Keyword

1 Introduction

Geophysical research in Egypt has been dominated by magnetometry, by far the most popular of the commonly applied methods. Electrical resistivity and other electromagnetic method, including ground-penetrating radar (GPR) and electromagnetic induction (EMI), being used only sporadically. This preference derives, inter alia, from the nature of the architectural building material contrasted with the geological background. The sun-dried mud brick, which was the main material in use throughout Egypt, in the Nile River valley and Delta as well as on the desert fringes, was made of Nile silt, which has a high iron oxide content giving it a high magnetic susceptibility (Hesse, 1967). In the case of archaeological sites on the desert fringes of the Nile valley, the sun-dried bricks are in a high contrast with the diamagnetic quartz sand matrix, and this was the reason of successful tracing of mud structures even when using instruments of low resolution (like proton magnetometers of ±1 nT resolution, Herbich, 2015). Experience gained over the past quarter of a century—when using the fluxgate and caesium instruments of higher resolution (in range of ±0.1–0.01 nT) has broadened the base for tracing mud-brick structures in the alluvial matrix of the river valley and Delta (Herbich, 2003).

The sporadic use of electrical resistivity method is due to their low time-effectiveness when collecting data in a sandy matrix characterised by very low humidity. Outcomes are successful only in very specific conditions: when stone-made structures are embedded in a humid alluvial matrix giving a high contrast between the resistivity of the structure and that of the alluvium. It should be kept in mind that because stone building material was always at a premium (including limestone specifically for burning lime), ancient sites always used to be a ready source, hence so few preserve elements of stone architecture.

As for GPR surveys, the logistics of its use in Egypt have always been problematic, as well as other adverse effects related to the type of geology and overall substrate. The shallow water table—barely a meter below the surface as a rule—on sites with an alluvial base and intensive irrigation practices, discourages the use of this method. Its effectiveness has been proved on a few sites, mainly in the Sudanese part of the Nile valley, where it has demonstrated great potential for detecting mud-brick structures (Obłuski et al., 2021). It has also shown success in revealing settlements in alluvial areas with a low ground water table (Ullrich & Wolf, 2015).

Geophysical work done in Egypt can be divided into two periods with the turnover occurring in the mid-1990s. The change is observed primarily when looking at manner to conduct magnetometer surveys. It is due to several factors, the most important: the resolution and measurement speed, both of which are directly related to apparatus development; enhanced data processing methods; and a growing body of experience from different kinds of archaeological sites located in regions with a different geological and pedological characteristics. Archaeological verification of geophysical data has also allowed to test the results and interpretation provided by such geophysical surveys, building a store of knowledge about different applications. The border date is 1996 when instruments automatically recording measurements were first used in Egypt, alongside software for digital map processing (Fassbinder et al., 1999; Becker & Fassbinder, 1999; Abdallatif et al., 2003). A similar phasing also applies to the use of the electrical resistivity.

2 Period I (1973–1996)

Malkata near Luxor was the first archaeological site to be tested in Egypt (Fig. 1). In February and March 1973, Elizabeth Ralph from the Museum of the University of Pennsylvania, a pioneer in the use of caesium magnetometers to survey archaeological sites, carried out research at the site of the New KingdomFootnote 1 harbour and settlement (Ralph, 1973). The measurements were made in a differential mode, meaning that the differences between the base magnetometer and the instrument moved along the traverses were noted and a total area of ~11 haFootnote 2 was covered. Despite a rather loose sampling grid (2 m traverse interval) the survey traced the remains of mud-brick structures within the settlement, but the results were not sufficient to reconstruct the plan of the port expected in the area.

Fig. 1
A satellite map of Egypt with the locations of archaeological sites marked. Some of the locations include Malkata, Berenike, Amarna, Giza, Abusir, Saqqara, Mendes, Sais, Buto, Marea, and Hierakonpolis.

