Nile waterscapes facilitated the construction of the Giza pyramids during the 3rd millennium BCE

Abstract

The pyramids of Giza originally overlooked a now defunct arm of the Nile. This fluvial channel, the Khufu branch, enabled navigation to the Pyramid Harbor complex but its precise environmental history is unclear. To fill this knowledge gap, we use pollen-derived vegetation patterns to reconstruct 8,000 y of fluvial variations on the Giza floodplain. After a high-stand level concomitant with the African Humid Period, our results show that Giza's waterscapes responded to a gradual insolation-driven aridification of East Africa, with the lowest Nile levels recorded at the end of the Dynastic Period. The Khufu branch remained at a high-water level (∼40% of its Holocene maximum) during the reigns of Khufu, Khafre, and Menkaure, facilitating the transportation of construction materials to the Giza Pyramid Complex.

Keywords: Giza Harbour; Great pyramid; Nile.

Fig. 1.

Location of the cores on the Giza floodplain. The two cores used to reconstruct Holocene variations in Khufu-branch levels (cores G1 and G4) are located where the Khufu basin was connected to the Nile. The Giza Pyramid Complex currently lies >7 km from the present-day Nile branch. Harbor remains were previously found at the site of G3; the fluvial history derives from the work of Lehner (5, 6, 10, 11).

Fig. 2.

Reconstructed Khufu-branch variations of the Nile during the last 8,000 y. (A) Pollen-derived Khufu-branch levels (K-1, blue line) expressed as Loess smoothing scores (with 2.5 and 97.5 percentiles) compared with Nile Delta sedimentation rates (red line) (19). Long-term trends are depicted by polynomial models (order 3, P_value < 0.001) (56). (B) Variations in Lake Victoria (White Nile) (22) and Lake Tana (Blue Nile) (21) reconstructed as a diatom record (red line) and Ti (mg/g; orange line). Long-term trends are depicted by polynomial models (order 3, P_value < 0.001) (56). The Khufu branch is represented by a shaded curve. (C) Summer insolation (56), migration of the ITCZ (25) shown as Wm⁻² (red line) and Fe (cps; purple line), and the Eastern African Monsoon shown as δD_precip (‰ VSMOW) (24). Long-term trends are denoted by polynomial models (order 3, P_value < 0.001) (56). The Khufu branch is illustrated by a shaded curve. The background colors show the transition from the African Humid Period to the drying trend in Egypt.

Fig. 3.

Khufu-branch levels during the Holocene compared with volcanic forcing, East African Monsoon and North African lakes. (A) Volcanic forcing (57) expressed as Wm⁻² and shown as events (red line), 100-y (shaded light purple) and 200-y (shaded dark purple) moving averages. (B) Pollen-derived Khufu-branch levels (K-1, blue line) expressed as 2.5 and 97.5 percentiles, and as a 200-y moving average. (C) East African Monsoon (green line) reconstructed from planktonic diatoms recorded at Lake Abiyata, Ethiopia (23). The long-term trend is depicted by a 200-y moving average (shaded blue line). (D) North African lake levels (blue: intermediate and purple: high lakes) (26) shown as relative abundances. Long-term trends are illustrated by polynomial models (order 3, P_value < 0.001) (56). (E) Nile Delta accretionary status (19) shown in mm per century (orange) with the SD (yellow). The background colors show the transition from the African Humid Period to the drying trend in Egypt.

Fig. 4.

Khufu-branch variations and North African archaeological sites. (A) Level of the Khufu branch (K-1, blue line) compared to the number of archaeological sites in Sudan and Egypt (light purple line) (28). Archaeological site numbers were transformed into z-scores. The two datasets are shown as 1,000-y averages (with a 95% confidence interval). Long-term trends are represented by polynomial models (order 3, P_value < 0.001) (56). (B) Level of the Khufu branch (blue line; Loess smoothing scores with 2.5 and 97.5 percentiles) compared to archaeological sites in Lower and Upper Egypt (orange line) (28). The number of archaeological sites is shown as a 1,000-y sum, transformed into z-scores. Long-term trends are illustrated by polynomial models (order 3, P_value < 0.001) (56). (C) Khufu-branch levels (2.5 and 97.5 percentiles) during the construction of the pyramids of kings Khufu, Khafre, and Menkaure (blue line). The long-term trend is represented by a polynomial model (order 3, P_value < 0.001) (56). The background colors show the transition from the African Humid Period to the aridification of Egypt (A and B).

Fig. 5.

Khufu-branch variations and Egyptian mummies geochemistry. (A) Pollen-derived Khufu-branch levels (K-1, blue line) expressed as Loess smoothing scores (with 2.5 and 97.5 percentiles). Long-term trends are depicted by a polynomial model (order 3, P_value < 0.001) (56). (B) δ¹⁸O ‰ values (with SD) from Egyptian mummies (teeth and bones; light red) (38). The data are shown as averages and with a polynomial model (order 3, P_value < 0.001) (56) to highlight the long-term trend. The background colors show the transition from the African Humid Period to the aridification of Egypt.