Angkor, in what is now Cambodia, was one of the most populated cities in the world from the 9th to 15th centuries CE. That said, archaeologists have struggled with knowing just how many people actually lived there. A comprehensive new study may have finally answered the question, affirming the ancient city as a bustling metropolis.
The Greater Angkor Region was home to between 700,000 and 900,000 people at its peak in the 13th century CE, according to research published today in Science Advances. That makes Angkor, the capital city of the Khmer Empire, one of the largest pre-modern settlements in human history.
“Determining the past populations of archaeological sites is a fundamental task for archaeologists,” Sarah Klassen, an archaeologist at the University of Leiden and a co-author of the new study, explained in an email. “While our research team had been working at Angkor for decades, no one had yet decided to tackle this fundamental question,” said Klassen, who conducted most of this work while a postdoctoral fellow at the University of British Columbia.
In an email, Alison Carter, an assistant professor in the Department of Anthropology at the University of Oregon and a co-author of the study, said “it’s only with the recent lidar surveys and Sarah’s machine learning algorithm work that we felt like we had a large enough dataset to tackle the question more comprehensively.”
Indeed, along with the recent lidar surveys, in which lasers were used to create three-dimensional maps of surface features, the team had access to 30 years’ worth of archaeological excavation data, radiocarbon dates, historical archives, and maps. Klassen’s computer algorithm was used to predict the chronology of ancient sites within the city, such as former homes and temples. Together, the multiple lines of evidence helped chronicle Angkor’s growth over time.
“I was amazed by the level of chronological and geographic demographic detail we were able to achieve by combining all these different datasets into a cohesive framework,” said Klassen.
Carter’s contribution was to use the archaeological data to estimate how big a typical household space might have been.
“One easy way to estimate population is to count houses and estimate how many people lived in these household spaces,” said Carter. “In Angkor, this is difficult, because people made their houses from organic materials that haven’t survived in the humid climate of Cambodia. However, the lidar clearly showed the mound spaces where people built their houses.”
Further to this, Carter and her colleagues conducted multiple sessions of fieldwork at residential mounds surrounding the temple at Angkor Wat, each of which measure about 6,450 square feet (600 square meters). Based on this field work, the team believes there was one house for every mound. Ethnographic data provided an estimate of five people per house, “so from there we could start plugging numbers into an equation,” said Carter. And because the amount of mound space was known for each time period, the scientists could track population growth and varying population densities over time.
Results show that Rome, er, Angkor, wasn’t built in a day, as it took hundreds of years for the city to reach its 13th-century peak. What’s more, population growth wasn’t linear, occurring at different rates within the three different zones of occupation: the civic-ceremonial center (home to royal family members and other elites), the metropolitan area, and the embankments.
“It was great to get some numbers to really quantify how huge Angkor was. We could always tell it was a massive, sprawling city, but this adds another layer,” said Carter. “The timing and growth of different parts of the city is really fascinating to me. It was amazing to see that the downtown part of Angkor didn’t get to its most dense occupation until the later periods.”
She suspects that particular rulers or historical events contributed to the city’s growth in various ways.
Carter was struck by the size and scale of Angkor. Accordingly, there’s “a lot we can learn about urbanism and the development and sustainability of cities by studying this site,” she said. Carter was also surprised by “how quickly the agriculturally focused metropolitan area grew early on in the city’s development,” adding that we “tend to think of cities by their high-density ‘downtown,’ but at Angkor this metropolitan area was a key part of the city as well.”
Interestingly, the estimate of 700,000 to 900,000 inhabitants during the city’s peak is consistent with other research. Archaeologist Eileen Lustig from the University of Sydney had previously estimated a population between 750,000 to 1,000,000 people, a number based on rice productivity and the carrying capacity of the landscape.
“Our numbers are pretty close to hers, so I think we’re on to something,” said Carter.
The peak population density at Angkor is comparable to those seen at Teotihuacan in Mexico and Anyang in China (two major pre-modern city centers), but still lower than Caracol, a 7th-century Mayan city in what is now Belize.
“My biggest fear is that all the news headlines will say, ‘One million people lived at Angkor,’ and that is not what our work is saying at all,” Carter said. “We think at its height there were between 700,000 to 900,000 inhabitants, but I think that the number was probably closer to 700,000 than 900,000.”
The reason for this (reasonably) wide estimate, explained Carter, has to do with uncertainties in the way certain spaces were inhabited, like the embankment, which is “still a big question mark for me and needs to be refined,” she said.
In terms of limitations, Klassen said they had to make numerous assumptions, but ones that were “strongly supported.”
“At the end of the day, this paper represents the work of a collection of archaeologists who have worked in this area—some of whom for decades—combining our data, debating back and forth, and coming to agreements on what we think is the most robust method to calculate the population of Angkor given what we currently know about the site,” Klassen explained. “This, of course, sets the stage for decades of work that will fine-tune and revamp these assumptions,” she said, adding that it’s “the nature of scientific inquiry, and it’s pretty exciting to be a part of it.”
Carter said this is “just a model,” and it will need to be tested with more research.
An exciting aspect of this research is that the method used by the team could now be used to study other ancient cities. The new results will also improve our understanding of social complexity, particularly in large urban settings.
“This work represents a crucial building block that allows us to move beyond traditional ‘historical’ research questions,” said Klassen. “With this fine level of detail of how humans aggregated on the landscape over the course of hundreds of years, how the urban centers formed, and how labor was divided across the landscape, we can start asking questions that are relevant to contemporary society,” namely questions having to do with human mobility, land value, and, very importantly, the question of whether or not “we can create resilient urban systems.”