Location of sites mentioned in the paper marked on a satellite image of Egypt. (Google)

Full size image

Proton magnetometers needed about 6 s to take a measurement with ±1 nT resolution, hence with one measurement per square meter it was not possible to cover more than 1000–1500 m2 in a day’s work. Most of these surveys were implemented over desert fringes of the Nile Valley, in Abusir, Giza and Saqqara, where the high contrast between the magnetic susceptibility of mud-brick structures and the matrix guaranteed the recording of structures invisible on the surface. At Giza, measurements were taken every 5–10 m in order to confirm the presence of structures without even thinking of tracing the layout (Dolphin et al., 1977). GPR and electrical resistivity surveys were also used at Giza within the framework of the same project. It was the first time that these geophysical methods were applied in Egypt and the objective was to trace stone structures around the pyramids and burial chambers (Dolphin et al., 1975).

The study carried out by Vladimir Hašek at Abusir for the Charles University (Prague, Czech Republic) expedition have been unduly forgotten and this study has not been cited neither in Egyptological nor in archaeo-geophysical literature. Fieldwork was preceded by a thorough laboratory examination of the magnetic and electrical properties of building materials used in mortuary architecture. Several magnetic anomalies were identified and interpreted as Late Period funerary shafts cut in bedrock with mud bricks lining their upper parts. The results of the surveys also served to draw a plan of the mortuary temple in front of the Raneferf pyramid. Verifying excavations in the following seasons fully supported Hašek’s interpretations (Hašek & Verner, 1981; Herbich, 2024, pp. 387–390).

The next prospection with a proton magnetometer was carried out by the University of Warsaw at Saqqara, west of Djeser’s Step Pyramid, in an area covered with sand bearing no evidence of human activity on the surface (Myśliwiec & Herbich, 1995). The surveys covered an area of ~1.5 ha, using two proton magnetometers in a differential mode over a one square meter grid. A few areas of anomalous values were identified and interpreted as clusters of mud bricks. The results were used to plan excavations. The largest anomaly turned out to be a tumble of mud bricks concealing the entrance to a funerary chapel cut in the rock of an unknown vizier named Meref-nebef from the reign of the late Old Kingdom king Teti (Herbich, 2003, pp. 16–18).

Among other studies using proton magnetometers, including Egyptian geophysicists (see Hussain, 1983), a survey carried out at Mendes, a site in the Nile Delta, stands out given the week magnetic contrast between subsurface mud-brick structures and the alluvial deposits. A square metre sampling grid was used with a measurement resolution of ±1 nT. The survey was effective in locating furnaces for pottery production, which were characterised by anomalies in the range of 50–100 nT due to additional thermoremanent magnetisation, but it failed to resolve remains of settlement architecture (Pavlish, 2004, pp. 87–100). The scarce contrast between mud-brick structures and the surrounding matrix, typical of sites in the Nile Delta, weighed heavily on future of magnetometer prospection in Egypt.

In the mid-80s, there were some sporadic surveys were carried out mainly in the desert fringes of the Nile Valley where a high contrast between the values of magnetic susceptibility of mud-brick structures and the surrounding matrix was expected (e.g., Mathieson, 1995). The lack of conviction continued until early 1990s, at a time when high-resolution instruments (between ±0.1 and ± 0.01 nT) started being used in Europe. These instruments allowed short measurement time, denser sampling, and larger-areas coverage.

Electrical resistivity surveys were also sporadically used during this period. Three surveys have been noted in publications: Abusir, Amarna and Tell Atrib in the Nile Delta. In Abusir, electrical resistivity and magnetometer surveys were used simultaneously to obtain complementary information to determine which building material was used (stone or mud-brick; Hašek & Verner, 1981). The survey in Amarna helped to determine the original extent of the desert, now covered by Nile deposits and the location of wells inside the city (Mathieson, 1989). An architectural complex of mud and red brick material was discovered at Tell Atrib (Myśliwiec & Herbich, 1988; Myśliwiec, 2013).

