Limescale deposits in the water channel were systematically removed / Maintenance was carried out in accordance with Roman recommendations
4 August 2023
The aqueducts constructed during the period of the Roman Empire can provide us with valuable insights on how they were once used and on the climate in the ancient environment. They can also tell us about the social dynamics and changes to population levels at the time. A research team including members of the University of Oxford and Johannes Gutenberg University Mainz (JGU) has demonstrated that the aqueduct of Divona near the city of Cahors in southwest France was subjected to regular maintenance work for a sustainable water supply over almost 100 years. Furthermore, they have been able to deduce what past developments formed in the local socio-economic life. Moreover, as the researchers state in their paper published in Scientific Reports, it would seem that the aqueduct carbonate has for the first time confirmed that such work was carried out as recommended in the treatise on aqueduct maintenance written by the Roman senator Frontinus.
Carbonate incrustations are records of Roman times
The aqueduct of Divona is among the oldest in France. It was originally constructed in the early 1st century CE to supply the inhabitants of the Gallo-Roman town of Divona Cadurcorum, Cahors, with water. The water was obtained from a spring and river in a side-valley of the river Lot near present-day Vers and transported over 31.6 kilometers by the aqueduct to the local public baths and fountains. The aqueduct channel was cut into solid rock over longer stretches and elsewhere was continued as a stone-lined conduit. It probably operated until the 4th or even 5th centuries CE. "The building of the aqueduct is an impressive technical achievement. Because it bore mineral-rich water, it can provide us with fascinating information on what was happening here almost 2,000 years ago," said Dr. Gül Sürmelihindi from the University of Oxford, the lead author of the article.
Roman aqueducts fed by springs and rivers in limestone regions typically conducted hard water that contained dissolved calcium carbonate. As carbon dioxide escaped from the water surface along the course of the aqueduct, this dissolved carbonate precipitated out of solution to create limescale deposits on the channel walls and floor. Over decades these could become several centimeters thick, clogging the aqueduct channel if they were not removed.
"At present, we are not able to say exactly when the current deposits formed in the case of the Divona aqueduct. But they must date from a late phase of the aqueduct, so presumably the 3rd or 4th centuries CE," added Dr. Gül Sürmelihindi, explaining how the detected deposits are being classified.
Roman aqueducts are archives of unwritten history
Such carbonate deposits can be used to study both the history of the use of aqueducts and the local climate in antiquity over long periods, of decades to centuries. However, their potential as environmental and climate archives may be impaired if parts of the deposits were removed during ancient maintenance work. At the same time, this apparent problem creates a new and unique opportunity to study the methods of Roman water management. "Aqueducts are in various ways archives that hold records of unwritten history," emphasized Sürmelihindi.
The research team led by Dr. Gül Sürmelihindi and Professor Cees Passchier, Senior Research Professor in Geoarchaeology at JGU, discovered evidence that the aqueduct of Divona underwent regular maintenance work during which the limescale deposits were partly removed using tools. In addition to marks left by the use of tools, the researchers also identified damage to the calcite crystals, so-called deformation twins of the type that can form when the crystals are struck with a heavy object, such as a mattock. The team first analyzed the seasonal temperature variations of water in the aqueduct and used this profile to count the annual layers, recording 88 years of aqueduct activity.
Reading carbonate stratigraphy like the pages of a book of water management
It was during this period that the aqueduct channel was regularly cleaned. The time intervals between cleaning events were 1 to 5 years, with a mean of 2.8 years. "It seems the cleaning activities themselves were conducted rapidly and took perhaps a month at most. Inherently, such work was never undertaken in summer when people needed most water," said Sürmelihindi. A pattern like this confirms to the advice provided by the Roman senator and author Sextus Julius Frontinus, who wrote a treatise on the maintenance of the aqueducts of Rome when he was responsible for their upkeep as curator aquarum – the only known surviving manual on this subject. The results obtained by the team could thus be the first and the only research to date which attests archaeologically the actual use of Frontinus' theoretical maintenance recommendations.
The team also found two horizons in the deposits where the aqueduct channel was repaired with a red Roman waterproof mortar called opus signinum. During these phases, the water flow was interrupted for longer periods. "It is curious to see such long breaks and imagine the way in which locals managed without a continuous water supply," said Dr. Gül Sürmelihindi. Overall, the conclusion drawn is that there were at least 28 cleaning and two repair events completed over a time equivalent of 88 years. In the concluding stages of use of the aqueduct of Divona, maintenance intervals grew longer and cleaning work was carried out less frequently, preceding the final abandonment of the water supply system. "This could well have been due to population decline and thus reduced water demand, or because fewer resources were available for maintenance," concluded Professor Cees Passchier. A change in the leadership of the maintenance team and hence in the cleaning strategy should also be considered as a reason for a decreasing frequency of cleaning.
New tools in geoarchaeology
"We consider that this form of research could be a powerful new tool in geoarchaeology to determine the status of the local economy and political stability," stated Sürmelihindi. "Regular maintenance can be taken as evidence for a well-structured organization of an ancient town, while less regular maintenance or a total lack thereof may indicate socio-economic stress."
Dr. Gül Sürmelihindi is a geoarchaeologist and uses in her research a combination of geochemical techniques and archaeological concepts to learn more about the history of humankind, natural processes and how they have interacted together. She has been a Marie Skłodowska-Curie Research Fellow at the School of Archaeology of the University of Oxford since October 2021. Previously she worked as a postdoc in the team headed by Professor Cees Passchier at Johannes Gutenberg University Mainz.