century later. This reservoir was excavated in places by as much as 1 meter, presumably deep enough to retain water throughout the dry season. What doomed the West Baray, Thung asserts, is something that Khmer engineers or astrologers could never have foreseen: geological uplift. He says the beds of the Siem Reap and other rivers unmistakably deepened during the Angkor era. This must have occurred as the upward movement of the crust gradually raised the flow gradient, making the meandering rivers run faster and cut deeper. River levels would have decreased relative to the surrounding land, until eventually the rivers were too low during the dry season to feed into the West Baray. “When the water was needed, the canals couldn't supply the reservoir,” Thung says.

Heart of the kingdom.

Canals diverted rivers in the north to the West Baray, shown here, and other giant reservoirs.

CREDITS: L. CREVELING/SCIENCE; MUTSUMI STONE

Whatever the problems with the barays, some researchers contend that their importance for irrigation has been vastly overstated. The engineered system would have provided water to grow enough rice to support between 100,000 and 200,000 people, less than half of the presumed population. The land between the barays and the Tonle Sap, a lake south of Angkor, is so flat that it's hard to imagine large amounts of water supplying a broad expanse of irrigated rice land, says archaeologist John Miksic of the National University of Singapore and an expert on early Indonesian cultures. Most of the populace tended bunded rice paddies that captured water during the monsoons or grew rice in saturated soil left after monsoon floodwaters receded.

Thung, Fletcher, and others do not dispute that baray-fed irrigation alone would not have sustained the populace. Rather, they suggest, the barays provided extra capacity that could have ensured survival rations during a poor harvest. The irrigation system “may have been a risk-management strategy for a bad monsoon year,” Fletcher says—a strategy that would have failed if the barays didn't fill. Another practical purpose of the barays might have been flood control by diverting waters from swollen rivers: “Insurance policies maybe, in case the monsoons were heavy,” Fletcher says. He proposes that the great embankments and canals are central features of an artificial wetland aimed at reducing the risk of flood damage.

Remarkably, the GAP team is the first to attempt a comprehensive assessment of Angkor's climate and waterworks. Penny, an expert on past climates, is analyzing pollen grains to uncover shifts in vegetation cover. From this he hopes to reconstruct changes in land use during Angkor's waning centuries. So far, he says, results indicate that Angkor's demise was “patchy, both over time and space,” undercutting traditional explanations for the city's last days. Other scientists are focused on the canals and barays—and tackling the perplexing riddle of the spillway.

Dowsing for the truth

The spillway might never have come to light if Khmer Rouge laborers hadn't dug an irrigation trench through it. Around the end of the 9th century C.E., with Angkor blossoming, engineers excavated a long canal that altered the course of the Siem Reap River, diverting it southward to the East Baray. The spillway extends westward from the northern end of the canal and would have functioned, Fletcher infers, to protect that canal from excess flooding, comparable to a bathtub's overflow slot.

The conundrum is that the spillway's meticulous construction is badly damaged. The Khmer Rouge trench has largely eroded away; the Angkorian workmanship was far superior. Nevertheless, most of the spillway's laterite blocks lie in a jumble under sandy soil. “It was torn apart,” says Fletcher. It's possible, he says, that an engineering flaw caused the spillway to give way. But Angkorian structures were built to last, which propels Fletcher toward a different conclusion. “It seems they ripped the spillway out themselves,” he says. “Perhaps something had gone wrong.”

Fletcher tends to see the hand of necessity, not neglect. He hypothesizes that the structure was damaged by flooding, then dismantled for its materials, after which Angkor workers buried the ruins to prevent the diversion canal from breaking out of its channel and flowing west.

The story of the spillway and other clues from the GAP team's work indicate that over the centuries, Angkor's vaunted water system grew ever more complex to support the sprawling city. “We know that something was going wrong, mechanically, with their water system,” Fletcher says. For instance, one of the great southern canals was filled with cross-bedded sand, indicating considerable and rapid water flow as well as sedimentation. “The sand buried the canals,” he says. The water infrastructure “became so inflexible, convoluted, and huge that it could neither be replaced nor avoided, and had become both too elaborate and too piecemeal.”

Mapping in style.

Donald Cooney's bird's-eye views of Angkor and its temples (top) have helped researchers detect subtle features, such as eroded canal earthworks.

CREDIT: R. STONE/SCIENCE

The more complicated and delicately balanced the system grew, the harder it would have been to compensate for unusual events, such as extreme flooding or drought. “They engineered a completely artificial environment, and it was a fragile environment,” Pottier says.

Add climate change to this volatile mix, and you have a recipe for disaster. The GAP team is probing whether monsoons became “really erratic” during the Little Ice Age, between 1300 and 1600 C.E. There's good evidence that cooling in the Northern Hemisphere not only weakened monsoons in mainland Southeast Asia during this period but also triggered sharp declines in crop yields in Europe. William Boyd, a geologist at Southern Cross University in Lismore, Australia, who's collaborating with Higham, finds GAP's scenario reasonable; he suspects that the Angkor region would have become drier.

The idea is “really intriguing,” says archaeologist Miriam Stark of the University of Hawaii, Manoa, who has studied the pre-Angkor Funan kingdom of Cambodia's Mekong delta. But archaeologists say more climate data are needed. “A definitive answer,” says Miksic, “would require a study of the ecology of the Tonle Sap Basin over the past 2000 years.”

Fletcher concurs, and he and Penny hope to receive permission from Cambodian authorities to go caving in search of evidence. Stalagmites add new layers every year; the isotopic chemistry of the layers captures a record of climatic conditions during accrual. (Stark and Paul Bishop of the University of Glasgow hope to gather similar data from caves in the Mekong delta region in the coming year.) Deciphering the isotope record in suitably old stalagmites should yield insights into past climate. Other evidence could be gleaned from changes in vegetation or lake water levels, and Chinese trade records, says Boyd.

As Angkor was rising, halfway around the world a similar loss of equilibrium brought the Maya to their knees. Overpopulation and environmental degradation had weakened their Mesoamerican cities. Beset by droughts and a “paroxysm of warfare,” the civilization crumbled, Coe says. “The demise of Angkor is directly comparable to the great Maya collapse,” he says. And it could happen again. “When populations in tropical countries exceed the carrying capacity of the land, real trouble begins,” Coe says.

Fletcher says there are lessons for developed nations as well. He compares Angkor's plumbing woes to modern cities having to cope at great cost with extensive, decaying sewers or cumbersome road systems.

On their quest to understand the end of Angkor, the GAP team will continue to excavate and analyze deposits from its water features and reconstruct its environmental history. Angkor, says Thung, “is a never-ending story.” Or, rather, it's an ending often rewritten, and with a lesson for society that gets gloomier each time.

View Abstract

Stay Connected to Science

Navigate This Article