3 Period II (After 1996)

New caesium and fluxgate magnetometers changed the attitude towards magnetometry in Europe in the 1980s, demonstrating the effectiveness of the method in tracing archaeological features (Herbich, 2015). Fluxgate gradiometers, produced then chiefly by Geoscan Research, measured the vertical gradient of the vertical component of the Earth’s magnetic field (Gaffney & Gater, 2003, pp. 61–64), whereas the caesium system (by Scintrex) applied in uncompensated so-called duo–sensor configuration, measured the variation of the total Earth’s magnetic field along two traverses (Becker, 1999). A team from the Bavarian State Department for Monuments and Sites in Munich, the Polish Academy of Sciences and National Research Institute of Astronomy and Geophysics (NIARG) in Helwan, Egypt, used the two types of instruments to survey the Polish concession in West Saqqara (Fassbinder et al., 1999; Abdallatif et al., 2019, pp. 149–150). The results clearly revealed square magnetic anomalies that typically corresponded to the response of mud-brick walls surrounding burial shafts in this area. The clarity of the image was the result of a dense sampling rate, with spacing between traverses equal to 0.5 m, in-line sampling 0.25 m for the fluxgate and about 0.10 m for the caesium instrument. The data were visualised as greyscale maps (256 shades, using Geoplot software).

Surveys in Dakhleh Oasis and the Nile Delta opened new perspectives for the use of these instruments. Fieldwork at the Old Kingdom settlement of ‘Ain al-Gazareen proved the great potential of magnetometry to trace mud-brick structures made of local silts devoid of the magnetic properties (Fig. 2). A 10 times higher measurement resolution compared to the proton magnetometers and a denser sampling rate (four measurements per square meter) enabled these results. Digital greyscale map analysis also helped to image the shape of structures differing by a fraction of nanotesla. Buried structures showed up as negative anomalies (features with low magnetic values in a matrix characterised by higher values). Higher values for the layers constituting the matrix were due to concentrations of ash and organic remains typical of cultural occupational deposits (Herbich & Smekalova, 2001).Footnote 3

Fig. 2
An aerial map and a layout map of the settlement. A. A map in gradient shades for magnetometer survey, with a few patches of dark shades and a highlighted excavation area. B. A plan with a horizontal scale of 0 to 20 meters. It includes mud-brick and reconstructed walls and fireplaces, etcetera.

‘Ain al-Gazareen. Left: results of the magnetometer survey of the eastern part of the settlement. Right: map of the settlement after excavation (after Mills, 2012, p. 178). 1—mud-brick walls; 2—walls reconstructed based on the survey; 3—fireplaces, hearths, kilns, ovens, ash dumps

Full size image

Instruments with higher sensitivity, like the caesium system applied in Saqqara, were tested by the expedition of the Pelizaeus Museum in Hildesheim at the site of Qantir in the Nile Delta, the late New Kingdom capital of Egypt Pi-Ramesse. The instrument proved to be effective in mapping mud-brick architecture on predominantly alluvial site (Becker & Fassbinder, 1999). The surveys carried out in 1996–2003 covered an area of more than 2 km2, mapping the city and distinguishing districts with chiefly palace architecture, residential, storage and religious areas, including the waterfront on the Pelusiac branch of the Nile (Pusch & Becker, 2017). The clarity of the mapping of mud-brick structures, in this case, usually as positive magnetic anomalies (characterised by values higher than those measured for the surroundings) is due, in part, to the fact that towns in the eastern Delta were frequently founded on sandy Pleistocene geziras not covered by Nile silt deposits in the Holocene (Sampsell, 2003, pp. 118–120). The lower magnetic susceptibility of the sand matrix provided a sufficient contrast to facilitate the identification of the archaeological remains as positive magnetic anomalies.

The successful outcome of the prospection in Qantir fomented interest in carrying out research on other sites in the Nile Delta. In 1999, the neighbouring site of Tell el-Dabca (ancient Avaris, capital of Egypt in the Second Intermediate Period) started to be explored. More than 10 years of research at this site brought many significant discoveries, not the least a palatial complex from the Hyksos period, a New Kingdom royal palace, the dimensions and plan of which were verified, fortifications from the times of Horemheb, domestic architecture from the Middle and New Kingdoms. Magnetometer survey results verified the course of the Pelusiac branch of the Nile and the extent of the floodplain provisionally traced as a result of a study of site geomorphology (Forstner-Müller, 2009; Herbich, 2024).

