Rows of massive columns in Karnak temple in Luxor, Egypt. (Photo by Jeff Kraft on Shutterstock)
In A Nutshell
- New geological research shows that Karnak Temple’s foundations formed on a natural Nile terrace by about 2520 BC ± 420 years, making it one of the oldest confirmed building sites in Thebes.
- The terrace created a small island surrounded by two Nile branches. Each flood season, the island appeared to rise from the water, an image that may have echoed Egyptian creation myths.
- As the river’s channels slowly filled with silt, temple builders expanded westward onto newly stable ground and even dumped sand to speed the process.
- The findings narrow a century-long debate about when the site became habitable, pointing to an Old Kingdom origin rather than a much earlier one.
SOUTHAMPTON, England — Karnak stands as one of ancient Egypt’s most monumental temple complexes, a sprawling religious center that dominated the landscape near modern-day Luxor for three millennia. Archaeologists have excavated the site for roughly 150 years, yet a basic question has remained stubbornly unresolved: when did people first settle there, and what did the landscape look like when they arrived?
A study published in the journal Antiquity provides the first answer backed by hard geological evidence. Using 61 sediment cores drilled across the site and surrounding floodplain, researchers have reconstructed how the Nile River shaped Karnak’s foundation and determined when permanent occupation became possible.
The evidence points to settlement beginning around 2520 BC, plus or minus 420 years, most likely during Egypt’s Old Kingdom period (roughly 2591-2152 BC). Before that time, the area was unsuitable for construction because fast-flowing water was actively depositing sand across what would become the temple grounds.
Karnak Was Built on a Sacred Island
The team, led by researchers from the University of Southampton and Uppsala University, discovered that Karnak was built on a natural river terrace, a segment of high ground carved out when the Nile incised channels on both its eastern and western sides. This created an island configuration covering approximately 10 hectares, stretching from what’s now called North Karnak to near the southern enclosure wall.
“The terrace segment upon which Karnak was founded is the only area of high ground surrounded by water thus far identified in the Theban area,” the researchers write in their study.
That geological reality may have held profound religious significance. Ancient Egyptian creation texts describe how the creator god manifested as high ground emerging from primordial waters. While earlier Pyramid Texts from the late Old Kingdom reference this concept, later documents like the Middle Kingdom Coffin Texts explicitly develop the idea of a “primeval mound” rising from the Waters of Chaos, known as Nun.
The researchers propose that Theban elites may have chosen Karnak’s location because it possibly reflected their cosmology, though practical factors like proximity to existing settlements across the river at at-Tarif could have been equally important. Each year as the Nile’s flood receded, the island would appear to rise from the water, recreating the moment of creation.
Surrounding the main terrace, lower land at about 71 meters above sea level would flood annually. As waters receded, this emerging ground effectively doubled the island’s seasonal area, completing the impression of land rising from the depths.
Two Major River Channels Flanked the Egyptian Temple
Previous reconstruction attempts, based on limited archaeological excavations and fewer than 10 detailed sediment cores, had proposed that a river channel likely existed east of Karnak at some point. Some researchers also suspected a second channel ran west of the temple’s main north-south axis during the Middle Kingdom (around 1980-1760 BC).
The new coring survey confirms both channels and provides precise dates for their activity. The eastern channel was massive, measuring 220 to 500 meters wide during the Third Intermediate Period (roughly 1076-664 BC), making it a major Nile branch throughout Karnak’s early history. This channel shifted eastward during the Middle Kingdom and Second Intermediate Period before finally silting up during or after the Macedonian/Ptolemaic period (332-30 BC).
On the western side, several minor river courses wound through what’s now the temple complex. During the early Middle Kingdom, one channel had its eastern bank near where the third pylon, a monumental gateway, would later stand. Archaeologists had previously identified ceramics from the First Intermediate Period (around 2152-1980 BC) at relatively deep levels near this location, suggesting a topographic depression that the new study confirms was an active waterway.
Ancient Engineers Manipulated Their Landscape
As the river channels naturally shifted and filled with sediment over centuries, Karnak’s occupants adapted. The temple’s westward expansion after the New Kingdom period (which began around 1539 BC) shows builders taking advantage of newly available land as western channels silted up. Major monumental gateways, the seventh through tenth and first through third pylons, were constructed on ground that had once been underwater.
But temple builders didn’t just wait for nature to cooperate. At one location immediately south of what would become the Hypostyle Hall, one of Karnak’s most famous structures, someone deliberately dumped 3.6 meters of desert sand into a channel. Ceramics found above and below this artificial fill date to around 1540 BC, placing the infill operation at the beginning of the New Kingdom.
The researchers note this action targeted a channel already silting up, with natural channel fill deposits lying underneath the dumped sand.
How Old Is Karnak Really?
Before this study, scholars disagreed about Karnak’s origins. Most favored a First Intermediate Period beginning based on excavated remains and a written reference to a temple of “Ra-Amun” from the reign of Intef II, traditionally dated to around 2066-2017 BC. The earliest excavated structures, found in the eastern part of the complex and southeast of the Sacred Lake, consist of mudbrick walls associated with ceramics from the First Intermediate Period or early Eleventh Dynasty.
Some researchers, however, argued for Predynastic occupation starting as early as 3900-3100 BC, based on artifacts from old excavations and objects found out of their original context.
“It’s tempting to suggest the Theban elites chose Karnak’s location for the dwelling place of a new form of the creator god, ‘Ra-Amun’, as it fitted the cosmogonical scene of high ground emerging from surrounding water,” says lead author Dr. Ben Pennington, a Visiting Fellow in Geoarchaeology at the University of Southampton, in a statement. “Later texts of the Middle Kingdom (c.1980–1760 BC) develop this idea, with the ‘primeval mound’ rising from the ‘Waters of Chaos’. During this period, the abating of the annual flood would have echoed this scene, with the mound on which Karnak was built appearing to ‘rise’ and grow from the receding floodwaters.”
