Received: 17 June 2021 | | Revised: 15 October 2021 Accepted: 21 October 2021 DOI: 10.1002/gea.21894 RESEARCH ARTICLE The hydro‐geomorphological setting of the Old Kingdom town of al‐Ashmūnayn in the Egyptian Nile Valley Willem H. J. Toonen1,2 Jan Peeters 5 | | Kylie Cortebeeck1 Harco Willems 1 Egyptology Unit, Faculty of Arts, KU Leuven, Leuven, Belgium 2 Earth & Climate Cluster, Vrije Universiteit Amsterdam, Amsterdam, The Netherlands 3 Hasselt University/PXL‐MAD, Hasselt, Belgium 4 Department of Prehistory, Western Asian and Northeast African Archaeology, Austrian Archaeological Institute, Austrian Academy of Sciences, Vienna, Austria 5 Asian School of the Environment, Nanyang Technological University, Singapore, Singapore Correspondence Willem H. J. Toonen, Egyptology Unit, Faculty of Arts, KU Leuven, Leuven, Belgium. Email: w.h.j.toonen@vu.nl Scientific editing by Kevin Walsh. | Stan Hendrickx3 | Bettina Bader4 | 1 Abstract The ancient Egyptian city of al‐Ashmūnayn (Minyā Governorate, Egypt) has been an important regional centre since at least the Old Kingdom (c. 2686–2160 BC). It is assumed to have been founded on the banks of the Nile, although no scientific evidence was hitherto available to support this claim. In this multidisciplinary study, the results of a geoarchaeological survey were combined with the study of pottery fragments. Boreholes placed at al‐Ashmūnayn produced thick layers of late Old Kingdom pottery in association with the local occurrence of river channel deposits, allowing us to conclude that it is very likely that al‐Ashmūnayn originated on the banks of the Nile River. The regional borehole survey demonstrates that major geomorphological reconfigurations of the fluvial landscape occurred throughout the late Holocene, notably by the process of river avulsion. An interconnectedness of changes in the natural Nile Valley landscape and cultural dynamics of the ancient Egyptian riverine society seems possible, based on the coincidence of river reconfigurations with shifts in the preferential locations for high‐status burials in the region. KEYWORDS Dayr al‐Barshā, Egyptology, fluvial geomorphology, Holocene, pottery, river avulsion 1 | INTRODUCTION of the settlement is, however, poorly known (see Section 2.1). The oldest monument visible today is what remains of a temple gate The present al‐Ashmūnayn is an agglomeration of villages in Middle (Figure 2: IV) from the reign of the pharaoh Amenemhat II Egypt (Figure 1a), located at the site of what was a major town in (c. 1911–1877 BC) during the Middle Kingdom (MK: c. 2055–1650 ancient times. The visible archaeological remains of the town mostly BC), which was uncovered during excavations carried out between date back to the New Kingdom (NK: c. 1550–1069 BC) (Figure 2: II), 1929 and 1939 (Roeder, 1959). Remains of an earlier temple were the Late Period (LP: 664–332 BC) (Figure 2: V), the Graeco‐Roman tentatively dated to the late Old Kingdom (OK: c. 2686–2160 BC), period (GRP: 332 BC to 395 AD) and the Late Roman period (LRP: but were poorly documented due to water‐logged conditions 395–641 AD; dates cf. Shaw, 2000, unless stated otherwise) (Steckeweh, 1937). More recent excavations, carried out by the (Figure 2: III). During these periods, the town was an important ad- British Museum, located a late OK to First Intermediate Period (FIP: ministrative and religious centre (Kessler, 1977). The pre‐NK history c. 2160–2055 BC) cemetery to the north of Seti II's Amun temple This is an open access article under the terms of the Creative Commons Attribution‐NonCommercial‐NoDerivs License, which permits use and distribution in any medium, provided the original work is properly cited, the use is non‐commercial and no modifications or adaptations are made. © 2021 The Authors. Geoarchaeology published by Wiley Periodicals LLC Geoarchaeology. 2021;1–17. wileyonlinelibrary.com/journal/gea | 1 2 | TOONEN ET AL. This assumption has become widely accepted, but was not factually based on concrete evidence. It seems, nonetheless, likely that the geomorphology of the Nile Valley landscape and particularly the location of river channels may have played an important role in creating suitable places for settlement, and this would have influenced the connectivity between sites in the wider region. Such interconnectedness of human and natural landscapes has been suggested for the Nile Delta (e.g., Ginau et al., 2019; Pennington et al., 2021) and other regions in the Nile Valley, for example in the Memphite region (Bunbury et al., 2017; Lehner, 2014; Tallet & Marouard, 2014), Thebes (Toonen et al., 2018) and in the Sudan (N. Spencer et al., 2012). The current study presents the results of a regional geoarchaeological survey, based on 33 newly drilled boreholes, complemented with the analysis of pottery fragments from nine dedicated cores at al‐Ashmūnayn. The aims of this study were (i) to determine the presence and extent of OK settlement deposits at al‐Ashmūnayn, (ii) to reconstruct the local environmental context of the earliest cultural deposits, (iii) to reconstruct changes in the regional fluvial landscape during the town's existence while developing an understanding of underlying morphodynamics that caused these changes and (iv) to compare changes in the regional cultural network of sites with changes in the regional fluvial network to establish potential correlations between both. 2 | C U L T U R A L A N D EN V I R O N M E N T A L SETTING 2.1 | Old Kingdom Khemenu The Arabic name for al‐Ashmūnayn is the grammatical dual of the earlier Coptic name Shmūn: ‘the two Shmūns’ (Gardiner, 1947). This may reflect the existence of two settlements in pharaonic times: Khemenu and Wenu. The exact meaning of Shmūn is still debated, but its etymological connection to ancient Egyptian Khemenu was demonstrated by Gardiner (1947). In Egyptian mythical narratives, Khemenu was considered to be the place where a group of eight primordial gods created the world. Wenu, ‘the Town of the Hare’ (Zibelius, 1978, pp. 64–67), refers to the nome (province) of the Hare, of which it was the capital (Figure 1b: 1). The name of Khemenu is earliest attested in the pyramid temple of pharaoh Sahure of the early Fifth Dynasty (c. 2487–2475 BC) (Zibelius, 1978, F I G U R E 1 (a) Location of the study area, indicated by the black box and (b) main places (as referred to in the text) within the Nome of the Hare (boundaries based on Kessler, 1981, pp. 120–185): (1) al‐Ashmūnayn (27°46ʹ53"N; 30°48ʹ14"E), (2) Dairūţ al‐Sharīf, (3) Itlīdim, (4) al‐Shaykh Saīd, (5) Hatnub, (6) Dayr al‐Barshā, (7) Dayr Abū Hinnis, (8) Mallawī, (9) al‐Shaykh ʽIbāda and Antinoöpolis, (10) Tūna al‐Jabal, (11) Amarna, and (12) Rairamun. Places outside of the Nome of the Hare; (13) Banī Ḥasan, (14) Ḥūr and (15) Balanṣūra [Color figure can be viewed at wileyonlinelibrary.com] p. 189; Zivie‐Coche, 2009). The name Wenu is even older, being attested for the first time in the Early Dynastic Period (EDP: c. 3100–2686 BC) (Zibelius, 1978, pp. 64–66). Although several large monolithic baboon statues representing Thoth, the main deity of Khemenu, were excavated from al‐Ashmūnayn (Figure 2: I), most information on Khemenu derives from burial sites in the region, which suggests that a major urban centre existed at least from the late Fifth and Sixth Dynasties onwards (c. 2400–2180 BC). The highest officials of the Hare nome of that time had their tombs built at al‐Shaykh (Figure 2c). Yet, no direct in situ evidence is available for the ex- Saʽīd (Davies, 1901), in the limestone cliffs that border the alluvial plain istence of an OK settlement at the site of al‐Ashmūnayn. (Figure 1b: 4). Some of the inscriptions found in these tombs explicitly Butzer (1976) argued that the OK town probably emerged on a refer to Khemenu. No tombs are known for the nomarchs of the FIP, but natural levee, in most years not reached by the Nile flood. in the alabaster quarries of Hatnub in the Eastern Desert (Figure 1b: 5), TOONEN ET AL. | 3 F I G U R E 2 (a) Study area with the locations of the borehole transects and individual boreholes, indicated with dots. (b) Satellite view of al‐ Ashmūnayn (Google Earth, 2020). (c) Research area focussed between Roeder's excavation trench and Nectanebo I's Thoth temple in the north, with borehole locations and the approximate location of the First Intermediate Period cemetery (A. J. Spencer, 1993). (I) Monolithic statue representing Thoth fronting the visitor centre (ex situ), (II) New Kingdom Seti II's Amun temple, (III) late Roman basilica, (IV) Middle Kingdom Amenemhet II's temple gate located in Roeder's trench (overlay by W. Hovestreydt, the Netherlands Institute for the Near East), (V) Late Period Nectanebo I's Thoth temple—the locations of these sites are marked in frames B and C [Color figure can be viewed at wileyonlinelibrary.com] several of the early inscriptions render nomarch names (Anthes, 1928, These texts refer to the tomb owners