Both Pi-Ramesse and Avaris are located today on intensively irrigated agricultural land, levelled from the beginning of the nineteenth century to increase the area of arable land; the levelling resulted in the removal of later occupational strata. Before the use of geophysical surveys to locate and map archaeological sites, potsherds concentrations were used as a proxy for the location of archaeological sites because of the plow pulling to the surface ceramics and other remains. Uncultivated sites in the Nile Delta are man-made mounds (“tells” or “koms” in Arabic) formed of accumulated cultural layers. These can be quite well preserved and be up to a few dozen meters high. Initially, their high/topography and the general low contrast between the magnetic properties related their buried archaeology and surrounding matrix discourage the implementation of magnetometer surveys. However, after the success of prospection in Qantir and Tell el-Dabca, these sites were also tested.

First on the list was Buto (of Predynastic and Old Kingdom date in the lower parts and Late Period and Ptolemaic-Roman in the upper ones) in the north-western Delta. In 1999, the surveys gave a distinct image of the subsurface structures related to the later phases of the site. Characteristic tower houses from this period, with square plans about 10–15 m to the side, had wide foundations, no less than 2 m wide, which undoubtedly aided in their perfectly clear mapping and period determination. Prospection in the following seasons produced images of structures from different periods which were verified by archaeological excavations: an exceptionally clear picture of a Roman-age pottery production centre (more than 50 furnaces) overlying a deserted Late Period and Ptolemaic settlement (Hartung et al., 2003). Earlier deposits (Early Dynastic period) were not reached as these were far too deep.

In 1999, fieldwork began at the Red Sea harbour site of Berenike. The site is located at an area of limestone and sandstone deposits (originated by erosion processes of the adjacent Red Sea Mountains) overlaying magmatic granite rocks. The main building material, as attested in excavations, was fossil coral reef chunks and blocks of gypsum/anhydrite, neither of which have diamagnetic or week ferromagnetic properties. Effective mapping was possible thanks to the higher magnetic susceptibility of the deposits in which these remains were embedded, and the resolution of the instrument used (at least ±0.1 nT) as the contrast difference was quite subtle (Herbich, 2007).

These projects and achievements reflect the new perspectives for geophysics in Egypt that appeared in the last 5 years of the twentieth century. Geophysical prospection, mainly using magnetometry, turned out to be an effective tool for recording archaeological structures not visible on the surface on all categories of sites in Egypt regardless of region, whether in the Nile Valley or Delta, the Mediterranean and Red Sea coasts, the deserts, and desert oases. A growing awareness of the new possibilities resulted in a considerable intensification of research.

The examples presented below concern sites primarily from the Nile Delta showing the application of geophysical methods in areas with predominantly Nile alluvium deposits. The presentation follows a division of sites by functional categories, matching a recent publication of surveys at desert sites (Herbich, 2019). Examples of landscape reconstruction around ancient sites are discussed in the last section.

3.1 Cities and Villages

Considering the results of the past 25 years of research in Egypt, it is evident that geophysics in general and magnetometry, in particular, have provided researchers with an effective tool for the study of ancient urban layout. Measurements of large areas in short periods of time opened the way to studies of cities and villages with an area from a few to a few hundred hectares. Before that, few cities were actually excavated over larger areas, and even then, rather in the sacral and palatial districts. This enabled various extreme hypotheses to be formulated, like the one that unlike Mesopotamia, Egypt did not develop cities as such (Wilson, 1960).

Prospection at urban sites had the objective of establishing the layout of given districts and, in a few cases, of the settlement as a whole. Of the sites where just districts were surveyed, one should mention Saïs, where sections of the city with buildings demonstrating plans typical of the Late Period were observed (Wilson, 2006, pp. 151–175), and Kom Firin, where the full plan of a large late New Kingdom-period walled complex centred around the temple was established (Spencer, 2014, Fig. 4, pp. 17–46). Regarding tell sites, measurements gave a reasonable image of a small settlement at Kom el-Gir, tracing plans of individual buildings, a network of streets over an area of 13 hectares and beside that, the first Roman fort to be known from the Delta (Schiestl, 2016). In the case of larger tells, either close to or exceeding 1 km2 in area such as Buto and Tanais, considerable parts of the cities were reconstructed, and urban layout analyses led to establish specific functions areas and even a chronology of development.