The new sediment evidence settles this debate. The river terrace that forms Karnak’s natural foundation contains no embedded ceramic fragments except at its very surface, indicating the sand was deposited before people occupied the area. In contrast, all the younger river deposits surrounding the terrace contain tens to hundreds of ceramic fragments per vertical meter, showing they formed while people lived at the site.
The oldest ceramics sitting on top of the terrace date from sometime between the Sixth Dynasty and early Eleventh Dynasty (around 2305-1980 BC), falling within either the First Intermediate Period or late Old Kingdom. Similar-aged ceramics appear in the surrounding river deposits, which are younger than the terrace itself.
Luminescence dating of the terrace sands provided an age of 2520 BC plus or minus 420 years. The lack of soil formation or wind-blown deposits between the river sand and the occupation layers suggests little time passed between when the terrace stopped forming and when people arrived.
Taking all evidence together, the researchers conclude that permanent settlement most likely began during the Old Kingdom, soon after the river carved out the island. A Predynastic origin is not supported by the data.
The discovery addresses a question that has persisted since French archaeologist Georges Legrain began systematic excavations at Karnak in the early 1900s. While archaeologists will continue debating details about specific structures and royal building programs, the geological data now provides firm boundaries for when the site could have been occupied at all.
Paper Summary
Methodology
Researchers drilled 61 sediment cores across Karnak and the surrounding floodplain using two techniques: an Eijkelkamp hand auger and a Cobra TT percussion corer. Cores averaged 6.4 meters deep, with the deepest reaching 11.65 meters. The team positioned cores at relatively even intervals along two main transects running west to east through the northern and southern parts of the site. They analyzed basic sediment characteristics in the field, including grain size, sorting, color, and composition. All recovered sediments were wet-sieved in roughly 100-millimeter intervals, and the resulting fractions were manually sorted into ceramic and non-ceramic material. Ceramic fragments were assigned ages within a Karnak-specific typology developed through decades of previous archaeological work. The team interpreted the sediment cores using standard geoarchaeological methods to identify river channels, floodplain deposits, terraces, and cultural layers. They contextualized their work using lithological and chronological information from 142 additional cores drilled across the wider Luxor area and 48 optically stimulated luminescence dates from the vicinity of Karnak.
Results
The survey revealed that Karnak sits on a foundation of sandy river channel deposits, designated Unit T, which formed a raised terrace at approximately 72 meters above sea level. Luminescence dating shows this terrace stopped forming after 2520 BC (±420 years). Subsequent channel incision on both sides created an island roughly 10 hectares in area. The oldest ceramics on the terrace date from sometime between the late Old Kingdom and First Intermediate Period (approximately 2305-1980 BC). A river channel existed east of the site, measuring 220 to 500 meters wide during the Third Intermediate Period. This eastern channel was active during the First Intermediate Period and early Middle Kingdom, then shifted eastward before finally silting up during or after the Macedonian/Ptolemaic period (332-30 BC). Multiple smaller channels existed west of the site, with one having its eastern bank near the third pylon during the early Middle Kingdom. These western channels progressively silted up from the Middle Kingdom through the New Kingdom periods. At one location, builders deliberately dumped 3.6 meters of desert sand into a western channel around 1540 BC to accelerate infilling. A smaller terrace segment existed in the southwest corner where the Opet Temple was later built. During the early Middle Kingdom, lower land at approximately 71 meters above sea level surrounded the main terrace, flooding annually but emerging as the inundation receded, effectively doubling the island’s usable area seasonally.
Limitations
The study faces several constraints. Sample collection from archaeologically sensitive areas was restricted, limiting the number of locations where deep cores could be drilled. Coring cannot precisely delimit the northern and southern margins of the original island, though the data suggests the northern limit was near North Karnak and the southern limit was near the southern enclosure wall. Dating relies primarily on ceramic typology, which provides broad chronological ranges rather than precise years. The earliest ceramics on the terrace span a 325-year period from roughly 2305 to 1980 BC, and the luminescence date for the terrace formation has an uncertainty of ±420 years at the 68.27 percent confidence level. The study cannot definitively prove that the island’s configuration influenced site selection for religious reasons, as practical factors like proximity to existing settlements may have been equally or more important. Five cores drilled near the Mut temple to the south encountered primarily cultural deposits rather than natural sediments, providing limited information about the landscape in that area. The cores provide snapshots at specific points rather than continuous records, so some short-term landscape changes may be missed. The correlation between physical landscape features and religious symbolism, while plausible, relies on interpretation of texts that may not have been well-known in Thebes during the proposed period of initial settlement.
Funding and Disclosures
The research was supported by the Knut och Alice Wallenbergs Stiftelse (KAW 2013.0163) and Uppsala Universitet (HUMSAM 2014/17) to Angus Graham as a Wallenberg Academy Fellow from 2014 to 2020, along with a small grant from M och S Wångstedts Stiftelse. The authors declare no competing interests. The research was conducted under the auspices of the Egypt Exploration Society (London) with permissions from the Egyptian Ministry of Tourism and Antiquities.
Publication Details
Pennington, B.T., Graham, A., Masson-Berghoff, A., Millet, M., Peeters, J., Toonen, W.H.J., Winkels, T.G., Sollars, L.H., Emery, V.L., Strutt, K.D., & Barker, D.S. (2025). “Conceptual origins and geomorphic evolution of the temple of Amun-Ra at Karnak (Luxor, Egypt),” published in Antiquity, October 6, 2025, pages 1-15. DOI: 10.15184/aqy.2025.10185