as a ‘revered one near the One who pp. 18–24). Names like Djehutinakht include the element ‘Djehuti’, which is in Khemenu’, suggesting that both al‐Shaykh Saʽīd and Dayr al‐Barshā is the name of the god Thoth, and thus suggests that the nomarchs were part of Khemenu's sphere of influence, and probably acted as sa- originated from Khemenu. tellites for burial activities. Other cemeteries in the region that may have At Dayr al‐Barshā, c. 5 km north of al‐Shaykh Saʽīd and c. 10 km east been connected to Khemenu are a late OK cemetery on the southern of Khemenu/al‐Ashmūnayn (Figure 1b: 6), there are several late OK in- outskirts of the modern village of Dayr al‐Barshā (Willems, 2013) and scribed tombs of officials that refer to a deity designated as ‘he who is in extensive early OK low‐status cemeteries in the foothills of the Eastern Khemenu’, which may again refer to Thoth (De Meyer, 2008, pp. 48–60). Desert between Dayr al‐Barshā and Dayr Abū Ḥinnis (Vanthuyne, 2017) 4 | TOONEN ET AL. (Figure 1b: 7). Moreover, an early Fourth Dynasty industrial site and Malouta (2012) applied this model also to the al‐Ashmūnayn region quarry settlements recently discovered near al‐Shaykh Saʽīd may also (Figure 3). Yet, the model has not been supported by targeted geomor- have been linked to Khemenu (Willems et al., 2009). phologic reconstruction of the past landscape neither in our area of interest, nor in the Sohag region for which it was initially formulated (Figure 1a). 2.2 | The fluvial environment The regional landscape may have played a pivotal role in Khemenu's and 3 | METHODOL OGY later al‐Ashmūnayn's importance as a major town and administrative centre throughout ancient Egyptian history. First, its location in one of the A multidisciplinary approach was used to reconstruct the past fluvial widest parts of the Nile Valley, measuring just over 15 km across, offered landscape and to date natural and cultural deposits. This reconstruction an extensive agricultural hinterland. A second factor of importance may was based on the general characteristics and architecture of deposits have been its supposed proximity to the river, allowing trade and facil- indicative for specific formative processes, agents of transport and the itating transport between the town and its satellite locations, and beyond. resulting landforms (cf. Bridge, 2003; Miall, 1985). As the shipping of Although the present Nile is positioned at the eastern fringe of its valley sample material for radiometric dating is restricted in Egypt, our main (Figure 1b), and is assumed to have been active in that zone throughout at source of dating information was the analysis of pottery fragments least the last few millennia (Verstraeten et al., 2017), it has been sug- contained in the sediments—following a similar methodology as that gested that the Nile flowed through the central part of the valley during used at many other sites in Egypt (see Section 3.2 and Supporting the OK (Butzer, 1976). His suggestion was based on the assumption that Information SA). floodplain settlements could only emerge on high landscape features, such as levees, that were beyond the reach of regular Nile floods. Accordingly, the location of Khemenu was inferred to imply the position of 3.1 | Survey of subsurface deposits the Nile in the centre of the valley. The main channel of the Nile would have migrated steadily in an eastward direction since the OK. Butzer's Boreholes were drilled at 33 locations in three main zones: (i) nine eastward channel migration premise has become a widely used model for boreholes distributed within the archaeological zone of al‐Ashmūnayn, the long‐term change of the fluvial landscape in Egypt; Bunbury and and two additional boreholes that flank the tell (town mound) on each F I G U R E 3 Historical map of the research area (composition made by Willems et al., 2017; adapted from Jacotin, 1821) with the location of waterways in blue: the main Nile channel in the eastern part of the Nile Valley, the Bahr Yussef in the west and various irregularly shaped canal systems in between. The black arrows indicate meander bend migration as suggested by Bunbury and Malouta (2012), based on the interpretation of anomalous field patterns (dashed lines). Al‐Ashmūnayn is located in the centre of the map: Ruines d'Hermopolis Magna [Color figure can be viewed at wileyonlinelibrary.com] TOONEN | ET AL. 5 side, (ii) 13 boreholes in a southwest–northeast‐oriented transect derived from the vertical succession throughout the borehole. In the passing the modern town of ʽIzbat Tabūt (Figure 2a) and (iii) another absence of an independent absolute radiometric dating control and nine boreholes in a west–east‐oriented transect in the floodplain area other (inscribed) finds tying samples to the dynastic history of Egypt, between the towns of Qulubba and Mallawī (Figure 2a). At al‐ the dating was largely based on comparative analyses of the materials Ashmūnayn, two boreholes (#1 and #2) were placed near the MK gate with well‐dated sequences of pottery and singular finds from other in the deepest part of Roeder's excavation trench (Figure 2c), from sites. Table SA1 presents a detailed overview of all the used which deposits of the last c. 4000 years had already been removed. analogues—OK pottery was mainly compared with pottery chron- Five boreholes (#3–7) were drilled between the NK Amun temple of ologies from the Memphite region just south of modern Cairo Seti II (c. 1200–1194 BC) and the LP Thoth temple of Nectanebo I (c. (Kytnarová, 2009; Rzeuska, 2006). 380–362 BC) to the north, to cover the area around the FIP cemetery Pottery derived from boreholes (#4–6) that demonstrated clear (Figure 2c). The remaining boreholes (not rendered in Figure 2—all mixing upon initial inspection was not considered in detail in this borehole locations provided in Supporting Information SB) were po- study. Mixing, most likely caused by the digging of foundations and sitioned east of the paved road leading to the visitor centre of al‐ the excavation of archaeological material, was usually easily identi- Ashmūnayn. fied inter alia by the presence of distinct brown LRP Amphorae The two borehole transects to the south of al‐Ashmūnayn were (Type 7) fragments, whose fragments dominate the present surface aimed at investigating the potential fluvial context of the settlement, deposits at al‐Ashmūnayn. The identification codes for diagnostics which is difficult to do systematically at the archaeological site due to were compiled from the borehole identification number, followed modern occupation and the thick overlying anthropogenic layers after a slash by its depth interval (in metres) and then an additional containing numerous stone blocks. The general west–east alignment character if multiple specimens were derived from the same interval, of the transects is perpendicular to the general south–north slope of for example, specimen F from a depth of 4.4–4.6 m in borehole #3 is the Nile Valley, and allows one to produce cross‐sections of the identified by the following code: #3/4.4‐4.6F. landscape that are used to identify past landforms. The general spacing of c. 200 m between individual boreholes ensures that no major geomorphologic elements in the subsurface, such as channel 4 | RESULTS belts, were missed. Boreholes were drilled using an Eijkelkamp percussion drilling set and a manual Edelman auger. At the site of al‐Ashmūnayn, only the percus- 4.1 | Sedimentary units and regional geomorphologic interpretation sion core sampler was deployed to penetrate pottery‐packed occupation layers. For both soil‐sampling kits, the maximum coring depth was c. 10 m The two lithological profiles to the south of al‐Ashmūnayn show a below the surface. The main sedimentary characteristics were described rather similar lateral and vertical succession of deposits (Figure 4). Six in the field following the USDA (2017) soil classification system (borehole main sedimentary units were recognized and interpreted as the basic logs provided in Supporting Information SB). The location of boreholes geomorphological elements of the past riverine landscape: a coarse‐ was recorded using a handheld Garmin GPS. A dumpy level and grained regional substrate (Unit A), sandy accretional channel de- topographer's rod were used to determine local relative elevations, posits (Unit B), heterogeneous channel‐fill deposits (Unit C), levees which were subsequently linked to a local Survey of Egypt point at (Unit D), fine‐grained floodplain deposits (Unit E) and minor channel al‐Ashmūnayn for Datum reference. and/or crevasse‐splay deposits (Unit F). Unit A represents the local substrate with its top located at a depth of 8–10 m below the present surface (c. 32–34 m + msl: above the mean 3.2 | Study of pottery fragments sea level). The medium to coarse sands (300–500 μm) were poorly sorted. Borehole collapse prevented deeper penetration; however, Attia (1954) Pottery fragments were