At Buto (Tell el-Fara’in), an important Pre- and Early Dynastic centre, abandoned around 2200 BC and reoccupied in the eighth century BC through early Islamic times, the objective of the surveys initiated by the German Archaeological Institute in Cairo in 1999, was to trace the remains of the first phase of occupation (Hartung, 2018). The structures did not show up well on the magnetometer survey results due to: the slightness of the contrast between the magnetic susceptibility of the walls and the surroundings; the unimpressive width of walls; and the depth at which the remains were found (usually more than 1 m). However, the later occupation phases were imaged perfectly and became a solid base for a program of excavations of the later phase of city occupation, especially the special role of a major pottery production centre from Ptolemaic and Roman times (Hartung et al., 2003, pp. 263–266; Ballet, 2018; Ballet et al., 2019). An area of 25.5 ha was surveyed in the western and northern parts of the site, recording a few dozen buildings on either side of a N–S street (Fig. 3). The buildings revealed typical casemate plans characteristic of the Late and Ptolemaic-Roman periods, with thick walls arranged in squares and a series of rooms observed at foundation level (see photograph in Fig. 3). The form of these multi-storied houses is known from iconographic representations and architectural models (Marouard, 2012). Analogous structures excavated in the city district dated to Ptolemaic times prove the continuity of building traditions independent of changing power models. The remains of two enclosures were traced, polygonal on the northeast, where walls were up to 5 m thick and the structures, of Late Period date and unknown function, covered an area of 8000 m2. Furthermore, a fragment of a wall 12 m thick, total length 400 m, located in the south-western part of the site, possibly surrounding a temple complex from Ptolemaic times. The different magnetic values demonstrated by the various structures inside the complex were explained by the different composition of the bricks. Once archaeological fieldwork had established the date of selected structures, the settlement chronology could be read from the magnetometer results, including the functional changes occurring in the different districts over time. The western part, inhabited until Ptolemaic times, was turned into a burial ground, while the northern part, also abandoned by the residents, was turned into a flourishing pottery production centre—both these functional changes were reflected in the results of the magnetometer survey (Hartung et al., 2007, 2009).

Fig. 3
An aerial map of the site, a layout, and a top-view photo. A map has a blend of dark and light shades, with the reconstruction area highlighted. A layout with arrows marks the house plans in sector E. The excavated houses' foundations with thick walls arranged in squares, and workers engage in work.

Buto/Tell el-Fara’in. Left: results of the magnetometer survey of the western part of the site; the box marks outline of the reconstruction map. Right, above: reconstruction map (after Marouard, 2012, p. 107), the arrows mark the Late Period houses with plans verified by excavation. Right, below: foundations of the houses excavated in sector E (courtesy of DAI, phot. Ulrich Hartung)

Full size image

At Tanis (Sân el-Haggar), close to a third of the 200 ha of the tell in its northern part was occupied by a sacral district, which was excavated. The rest of the site was expected to be of a residential nature, an idea supported by satellite imagery showing up in a few places the outlines of buildings typical of the Late Period (Leclère et al., 2016, pp. 41–42). The results of the magnetometer survey showed that there are two clearly different districts, both in terms of layout and the kind of building material used (Fig. 4). On the western side of the central part of the tell and in the southern part the architecture, consisting of casemate buildings, follows an irregular street grid typical of the first millennium BC (Marouard, 2012). Stable values of the readings indicate dried mud-brick as the building material (Fig. 4a). In the eastern district, the architecture is predominantly of an insular kind separated by a regular network of streets, close to a square in shape. A series of high-magnetic anomalies is typical of the magnetic response of walls made of red brick (Fig. 4b). The layout suggested a Roman date for this district, and this agreed with the chronology of surface pottery (Defernez, 2015; Leclère, 2015). The analysis also confirmed the settlement chronology: Late Period and Ptolemaic ceramics on the surface are an indication of the district with Late Period architecture not being inhabited after Hellenistic times; and the settlement retreating to the highest central part of the tell around the religious enclosure (Leclère, 2019).

Fig. 4
An aerial map of the site and 2 aerial inset maps of the site parts, A and B. A map with survey areas highlighted and a horizontal scale ranging from 0 to 200 meters. A and B present the results of the magnetometer survey with patches of light and dark shades and a horizontal scale of 50 meters.