collected and bagged in the field per c. suggests a thickness of this regional substrate, which consists of coarse 10–20 cm interval. After rinsing and drying, fragments larger than sands and gravels, of at least 35 m. Similar deposits exist throughout the 5 mm were selected for further study. The bulk of these pottery finds Egyptian Nile Valley, and are thought to have formed in a distinctly dif- consisted of body fragments, but also included a scatter of diagnostic ferent fluvial environment that existed in the Early Holocene specimens, such as rims, bases and handles (Figures 6–8). These were (Butzer, 1998; Said, 1993). The transition with the overlying sedimentary described and drawn if sufficiently large and well enough preserved. units was characterized by the frequent occurrence of calcium carbonate This type of study unlocks chronological information by scrutiny of nodules (rhizoliths), and iron and manganese flecks. Such pedogenic fabrics, surface treatment, manufacturing technology and shape—all features suggest that the top of the regional substrate has been an ex- of which changed over time (Bourriau, 2006). The classification sys- posed surface level for a prolonged period (cf. Klappa, 1980); similar tem is presented in Supporting Information SA. features have been observed in the Luxor region (Toonen et al., 2019) Pottery dates were based on the information from individual and in the Nile Delta (e.g., Chen & Stanley, 1993). diagnostic fragments combined with the characteristics of the as- Unit B consists of a c. 800–1000 m wide zone of (loamy) sands semblage of finds at the same level and relative age indications that rises to max. 3 m above the regional substrate. These quartz‐rich 6 | TOONEN ET AL. F I G U R E 4 Generalized lithological profiles of the two borehole transects south of al‐Ashmūnayn—location in Figure 1. Interpreted geomorphologic elements/units; (A) regional substrate, (B) channel belt, (C) residual channel fill, (D) levee, (e) floodbasin and (F) crevasse splay sands fine upward, from 500 μm at their base to c. 100 μm in their This unit flanked Unit B in 200–300 m wide zones, where it directly top levels. Compared to the substrate, these sands are much more overlies the regional substrate and reached up to similar levels as heterogeneous, with 1 cm‐scale loam and clay laminations. Mm‐scale Unit B. It fined and thinned at increasing distance from the channel sands. layering within the sands was interpreted as cross‐bedding features Based on its configuration, the texture of deposits and the lateral fining, that formed within the active river channel, as internal structures this unit was interpreted as the levees of the aforementioned river found in bars or dunes (Miall, 1985). The width of this sand body channel. reflects the maximum possible width of the active river channel. Unit E is dominated by homogeneous brown‐grey‐coloured Presumably, the active channel would only have occupied part of the silty clays that often lacked clear sedimentary structures. Their sand body, with lateral movement leaving the sandy deposits at its mean grain size ranged between c. 10 and 30 μm. Unit E was found flanks. Given the low‐sinuosity and cohesive clayey banks of the Nile, in all boreholes, but its thickness varied widely; where Units B and F the amount of lateral migration would probably have been limited, were absent, the silty clays reached from the substrate to the suggesting the presence of a large (primary) channel of the Nile close present surface. These alluvial fines were interpreted as common to al‐Ashmūnayn. A weak palaeosol was encountered in the top of floodplain fines that were deposited as overbank material during Unit B at the transition with overlying fine‐grained sediments. At this Nile floods. level, c. 6 m below the present surface, many charcoal flecks and Unit F is a group of (silt‐)loam to fine‐sand deposits found in degraded pottery fragments were found. This suggests that the re- the upper metres of the floodplain. Their thickness was usually c. latively high grounds of the channel belt and its levees were a pre- 1–1.5 m, and the dominant texture was loam. At a few places, the ferred location for human activity. thickness of this unit increased to c. 3 m (Figure 4), where sedi- Unit C represents local anomalies in the sandy deposits of Unit B ments also coarsened to fine (loamy) sands (100 μm). The location (Figure 4). Instead of sands, up to 3 m thick sequences of strongly of these roughly matches those of irregularly shaped waterways laminated loams were found, at 6–9 m depth. A slight downward indicated on historical maps of the region (Jacotin, 1821; Willems coarsening occurred towards the bottom, where such sequences et al., 2017) (Figure 3). Based on our observations, it seems plau- graded into the sands of Unit B. The mm‐scale laminations can be sible that parts of the modern drainage system have emerged from linked with individual flood events, as recorded in the sedimentary fill a pre‐existing network of minor channels, whose configuration, of residual channels (Toonen et al., 2012). together with the presence of relatively thin sheets of loamy ma- Unit D is composed of loamy sediments, ranging from loamy sands to silty clay loams (30–80 μm), that often featured mm‐scale laminations. terial, resembles that of crevasse‐splay et al., 1989; Toonen et al., 2015). complexes (Smith TOONEN 4.2 | ET AL. | Deposits at al‐Ashmūnayn 7 and again between 5.3 and 6.3 m depth (Figure 5); yet, pottery fragments were relatively rare. At al‐Ashmūnayn, the deepest core in Roeder's excavation trench At borehole #7 (c. 41.5 m + msl), located at c. 50 m northwest of (Figure 2c: #1), reached 9 m below the local surface (at 41 m + msl; Seti II's Amun temple, relatively coarse sands were encountered Figure 5). The lowest 2 m of deposits consisted of fine to medium below 7.5 m depth (33.5 m + msl). The upper 7.5 m of sediments were sands (up to c. 300 μm) with mm‐scale layering and intercalated clay dominated by loams, but highly variable, with a texture ranging be- loams. Charcoal and pottery fragments were found in the top half of tween silty clay loam (20 μm), sandy loams (50 μm) and occasionally c. these sands. Above the sands, a c. 3 m thick sequence of laminated 30 cm thick intervals of sands (200–300 μm). The entire interval clay loams was found (34–37 m + msl), again with pottery and an above 37 m + msl was marked by an increase in the occurrence of abundance of charcoal fragments. The upper 4 m of sediment con- pottery fragments, charcoal flecks and chunks of clay (Figure 5). sisted of sandy deposits with many pottery fragments and occa- Other boreholes at al‐Ashmūnayn did not penetrate the mixed sionally some charcoal fragments. The loose structure and high water top‐soil or had mixed stratigraphies and are therefore not discussed content of these sands resulted in a rather poor sample recovery in here; their data are presented in Supporting Information SB. As they the bottom metre of this unit. did not yield significant geogenetic or chronological information, they Borehole #2 had a lower depth. The lowermost deposits around were excluded from further analysis and interpretation. 6 m depth (35 m + msl) were clay loams with many charcoal frag- Borehole #10 was placed in an agricultural plot c. 150 m west of ments. These correlate to the deposits at similar depth in the nearby al‐Ashmūnayn (Figure 2b). At least 8.8 m (42.0–33.2 m + msl) of fine‐ borehole #1 (Figure 5). Above this level, relatively fine‐grained sandy grained deposits, predominantly silty clay loams, were found. The top deposits were encountered all the way up to the surface. The lower 6 m was characterized by the occurrence of loamy laminations and part of these sands contained many charcoal and pottery fragments. small pottery fragments. Between 6 and 7 m depth (36–35 m + msl), a Between 2.5 and 4.5 m depth (37.5–39.5 m + msl), the deposits were loamy sand interval (100 μm) was found, devoid of any archaeological loamier and strongly laminated. The top of this sequence consists of material. The stiff silty clay loams in the lowest metre featured small unconsolidated sands, similar to borehole #1. pottery fragments. Borehole #3, with a surface elevation of c. 41.5 m + msl, featured Borehole #11 was located c. 125 m east of al‐Ashmūnayn comparable deposits as those found in boreholes #1 and #2. Medium (Figure 2b). The top 9 m (42.0–33.0 m + msl) had a texture ranging sands at the bottom (33.5 m + msl) were occasionally layered and from silty clay loam to sandy loam, with some pottery fragments contained some charcoal fragments, but only a few pottery frag- occurring in the upper metre and between 1.7 and 7 m depth. The ments. The number of charcoal fragments increased in the overlying basal metre (33–32 m + msl) consisted of sandy deposits that coar- c. 1 m of silty clay loams. After another metre of medium sands, again sened downward from 80 to 300 μm. 50 cm of clay loam was found. Both contained an abundance of charcoal fragments and pottery. Above c. 37 m + msl (Figure 5), unconsolidated sandy deposits reached to the present surface. Locally, these sands were characterized by many laminations, charcoal and 4.3 | Pottery finds and chronologies from al‐Ashmūnayn chunks of clay. The latter are probably rip‐up clasts, the remains of clayey deposits or laminations that were eroded by river flow and its Four boreholes at al‐Ashmūnayn produced reliable continuous fragments rounded during transport. chronologies: #1‐3, and #7 (Figure 2c). Despite the presence of many Borehole #4 was farthest north and placed in the excavation pottery fragments, borehole #1 produced only 12 diagnostic frag- trench of the Nectanebo I's Thoth temple at c. 4 m (40 m + msl) below ments, of which eight were well enough preserved and sufficiently the surrounding surface levels. Mixed sandy deposits started at a large to be drawn (Figure 6a). The deepest segment of the core that depth of 1.5 m, but pristine sands appeared from c. 2.5 m yielded pottery, between 6.0 and 8.0 m depth (33–35 m + msl), pre- (37.5 m + msl; Figure 5) and continued down to the base of the dominantly contained medium coarse Nile B2/C1 fabric specimens borehole at 8 m depth. The sandy deposits demonstrated an overall (similar to Nile C1‐2 in the Vienna System; Nordström & coarsening‐down trend from 300 to 800 μm with an increase in small Bourriau, 1993, pp. 173–174, pl. II.e‐i). For these, it was not possible gravels. The upper 1.5 m of deposits was heavily disturbed and to determine the original vessel shape. In addition, a few fragments contained a few pottery and limestone fragments. were recovered of a fine Nile clay fabric (Nile A‐B1), belonging to Borehole #5 (c. 43 m + msl) was positioned north of the FIP red‐slipped and polished (inside and outside) OK bowls or plates, cemetery (Figure 2c). Sands occurred below c. 6.5 m depth most evident from a rim specimen (#1/6.9‐7.0: Figure 6a). A more (36.5 m + msl). These coarsened down from 300 μm in the top to precise date could not be suggested because this type of bowl oc- 500 μm in the basal parts of the borehole at 10 m depth. The strongly curred throughout the OK (c. 2686–2160 BC). internally laminated top 6 m of sediments was very heterogeneous, Except for a single beer jar body fragment at 5.4 m, no diag- with layers of (silty) clay loam to loamy sediments (20–50 μm) in- nostic pottery was retrieved between 5.2 and 6.0 m depth tercalating with more sandy loams and loamy sands (60–80 μm). A (35–35.8 m + msl). Many fragments were found between 4.2 and concentration of charcoal flecks was found between 2.8 and 3.4 m 5.2 m depth (36.8–35.8 m + msl), with a suggested late OK date. 8 | TOONEN F I G U R E 5 Sedimentary logs of the al‐Ashmūnayn boreholes; for their location, see Figure 2c [Color figure can be viewed at wileyonlinelibrary.com] ET AL. TOONEN | ET AL. FIGURE 6 9 Pottery drawings of boreholes #1 and #2—see Supporting Information SC for grouping per pottery type Most body fragments there belonged to beer jars; a single diag- several minor intervals. These were caused by poor sediment recovery nostic sherd (#1/4.8‐4.9A) was similar to those of phase III–IV at due to unconsolidated deposits and do not, therefore, represent gaps in Saqqāra (Rzeuska, 2006, p. 382), corresponding to the reign of the cultural occupation history. The lowest section (5.5–6.0 m depth; Pepy II (c. 2278–2184 BC) until the end of the Sixth Dynasty. This 35.0–35.5 m + msl) of the borehole was dominated by medium coarse date was supported by a large Nile B1 bowl fragment with a Nile B2/C1 fabrics, similar to borehole #1, and were interpreted as body modelled rim (#1/4.8‐4.9C). Two fragments of Nile B1‐B2 jars with fragments from beer jars of types that occurred from the late EDP to a modelled rim but without an actual neck (#1/4.3‐4.5 and #1/4.8‐ the late OK (cf. OK examples in Rzeuska, 2006, pp. 382–383). The 4.9B) may have been part of medium‐sized or small jars compar- combination of these with fragments of fine red‐slipped and polished able to types that occurred rather frequently during the Sixth plates and dishes (Op de Beeck, 2004; Rzeuska, 2006, pp. 198–260), as Dynasty at Saqqāra and more exceptionally at Abū Ṣīr (Table SA1). well as bread moulds (Faltings, 1998; Jacquet‐Gordon, 1981, pp. Another red‐slipped and polished plate fragment with a modelled 125–136) found between 5.5 and 5.7 m supports a general late OK date rim (#1/4.2‐4.3) was equally characteristic for the late OK in the (Fifth or Sixth Dynasty). The diagnostics at this level are not particularly Memphite region and further south at Iḥnāsīya al‐Madīna and al‐ large or distinctive, but the fabrics and surface treatments support the Kāb (Table SA1). Between 3.0 and 4.2 m depth, no pottery was proposed age (e.g., #2/5.6‐5.7A with a parallel at Abū Ṣīr; Table SA1). found as a result of poor sample recovery. At 2.9–3.0 m depth, Similar pottery material continued to be found up to a 4.0 m some beer jar fragments were found, but no other diagnostic depth (37 m + msl), with a general OK to late OK date. OK beer jars material in the upper 3 m (38–41 m + msl). The present surface appeared prominently, along with red‐slipped and polished open level at the borehole locations in Roeder's excavation trench cor- vessels and wide bread moulds (Jacquet‐Gordon, 1981, fig. 2–3; responded with the floor level of MK monuments; overlying Faltings, 1998, pp. 129–134). Diagnostics found between 4.0 and younger deposits were excavated and removed in the past. 4.5 m, and a single specimen at 4.8–4.9 m (Figure 6b), pointed to a Borehole #2 was drilled from the same surface level as borehole #1. more specific final OK to FIP date. These included a wide bread It produced 21 diagnostic pottery fragments, of which 18 were drawn mould (#2/4.4‐4.5B), a ‘Maidūm’ bowl (#2/4.2‐4.4C), shallow plates (Figure 6b). A high density of pottery, including many large fragments, or dishes with parallel finds in the western Nile Delta, at Giza, Saq- was found continuously throughout the 6 m core, with the exception of qāra, Abū Ṣīr, Iḥnāsīya al‐Madīna and al‐Kāb, with a groove under the 10 | TOONEN ET AL. interior of the rim (#2/4.0‐4.2 and #2/4.2‐4.4A) and sometimes also of 1.9 m. While the material found above 1.8 m was different from on the exterior (#2/4.4‐4.5A; similar to #1/4.2‐4.3 and #2/4.8‐4.9)— that below, a specific age assignment was difficult; the sherds do not see parallels from other studies listed in Table SA1. appear to be as late as the NK or later, but there no unequivocal MK The number of beer jar fragments declined between 3.4 and material was present. A single diagnostic (#2/1.7‐1.8) of the base of a 4.0 m, and almost no pottery was recovered from between 3.0 and hand‐modelled and trimmed closed vessel with a fine version of Nile 3.4 m depth (37.6–38 m + msl), hampering any age inference for this B2 fabric might indicate a FIP to early MK date based on its manu- level. The lack of pottery may relate to a chronological change or to a facturing technology. This could not, however, be confirmed as no temporal shift in local occupation or use of this site. Overlying pot- exact external parallels are yet known for this type of vessel. tery fragments, between 2.0 and 3.0 m depth (38–39 m + msl), were, Borehole #3 contained a total of 114 diagnostics, amidst hun- however, still roughly attributable to the OK, based on their similar dreds of nondiagnostic pottery fragments, of which 41 were drawn characteristics as material from deeper strata. This included beer jars, (Figure 7). At depths between 4.8 and 7.7 m (33.8–36.7 m + msl), red‐ bread moulds and fine red‐slipped and polished body fragments of slipped and polished pottery of a Nile B1‐B2 fabric were found to- open vessels, and diagnostics that were comparable with some dis- gether with Nile B2/C1 fragments. The former belong to plates and cussed above (#2/2.3‐2.4 being similar to #2/4.4‐4.5A; Figure 6b). dishes, and the latter belong to bread moulds and beer jars. The The amount of pottery remained low in the upper 2.5 m of the pottery assemblage suggests a date in the (late) OK, in accordance borehole. No coarse Nile C diagnostics or body fragments of bread with the finds in borehole #2. The first indisputable OK diagnostics moulds and beer jars were identified. However, a few fragments of supporting this date were a rim fragment of a ‘Maidūm’ bowl finer open vessels with red slip and polish were found up to a depth (#3/6.4‐6.6) and a rim of a dish with a polished red slip on the interior FIGURE 7 Pottery drawings of borehole #3—see Supporting Information SC for grouping per pottery type TOONEN | ET AL. 11 and exterior (#3/5.8‐6.0) (Table SA1). Other fragments supporting (#3/2.2‐2.4B) (Table SA1). The latter is diagnostic for the classic MK the late OK date include a partially preserved spout (#3/5.4‐5.6), (Twelfth Dynasty; c. 1985–1773 BC). Several diagnostics belonged to which is typical for several OK vessel types (Table SA1)—though dishes that were only slipped on the interior (#3/3.2‐3.4B, #3/1.8‐ spouted forms continued to be used into the MK and later (e.g., 2.0), excluding the top of the rim, of which some examples have been Aston, 2004, pp. 28–32; Czerny, 1999, pp. 153–154). The highly found in MK shaft burials at Dayr al‐Barshā (e.g., the shaft tomb of quartz‐tempered marl clay body fragments found at this interval re- Ahanakht I; unpublished). The fabrics and surface treatments of sembled the Marl C fabric of large handmade storage jars, of which pottery in the upper 0.7 m indicated an LP to LRP age, for example parallels were found at Saqqāra and Iḥnāsīya al‐Madīna (Bader, 2009, Marl A5 jars, LRP Amphora 7 and greenish marl clays. figs. 7e,f, 12b; Rzeuska, 2006, pp. 122–125). This fabric occurred in Borehole #7 produced 18 diagnostics that were well enough the OK at Giza and Dahshūr, but its use continued later as well (e.g., preserved for drawing (Figure 8)—13 of these belonged to bowls and Bader, 2001, p. 41; Bader, 2002, p. 29; Bader, 2009, p. 34; Nord- dishes of types that have already been discussed for other boreholes. ström & Bourriau, 1993, pp. 179–181; Ownby, 2009, pp. 120–122). Other diagnostics were present, but were generally less useful for Between 4.6 and 6.0 m (35.5–36.9 m + msl), body fragments of dating purposes compared to those discussed above, as most diag- red‐polished dishes and rough bread moulds or beer jars were found in nostics related to Nile B1, B2 or rougher B2/C1 open vessel types, association with fragments of a suggested FIP to early MK date. These with a range of surface treatments, but morphologically rather con- were fragments of uncoated dishes and body fragments with scrape sistent with direct rims. marks on the exterior. The scrape marks were clearly identified as a FIP Pottery from 8.4 to 4.8 m depths (33.1–36.7m + msl) included Nile to early MK style, as they are hardly ever observed on OK pottery (cf. C1/2 body fragments, which could be related to (late) OK beer jars. Rzeuska, 2006 and at other hallmark OK sites). Hence, this interval Found with these were fragments of red‐slipped and polished open represents an overlap of periods or a transition from the late OK to FIP vessel types, present from 7.0 to 3.4 m. Despite these finds, it was not and the early MK. Above this interval, most of the diagnostics and the possible to unequivocally determine whether the material found in the documented vessel technology of the body fragments suggest a post‐ deeper part of this borehole is attributable to the OK or the FIP. OK date; polished open vessel types, handmade pottery and potential rough bread moulds or beer jars are absent. Around 4.8 m and continuing upward to a depth of 2.7 m (38.2–40.3 m + msl), pottery fragments with red slip on the interior Between 3.4 and 4.6 m (36.9–38.1 m + msl), many diagnostics and and rough scraping on the exterior began to appear in combination body fragments, with red slip on both inner and outer sides, appeared with late OK types. The pottery assemblage was tentatively dated as to belong to a transitional phase in pottery‐making traditions during transitional to the FIP or the early MK. Scraped pottery fragments the FIP and early MK: #3/4.4‐4.6B, #3/4.2‐4.4, #3/3.8‐4.0 A, #3/3.4‐ with red slip on the interior continue up to 1.2 m (36.7 m + msl), 3.6 A, #3/3.4‐3.6 C, #3/3.4‐3.6D (Table SA1). They are not polished, where earlier material is considered to be residual. Unfortunately, which is rather uncommon for OK open vessel types. A considerable securely datable MK specimens were rare. Diagnostic material number of pottery fragments of open vessels showed red slip on only included two fragments of dishes (Nile B2) with red slip on the in- one side and/or scrape marks on the exterior, which are clear indica- terior, which stops on top of the rim (#7/1.8‐2.0 and #7/2.0‐2.2B), tions for a FIP to early MK date, of which examples were found up to and one of a Nile C2 fabric with red slip on the interior and a rope 1.5 m. Among the diagnostics were several direct rims from plain cups/ impression on the exterior (#7/2.0‐2.2A). Schiestl and Seiler (2012) bowls (#3/4.6‐4.8A, #3/4.4‐4.6E, #3/3.8‐4.0C, #3/3.8‐4.0F), a rim of a dated comparable specimens to the full span of the MK, while Bader carinated bowl with a groove inside (#3/4.6‐4.8B), a rim of a red‐ (2009) attributed roughly similar specimens to a more confined range slipped (interior and exterior) carinated bowl (#3/4.4‐4.6A), open of the FIP to early MK, which was possible due to the specific context vessel types with red slip inside and a flattened rim (#3/4.4‐4.6C/D) of their finds (Table SA1). Two Nile B2/C1 fragments (#7/7.6‐7.8B and a bowl with a flaring rim (#3/3.4‐3.6B) (Table SA1). The base and #7/3.8‐4.0B) were identified as probably belonging to tall stands, fragment of a miniature tubular bread mould (#3/4.4‐4.6F) and body based on the manufacturing technology. No exact parallels are fragments presumably of the same vessel type were found at a depth available for this specific type, although stands were common during of 4.6 m. Parallels for these vessels are hitherto only known from the MK (Table SA1). Pottery fragments in the top of the borehole nearby Dayr al‐Barshā (Table SA1). Also, a modelled rim fragment of a (0.2–0.4 m depth) were determined as LRP Amphora 7 (c. third larger tubular bread mould (#3/3.8‐4.0E) was found, of which parallels to eighth century AD) and a single assumed ex situ Marl A2 rim of are known from FIP to early MK contexts at Iḥnāsīya al‐Madīna, Dayr an NK jar. al‐Barshā, Dandara, ʽIzbat Rushdī and Asyūṭ (Table SA1). Diagnostic pottery finds from 0.8 to 3.4 m depth (38.1–40.7 m + msl) represent MK types. The assemblage included hemispherical 5 | D IS CU SS IO N cups or their precursors (#3/3.2‐3.4C, #3/2.4‐2.6A, #3/2.2‐2.4A, #3/ 2.0‐2.2; their fragmentary nature does not allow to relate them to 5.1 | The origin of al‐Ashmūnayn regional typo‐chronologies), rim fragments of tubular bread moulds (#3/3.2‐3.4A, #3/2.6‐2.8), necked jars (#3/2.8‐3.0B), a jar with a The occupational history revealed by qualitative analysis of the pottery rounded lip rim (#3/1.3‐1.5) and a Marl C rim of a large storage jar provides an OK to late OK age in borehole #1 below c. 37 m + msl, 12 | FIGURE 8 TOONEN ET AL. Pottery drawings of borehole #7—see Supporting Information SC for grouping per pottery type while borehole #7 shows a similar age below c. 36.5 m + msl and A transition from the late OK to the early MK can be observed in borehole #3 below c. 36 m + msl (Figure 5). Borehole #2 reached late the material from boreholes #3 and #7. Both have an overlap of late OK pottery at c. 37 m + msl (Figure 5). Hence, the newly uncovered OK and FIP material for c. 2 m (#7 at 36.5–38.5 m + msl; #3 at cultural material from al‐Ashmūnayn provides clear evidence of human 36.2–38.2 m + msl), with FIP and MK pottery from 38.5 m + msl up- activity during the OK period, particularly during the Fifth‐Sixth Dy- wards in borehole #7 and from 38.2 m + msl upwards in borehole #3. nasties (c. 2494–2181 BC). Boreholes #1–3 and potentially also #7 Boreholes #1 and #2 did not cover these chronological transitions produced thick sequences of up to c. 4 m, packed with pottery, that because post‐OK material had already been excavated. The gradual can be associated with this age, suggesting a focal area of human transition from the late OK period pottery into the early and later MK activity of this period roughly in the same zone as where MK deposits material, possibly without hiatuses, suggests that there may have been had been found earlier (and excavated) at higher levels (Roeder, 1959). continuous human presence (and likely occupation) at al‐Ashmūnayn To the north and northeast of the Amun temple of Seti II and near the during the FIP. This is an important observation considering the on- FIP cemetery (boreholes #4–6), the amount of OK finds diminished to going debate on the severity of cultural decline and depopulation in the an extent that did not allow a clear age assignment or to infer intensive Nile Valley following climatic instability at the end of the OK and early human usage of this area. FIP (cf. Butzer, 1984, 2012; Moeller, 2005; Moeller & Marouard, 2018). TOONEN | ET AL. 13 The current survey covered only a limited part of al‐Ashmūnayn; Boreholes #10 and #11 provide further understanding of the other sections are difficult to access due to modern settlement and configuration of the fluvial setting at al‐Ashmūnayn. Their overall the thick cultural deposits of the last four millennia. Therefore, it was fine‐grained deposits with some laminated loamy intervals link them neither possible to estimate the size or extent of the former OK town to Unit D or E, reflecting a distal part of the levee or