Tanis/Sân el-Haggar. Results of the magnetometer survey of the central and southern part of the site. White lines mark the extent of the survey areas A and B (A: dynamics −6/+6 nT; B: dynamics −4/+4 nT)

Full size image

The study at Pelusium (Tell Farama), a Graeco-Roman town at the mouth of the now defunct Pelusiac branch of the Nile where it flows into the Mediterranean, is an example of the integration of magnetometer and electrical resistivity surveys. The area east of the Byzantine-period citadel was surveyed, between the foundations of a Roman theater and the old lagoon shoreline (today the site is 4 km from the sea) (Jakubiak, 2009). The magnetometer results enabled a reconstruction of city quarters with streets and established the orientation of the architecture (Fig. 5). However, the red brick used as building material, characterised by high magnetic susceptibility, made difficult to reconstruct the plans of individual structures apart from the outer shape. The bulk of the walls were recorded in the negative: trenches where walls were dismantled down to the brick foundations (used in the Byzantine fortress) (Herbich, 2021). The electrical resistivity results often produced more detailed plans of streets and buildings (or confirmed those revealed by magnetometry, e.g. a circular building interpreted as a bouleuterion), possibly thanks to the good contrast provided by the soil high humidity and salinity against the building materials. The use of two methods concurrently allowed the type of building material to be distinguished (whether red brick, mud brick or stone). This was the case of a round feature in the western part of the map (red arrow in Fig. 5), characterised by low magnetic susceptibility and high resistivity, the combination of which suggested stone which has diamagnetic or weak ferrimagnetic properties of its own as the building material of choice (Herbich, 2020).

Fig. 5
2 aerial maps of the site. A map has a blend of dark and light shade patches, with the resistivity survey area highlighted. A map of the highlighted site with large light-shade patches and a few dark ones. Streets, a stone structure, and a dirt road are marked by arrows and a dashed line.

Pelusium/Tell el-Farama. Above: results of the magnetometer survey of the north-eastern part of the town. The white box marks extent of the resistivity survey. Yellow arrows point to the streets. The red arrow indicates a structure built of stone (well?). Dashed line marks the modern dirt road

Full size image

3.2 Palatial Centres

In the case of prospection covering large areas, analysing town plans can lead to the discovery of expected buildings the location of which had not been known, such as a royal seat of power. Avaris again is a good example, the city being the capital of Egypt during the 15th Dynasty (1650–1550 BC). A royal palace, measuring roughly 1 ha in size, was identified. It had two rows of building sections, each of different size and containing units of different size, arranged around courtyards (Fig. 6). The plan bears no resemblance to Egyptian palaces with their consecutive room arrangements and seems to owe its origins to the Near Eastern concept, indicating its link with the foreign Hyksos dynasty (Bietak et al., 2007). Excavations confirmed the Hyksos attribution of this building, bringing to light both: seal impressions with the name of king Khayan from the 15th Dynasty; and 6000 pottery vessels from the period found in pits. The largest of these ceremonial pits corresponds to an oval anomaly (more than 5 m of diameter) with high magnetic values (Bietak, 2011; Bietak et al., 2013).

Fig. 6
An aerial map and a layout plan of the site. A. A map in gradient shades for magnetometer survey, with patches of light and dark shades and a highlighted excavation area. B. A plan with a horizontal scale of 0 to 50 meters. It includes offering pits, courts, throne rooms, gates, and magazines.

Avaris/Tell el-Dabca. Left: results of the magnetometer survey of the Hyksos period palatial complex. White line marks the extent of the excavation grid. The white arrow marks an anomaly corresponding to a ceremonial pit filled with pottery. Right: plan of the palace after excavation. (Courtesy of Manfred Bietak, ÖAW)

Full size image

At the same site, magnetometer surveys were also useful in precising the plan of a Thutmosid palatial complex from the early New Kingdom (c. 1479–1400 BC). A preliminary tracing of the southern palace G showed to be twice as big as the northern palace F, discovered earlier. Another monumental building appeared to have stood southwest of palace G (Fig. 7). The palaces were built on platforms corresponding in plan to the buildings raised on them. The good results allowed different parts of the structure to be assigned provisional functions in analogy to the palace F (courtyard, vestibule, sanctuary, throne hall). The extent of the building was also determined. (Bietak et al., 2001, pp. 74–85).