proximal nor to pinpoint various functional areas based on the current in- floodplain setting (Figure 4). Following the reconstructed channel belt formation. Furthermore, the mix of cultural material consisted mainly architecture to the south of al‐Ashmūnayn, such deposits are ex- of pottery and charcoal fragments, while somewhat surprisingly, pected at either side of the main sandy channel belt. Moreover, basal common settlement waste material such as animal bone was missing. sands occur below c. 33 m + msl at either side of al‐Ashmūnayn, thus Nonetheless, the chronological continuity in the archaeological ma- matching the level of the regional substrate (Unit A). As channel terial suggests a prolonged concentration of human activities, al- sands (Unit B) are missing at boreholes #10 and #11, these locations though its exact nature and expanse are hard to interpret from the thus demarcate the westernmost and easternmost limits to the current data. possible location of the channel belt at al‐Ashmūnayn. To understand the environmental context of the cultural de- Assuming that the extensive cultural deposits reflect the pre- posits, the local borehole lithologies at and near al‐Ashmūnayn were sence of a nearby settlement, these results broadly confirm the hy- compared with the sedimentary units from the two transects. This pothesis that ancient Khemenu was founded on the banks of a river only applied to strata found in the lower parts of the boreholes channel (cf. Butzer, 1976). The deposits are, however, more char- placed on the tell, as the upper sections were heavily disturbed by acteristic for an active channel zone rather than its (finer‐grained) human activity. Sandy deposits found in borehole #2 resemble Unit B levees where human activity would have concentrated, and it seems (Figures 4 and 5), and are found at comparable levels: 37.5 m + msl at therefore unlikely that the encountered pottery fragments belong to al‐Ashmūnayn versus c. 36–37 m + msl in the transects to the south. in situ settlement strata. As almost no pottery fragment showed The laminated fines found in borehole #1 fit well with Unit C and the evidence of abrasion or erosion, the pottery does not seem to have sands further down again with Unit B. Units B and C were found to been exposed to transport in water and must have come from a intercalate at borehole #3, before reaching Unit B in its basal part. source very close to the borehole locations. Based on the similarity to deposits found in the transects, the environmental context of the investigated zone at al‐Ashmūnayn can be interpreted as channel deposits; the heterogeneous deposits with frequent layering between sands and silty clays relating to the infill of 5.2 | Development of the fluvial environment and links with the cultural landscape residual channel depressions, while homogeneous sand units, as found for example at Nectanebo I's Thoth temple, are more likely to Based on the borehole survey and information from the pottery represent bars in the channel zone. Although markedly different in analysis, it seems likely that a major channel of the Nile was active at appearance—representing the variability that is possible at a local al‐Ashmūnayn during the OK (Figure 9a). The MK gate of Amenemhat scale—both form integral parts of the same overall geomorphological II (c. 1911–1877 BC) and later monuments at al‐Ashmūnayn were setting. built directly on top of what used to be the active channel zone (Units F I G U R E 9 Palaeoenvironmental reconstruction of the Nile Valley landscape and hypothetical travel routes. (a) Old Kingdom setting with an active channel belt that is assumed to have connected al‐Ashmūnayn (1) with upstream al‐Shaykh Saʽīd (2). (b) The Middle Kingdom setting with an active channel belt located between Rairamun (3) and Dayr al‐Barshā (4). (c) Successive channel belts of the Nile in the eastern section of the valley (after Verstraeten et al., 2017) with periods of activity roughly between the New Kingdom and the Graeco‐Roman period and the last millennium [Color figure can be viewed at wileyonlinelibrary.com] 14 | TOONEN ET AL. B and C). Therefore, it can be concluded that the active Nile had configuration of geomorphological units (Smith et al., 1989). Low‐ largely shifted its position before the 19th century BC. relief floodplains of large river systems do often develop a convex When the channel was active, it must have run in close proximity topography due to the relatively high rates of sediment accumulation of the modern towns of Qulubba and Mallawī. These towns have in areas approximate to the main channel and lower rates of de- previously not been associated with occupation before the GRP and position in distal backswamp areas (Lewin & Ashworth, 2014). Byzantine period based on a study of their toponyms and isolated Eventually, this promotes the establishment of alternative flow paths surface finds (Kemp, 2005; Willems et al., 2017). Yet, based on their with a more favourable gradient compared to the existing one. The configuration alongside the OK channel route, these places could trigger for such changes to occur is generally a levee‐breaching flood. have been suitable settlement locations since the OK as the levees Larger floods and/or periods of enhanced flood activity with higher would have remained relatively high locations out of reach of annual rates of sediment accumulation would have increased the chances of Nile floods. Therefore, earlier occupation levels may have been pre- occurrence of an avulsion; the changes observed in the al‐ served at the basal levels of their high tells (Figure 9), coinciding with Ashmūnayn region coincide broadly with such a phase during the FIP the top levels of the channel sands and levee deposits (Units B and D; and early MK (Hassan et al., 2012; Macklin et al., 2015; Figure 4). Seidlmayer, 2001; Toonen et al., 2018). The modern urban sprawl of Mallawī did not allow us to trace the The time required to complete a full diversion of discharge to a channel belt in an upstream direction. Extending the channel belt in new channel route may be very short—for example several decennia the same general direction to upstream regions, however, does bring —but it may also be a very gradual process that takes several cen- it to the location of the late OK nomarchs' tombs at al‐Shaykh Saʽīd turies (Kleinhans et al., 2011). It is, therefore, possible that al‐ (Figure 9a). Albeit unconfirmed by actual observations, this situation Ashmūnayn temporarily retained a connection with the relocated would imply that apart from a cultural connection between al‐Shaykh main channel of the Nile through a small residual channel. The re- Saʽīd and Khemenu, they were also physically connected in the riv- sultant geomorphology of past fluvial activity would continue to offer erine landscape of the OK. The dated silting up of the channel, before advantages for cultural activity; the sandy deposits at al‐Ashmūnayn the 19th century BC, roughly corresponds with the shift of high‐ were suitable as foundations for later monuments, and the relatively status burials away from al‐Shaykh Saʽīd, suggesting that pivotal high levees of the channel belt would still be suitable for overland changes in the landscape influenced travel routes and the ritual travel (Figure 9b). The new course of the Nile and existing overland connection between sites in the region. routes may thus have made Dayr al‐Barshā an attractive place for The current survey also did not extend to the north of al‐ burials. The location and suggested pharaonic origin of the town of Ashmūnayn, so the downstream continuation of the OK river channel Rairamūn (Willems et al., 2017) fit the revised configuration of the remains unclear. Suspected OK sites on the western side of the natural and cultural landscape, as it would have been an ideal valley, such as Ḥūr and Balanṣūra (Willems et al., 2017) (Figure 1b: waystation between settlement and necropolis, and a suitable place 14, 15), could potentially mark a continuation of its north‐western to embark and cross the river (Figure 9b). direction. Yet, this suggestion remains largely speculative without further detailed geoarchaeological study. For the last two millennia, it has been amply demonstrated by 6 | CONCLUSIONS archaeological finds, from textual sources and historical maps (Willems et al., 2017) and by geological surveys near Dayr al‐Barshā The main goals of this study were to establish the presence of Old (Verstraeten et al., 2017), that the Nile was located at the eastern Kingdom settlement remains at al‐Ashmūnayn, to reconstruct the limit of the valley (Figure 9c). A few pottery fragments dated to the local environmental setting and to compare changes in the regional NK found in channel deposits near Dayr al‐Barshā also suggest an landscape with changes in the configuration of cultural sites that are earlier presence of the Nile in the eastern part of the floodplain associated with al‐Ashmūnayn. Diagnostic pottery fragments and (Verstraeten et al., 2017). The tomb of Djehutihotep (c. 1900 BC) at the wider assemblage provided evidence for continuous human Dayr al‐Barshā potentially provides even earlier evidence, as the activity at the site of al‐Ashmūnayn since at least the late Old colossal monolithic statue depicted on one of its murals is mentioned Kingdom. The densely packed pottery fragments occurred in a se- to overlook a channel of the Nile (Willems et al., 2005). This early diment matrix of what was interpreted as river channel deposits. date for channel activity fits well with the suggested date of channel Based on the minimal abrasion of pottery fragments and their sheer abandonment at al‐Ashmūnayn (Figure 9b). quantity, it seems likely that Old Kingdom settlement activities were The sedimentary archive supports a fairly rapid relocation of the main channel of the Nile, but cannot be explained using Butzer's located very close to the position of the boreholes, likely on local river banks. (1976) eastern migration premise. Both borehole transects are Previous claims of al‐Ashmūnayn's origin at the banks of an ac- dominated to the east by fine‐grained deposits that accumulated as tive channel can thus be considered valid. The channel sands found at overbank material during floods. In the absence of sandy channel the site occur at similar depths and in a roughly similar configuration deposits, a gradual eastward shift by lateral channel migration can be as off‐site research locations. In two transects, all geomorphologic ruled out. Instead, the process of avulsion fits the observed elements of a channel belt (i.e., channel sands, levees, residual TOONEN | ET AL. 15 channel fills) and adjacent floodplains were identified based on the RE F ER EN CES architecture of lithological units. The 800–1000 m width of the Anthes, R. (1928). Die Felseninschriften von Hatnub. Untersuchungen zur Geschichte und Altertumskunde Aegyptens 9. Hinrichs, Leipzig. Aston, D. (2004). Tell el‐Dab'a 12: A Corpus of late Middle Kingdom and Second Intermediate Period Pottery. Verlag der Österreichischen Akademie der Wissenschaften, Denkschriften der Gesamtakademie 28. Vienna. Attia, M. I. (1954). Deposits in the Nile valley and delta. Geological Survey of Egypt, Cairo. Bader, B. (2001). Tell el‐Dab'a 13: Typologie und Chronologie der Mergel C‐Ton Keramik: Materialien zum Binnenhandel des Mittleren Reiches und der Zweiten Zwischenzeit. Verlag der Österreichischen Akademie der Wissenschaften, Denkschriften der Gesamtakademie 22. Vienna. Bader, B. (2002). A concise guide to Marl C‐pottery. Ägypten und Levante, 12, 29–54. Bader, B. (2009). The Late Old Kingdom in Herakleopolis Magna? An interim interpretation. In T. Rzeuska & A. Wodzińska (Eds.), Studies on Old Kingdom pottery (pp. 13–41). Neriton. Bourriau, J. (2006). Technology in the pottery of the Middle and New Kingdoms: An underrated tool in the archaeologist's armory. In B. Mathieu, D. Meeks, & M. Wissa (Eds.), L'apport de l'Égypte à l'histoire des technique (pp. 31–41). Institut Français d'Archéologie Orientale du Caire. Bridge, J. S. (2003). Rivers and floodplains: Forms, processes, and sedimentary record. Blackwell Science. Bunbury, J., & Malouta, M. (2012). The geology and papyrology of Hermopolis and Antinoopolis. eTopoi, 3, 119–122. Bunbury, J., Tavares, A., Pennington, B., & Gonçalves, P. (2017). Development of the Memphite floodplain. In H. Willems & J. Dahms (Eds.), The Nile: Natural and cultural landscape in Egypt. Mainz Historical Cultural Sciences 36 (pp. 71–96). Transcript. Butzer, K. W. (1976). Early hydraulic civilization in Egypt: A study in cultural ecology. University of Chicago Press. Butzer, K. W. (1984). Long‐term Nile flood variation and political discontinuities in Pharaonic Egypt. In Clark, J. D. & Brandt, S. A. (Eds.), From hunters to farmers (pp. 102–112). University of California Press. Butzer, K. W. (1998). Late Quaternary problems of the Egyptian Nile: Stratigraphy, environments, prehistory. Paléorient, 23, 151–173. Butzer, K. W. (2012). Collapse, environment, and society. Proceedings of the National Academy of Sciences of the United States of America, 109, 3632–3639. Chen, Z., & Stanley, J. D. (1993). Alluvial stiff muds (Late Pleistocene) underlying the lower Nile Delta plain, Egypt: Petrology, stratigraphy and origin. Journal of Coastal Research, 9(2), 539–579. Czerny, E. (1999). Tell el‐Dab'a 9: Eine Plansiedlung des frühen Mittleren Reiches. Verlag der Österreichischen Akademie der Wissenschaften, Denkschriften der Gesamtakademie 16. Vienna. Davies, N. de G. (1901). The rock tombs of Sheikh Saïd. Archaeological survey of Egypt 10. Egypt Exploration Fund. De Meyer, M (2008). Old Kingdom rock tombs in Dayr al‐Barshā. Archaeological and textual evidence of their use and reuse in Zones 4 and 7 (PhD thesis). Leuven University. Faltings, D. (1998). Die Keramik der Lebensmittelproduktion im Alten Reich. Ikonographie und Archäologie eines Gebrauchsartikels, Studien zur Archäologie und Geschichte Altägyptens 14. Heidelberger Orientverlag. Gardiner, A. H. (1947). Ancient Egyptian Onomastica II. Oxford University Press. Ginau, A., Schiestl, R., & Wunderlich, J. (2019). Integrative geoarchaeological research on settlement patterns in the dynamic landscape of the northwestern Nile Delta. Quaternary International, 511, 51–67. Google Earth. (2020, April 27). al‐Ashmūnayn, Egypt. CNES/Airbus 2012. http://www.earth.google.com channel belt, forming a maximum estimate for the contemporary active channel, is comparable to that of the main active channel of the (present) Nile. The abandonment of the river channel at al‐Ashmūnayn is dated to have happened at some time between the late Old Kingdom and the early Middle Kingdom. This range is constrained by the lithological context of late Old Kingdom pottery in an infilling residual channel and the age of monuments that were founded on top of the palaeo‐channel sands, which implies that there was no imminent threat of fluvial erosion or flooding at the time of their construction. From our regional survey, it can be concluded that the relocation of the active channel was relatively fast and forced by an upstream avulsion. This interpretation is at odds with traditional models that infer a gradual lateral eastward river channel migration in the Egyptian Nile Valley. A major implication of an avulsive migration model is that substantial areas across the valley that were previously thought to have been disturbed by fluvial erosion can still have in situ archaeological remains preserved. The revised configuration of the past landscape equally sheds new light on the potential ancient origins of several towns in the region based on favourable settlement locations along fluvial routes in the past. The rapid transition of the fluvial environment is likely to have triggered substantial changes in the cultural and ritual landscape, as contemporary shifts in preferred burial locations at satellite locations of al‐Ashmūnayn are observed. Such coincidental dynamics may indicate that the ancient Egyptian riverine civilization was closely connected to the fluvial environment and not only susceptible to large‐scale climate‐driven changes in hydrology but also to the dynamics of the regional fluvial landscape. A C KN O W L E D G E ME N T S All data are included with the current publication as supporting information. This study was carried out as part of the KU Leuven/NVIC mission to Dayr al‐Barshā, and was funded by an F + fellowship (F+/14/16) to H. W. and W. H. J. T. and an FWO‐fellowship (1177218N) to K. C. We express our gratitude to the Permanent Committee of the SCA for permitting us to work at the site, and specifically to Dr. Khālid al‐Inānī, Dr. Muṣṭafā Wazīrī and Dr Muḥammad Ismaʽīl, Mrs. Nashwa Jābir, Mr. Jamāl Abū Bakr, Mr. Maḥmūd Ṣalāḥ, Mr. ʽĀlī al‐Bakrī, Mr. Fatḥī ʽAwwad Rīyāḍ and Mr. Ḥamāda Muḥammad ‘Abd al‐Mu‘īn Kallāwy. We sincerely thank the inspectors of the Egyptian Ministry for Antiquities (ʽAbd Allāh ʽĀlī ʽAbd al‐ Raḥmān Maʽrūf, Sayyid ʿAbd al‐ʿAzīz al‐Jahmī, Rāmī Isḥāq and Marwā ʿUthmān), and the local workmen and landowners for making this study possible. Detailed drawings of diagnostic pottery were made by Dorien Ceyssens, Nina Troosters and Merel Eyckerman (Hasselt University/PXL‐MAD, Hasselt, Belgium). David Aston is thanked for sharing his expertise on New Kingdom to Ptolemaic and Late Antique pottery. Two anonymous reviewers and editor Dr. Sarah Sherwood are thanked for their supportive comments. 16 | Hassan, F. A., Hamdan, M. A., Flower, R. J., & Keatings, K. (2012). The oxygen and carbon isotopic records in Holocene freshwater mollusc shells from the Faiyum paleolakes, Egypt: Their environmental and paleoclimatic implications. Quaternary International, 266, 175–187. Jacotin, M. (1821). Carte Topographique de l'Égypte et de plusieurs parties des pays limitrophes; levée pendant l'Expédition de l'Armée Française par les Ingénieurs—Géographes, les Officiers du Génie Militaire et les Ingénieurs des Ponts et Chaussées; assujettie aux Observations des astronomes, Construite par M. Jacotin. Gravées au depot Général de la Guerre, à l'Echelle de 1 Millimètre pour 100 Mètres. Publiée par ordre du Gouvernement. Jacquet‐Gordon, H. (1981). A tentative typology of Egyptian bread moulds. In D. Arnold (Ed.), Studien zur altägyptischen Keramik (pp. 11–24). Philipp von Zabern. Kemp, B. (2005). Settlement and landscape in the Amarna Area in the Late Roman period. In J. Faiers (Ed.), Late Roman pottery at Amarna and related studies. Egyptian Exploration Society Excavation Memoirs 72 (pp. 11–56). Egyptian Exploration Society. Kessler, D. (1977). Hermupolis Magna. In W. Helck & E. Otto (Eds.), Lexikon der Ägyptologie II (pp. 1137–1147). Harrassowitz. Kessler, D. (1981). Historische Topographie der Region zwischen Mallawi und Samalut. Harrassowitz. Klappa, C. F. (1980). Rhizoliths in terrestrial carbonates: Classification, recognition, genesis and significance. Sedimentology, 27, 613–629. Kleinhans, M. G., Cohen, K. M., Hoekstra, J., & IJmker, J. M. (2011). Evolution of a bifurcation in a meandering river with adjustable channel widths, Rhine delta apex, the Netherlands. Earth Surface Processes and Landforms, 36, 2011–2027. Kytnarová, K. (2009). Pottery of the Fifth and Sixth Dynasty from the excavations of the Czech Institute of Egyptology in the cemeteries of Abusir South (PhD thesis). Univerzita Karlova v Praze, Prague. Lehner, M. (2014). On the waterfront: Canals and harbors in the time of Giza Pyramid‐building. Aeragram, 15, 14–23. Lewin, J., & Ashworth, P. J. (2014). The negative relief of large floodplains. Earth‐Science Reviews, 129, 1–23. Macklin, M. G., Toonen, W. H. J., Woodward, J. C., Williams, M. A. J., Flaux, C., Marriner, N., Nicoll, K., Verstraeten, G., Spencer, N., & Welsby, D. (2015). A new model of river dynamics, hydroclimatic change and human settlement in the Nile valley derived from meta‐ analysis of the Holocene fluvial archive. Quaternary Science Reviews, 130, 109–123. Miall, A. D. (1985). Architectural‐element analysis: A new method of facies analysis applied to fluvial deposits. Earth‐Science Reviews, 22, 261–308. Moeller, N. (2005). The first intermediate period: A time of famine and climate change? Ägypten und Levante, 15, 153–167. Moeller, N., & Marouard, G. (2018). The development of two early urban centres in upper Egypt during the 3rd Millennium BC: The examples of Edfu and Dendara. In J. Budka & J. Auenmüller (Eds.), From microcosm to macrocosm: Individual households and cities in ancient Egypt and Nubia (pp. 29–58). Sidestone. Nordström, H. Å., & Bourriau, J. D. (1993). Ceramic technology: Clays and fabrics. In Arnold, D. & Bourriau, J. D. (Eds.), An introduction to ancient Egyptian pottery. Sonderschrift des Deutsches Archäologisches Institut, Abteilung Kairo: Sonderschriften 17 (pp. 140–186). Philipp Von Zabern. Op de Beeck, A. (2004). Possibilities and restrictions for the use of Maidum‐bowls as chronological indicators. Cahiers de la Céramique Égyptienne, 7, 239–274. Ownby, M. F. (2009). Petrographic and chemical analyses of select 4th dynasty pottery fabrics from the Giza Plateau. In T. Rzeuska & A. Wodzińska (Eds.), Studies on Old Kingdom pottery (pp. 113–137). Neriton. TOONEN ET AL. Pennington, B. T., Wilson, P., Sturt, F., & Brown, A. G. (2021). Landscape change in the Nile Delta during the fourth millennium BC: a new perspective on the Egyptian Predynastic and Protodynastic Periods. World Archaeology, 52, 550–565. https://doi.org/10.1080/ 00438243.2020.1864463 Roeder, G. (1959). Hermopolis 1929‐1939. Ausgrabungen der Deutschen Hermopolis‐Expedition in Hermopolis, Ober‐Ägypten. Gebr. Gerstenberg, Hildesheim. Rzeuska, T. I. (2006). Pottery of the late Old Kingdom: Funerary pottery and burial customs: Saqqara 2. Neriton. Said, R. (1993). The river Nile: Geology, hydrology and utilization. Pergamon Press. Schiestl, R., & Seiler, A. (2012). Handbook of the pottery of the Egyptian Middle Kingdom I: The corpus volume. Österreichische Akademie der Wissenschaften. Denkschriften der Gesamtakademie 72. Contributions to the Chronology of the Eastern Mediterranean 31. Verlag der Österreichischen Akademie der Wissenschaften. Seidlmayer, S. J. (2001). Historische und moderne Nilstände: Untersuchungen zu den Pegelablesungen des Nils from der Frühzeit bis in die Gegenwart. Achet A1. Achet Verlag. Shaw, I. (2000). The Oxford history of ancient Egypt. Oxford University Press. Smith, N. D., Cross, T. A., Dufficy, J. P., & Clough, S. R. (1989). Anatomy of an avulsion. Sedimentology, 36, 1–23. Spencer, A. J. (1993). Excavations at el‐Ashmunein III. The Town. British Museum. Spencer, N., Macklin, M. G., & Woodward, J. C. (2012). Re‐assessing the abandonment of Amara West: The impact of a changing Nile? Sudan and Nubia, 16, 37–43. Steckeweh, H. (1937). B. Die Ausgrabung. In H. Roeder, A. Hermann, H. Kortenbeutel, H. Steckeweh, & J. Werner (Eds.), Bericht über die Ausgrabungen der Deutschen Hermopolis‐Expedition 1935. Mitteilungen des Deutschen Instituts für Ägyptische Altertumskunde in Kairo 7 (pp. 10–16). Reichsverlagsambt. Tallet, P., & Marouard, G. (2014). The harbor of Khufu on the Red Sea coast at Wadi al‐Jarf, Egypt. Near Eastern Archaeology, 77, 4–14. Toonen, W. H. J., van Asselen, S., Stouthamer, E., & Smith, N. D. (2015). Depositional development of the Muskeg Lake crevasse splay in the Cumberland Marshes, Canada. Earth Surface Processes and Landforms, 41, 117–129. Toonen, W. H. J., Graham, A., Masson‐Berghoff, A., Peeters, J., Winkels, T. G., Pennington, B. T., Hunter, M. A., Strutt, K. D., Barker, D. S., Emery, V. L., Sollars, L., & Sourouzian, H. (2019). Amenhotep III's mansion of millions of years in Thebes (Luxor, Egypt): Submergence of high grounds by river floods and Nile sediments. Journal of Archaeological Science: Reports, 25, 195–205. Toonen, W. H. J., Graham, A., Pennington, B. T., Hunter, M. A., Strutt, K. D., Barker, D. S., Masson‐Berghoff, A., & Emery, V. L. (2018). Holocene fluvial history of the Nile's west bank at ancient Thebes, Luxor, Egypt, and its relation with cultural dynamics and basin‐wide hydroclimatic variability. Geoarchaeology, 33, 273–290. Toonen, W. H. J., Kleinhans, M. G., & Cohen, K. M. (2012). Sedimentary architecture of abandoned channel fills. Earth Surface Processes and Landforms, 37, 459–472. USDA. (2017). Soil survey manual. In C. Ditzler, K. Scheffe, & H. C. Monger (Eds.), USDA handbook 18. Soil Science Division Staff, Government Printing Office. Vanthuyne, B. (2017). Early Old Kingdom rock circle cemeteries in the 15th and 16th nomes of Upper Egypt. A socio‐archaeological investigation of the cemeteries in Dayr al‐Barshā, Dayr Abū Ḥinnis, Banī Ḥasan al‐Shurūq and Nuwayrāt (PhD thesis). Leuven University. Verstraeten, G., Mohamed, I., Notebaert, B. & Willems, H. (2017). The dynamic nature of the transition from the Nile floodplain TOONEN | ET AL. to the desert in Central Egypt since the Mid‐Holocene. In H. Willems & J. Dahms (Eds.), The Nile: Natural and cultural landscape in Egypt. Mainz Historical Cultural Sciences 36 (pp. 239–254). Transcript. Willems, H. (2013). A note on the ancient name of Dayr al‐Barshā. Zeitschrift für Ägyptische Sprache und Altertumskunde, 140, 188–192. Willems, H., Creylman, H., De Laet, V., & Verstraeten, G. (2017). The analysis of historical maps as an avenue to the interpretation of pre‐ industrial irrigation practices in Egypt. In H. Willems & J. Dahms (Eds.), The Nile: Natural and cultural landscape in Egypt. Mainz Historical Cultural Sciences 36 (pp. 255–344). Transcript. Willems, H., Peeters, C., & Verstraeten, G. (2005). Where did Djehutihotep erect his colossal statue? Zeitschrift für ägyptische Sprache und Altertumskunde, 132, 173–189. Willems, H., Vereecken, S., Kuijper, L., Vanthuyne, B., Marinova, E., Linseele, V., Verstraeten, G., Hendrickx, S., Eyckerman, M., Van den Broeck, A., Van Neer, W., Bourriau, J., French, P., Peeters, C., De Laet, V., Mortier, S., & De Kooning, Z. (2009). An industrial site at al‐Shaykh Sa‛īd/Wādī Zabayda. Ägypten & Levante, 19, 293–331. 17 Zibelius, K. (1978). Ägyptische Siedlungen nach Texten des Alten Reiches. Beihefte zum Tübinger Atlas des Vorderen Orients, Reihe B (Geisteswisserschaften, Nr 19). Wiesbaden. Zivie‐Coche, C. (2009). L'Ogdoade à Thèbes à l'époque Ptolémaïque et ses antécédents. In C. Thiers (Ed.), Documents de theologies Thebaines Tardives (DAT 1) (pp. 167–255). Montpellier. SUPP ORTING INFO RM ATION Additional supporting information may be found in the online version of the article at the publisher’s website. How to cite this article: Toonen, W. H. J., Cortebeeck, K., Hendrickx, S., Bader, B., Peeters, J., & Willems, H. (2021). The hydro‐geomorphological setting of the Old Kingdom town of al‐Ashmūnayn in the Egyptian Nile Valley. Geoarchaeology, 1–17. https://doi.org/10.1002/gea.21894