Fig. 7
An aerial map and a layout plan of the palace. A. A map in gradient shades for magnetometer survey, with the majority of the area having a darker shade. B. A plan with a horizontal scale of 0 to 50 meters. The shrine is at the top center, with a ramp and a courtyard below. A bath is on the right.

Avaris/Tell el-Dabca. Left: results of the magnetometer survey of the Palace G (Thutmosid period). Right: plan of the Palace G, based on results of excavation and the magnetometer survey. (Courtesy of Manfred Bietak, ÖAW)

Full size image

3.3 Cemeteries and Cult Places

Tombs in Upper and Middle Egypt were located on the desert fringes neighbouring with the Nile valley, making them easily traceable with magnetometry because of the high contrast between magnetically susceptible mud-brick walls and the less magnetic desert matrix (Herbich, 2019, pp. 221–232). In the Delta, however, where cemeteries were founded near settlements, in Nile alluvial soils, the case is different because they are rarely marked on the surface and hard to find on cultivable land. Hitherto most of the cemetery discoveries are made on tells, which are not under cultivation, or else when surveys cover large swathes of agricultural land (as was the case of the Tell el-Dabca site, Forstner-Müller, 2009).

Buto and Tell el-Farkha are good examples of burials discovered by geophysical prospection on tell sites. At Buto the cemetery occupied the area of an abandoned earlier settlement. Saite Period tombs were located at the northeastern edge of the tell, in the ruins of the Old Kingdom site. Anomalies of rectangular form, with reduced magnetic contrast than their surroundings, correspond to tomb structures (Fig. 8a). The specificity of this type of response was demonstrated once excavations revealed that there was a greater amount of sand in the layers filling the tomb structures (Hartung et al., 2009, pp. 91–94, Fig. 2) (Fig. 8b). Remains of residential structures from the Late Period are also clearly visible south and east of the cemetery. Clusters of dipole anomalies with values ranging −20/+20 nT (Fig. 8a) correspond to Roman-age burials, which were placed in the ruins of buildings a few centuries older in date. These burials were made in terracotta coffins, a material characterised by a high magnetic susceptibility and thermoremanent magnetisation (Hartung et al., 2003, pp. 253–254) (Fig. 8d).

Fig. 8
An aerial map, an aerial layout, and 2 top-view photos. A. A map highlights trench J 2 and burial extent with dark and light shades. B. A layout of trench J 2 with walls and a tomb, 1 to 4 marked. C. A burial chamber with a tomb. d. A skeleton in a burial container with a scale and slate nearby.

Buto/Tell el-Fara’in. (a) results of the magnetometer survey of the north-eastern part of the site, with the location of trench J2. Dashed yellow line marks the extent of the Roman-period burials in terracotta coffins. (b) map of trench J2; 1—Third Intermediate Period walls; 2—Old Kingdom walls; 3—tomb J2/89 with limestone sarcophagus and sand filling; 4—probable Saite walls (after Hartung et al., 2009, fig. 2, p. 86). (c) Tomb J/89 seen from the west. D –Roman-period burial in a terracotta coffin. (c and d: courtesy of DAI, phot. Ulrich Hartung)

Full size image

At Tell el-Farkha, just as in Buto, excavations revealed the nature of the identified rectangular magnetic anomalies. These rectangles never exceeded 5 m in length and turned out to be Predynastic burials. Their study has contributed extensively to broadening current knowledge of burial practices in this period (Dębowska-Ludwin, 2012). The anomalies were caused by the different magnetic susceptibility of the backfill of the rectangular chambers compared to the values yielded by the mud-brick chamber walls and the surrounding matrix. The presence of large sets of pottery vessels inside the burial chambers also tended to change the magnetic value of the backfill of these rectangular features; the arrangement of the anomaly enabled a reconstruction of where the vessels were located in these tombs (Herbich, 2004).

The graves found at Tell el-Dabca were not verified by archaeological excavations and are dated based on the recorded shape of the structures, which can be compared to excavated burials and dated accordingly, as is often the case in the investigation of cemeteries in Upper Egypt (Forstner-Müller, 2009).

The effectiveness of magnetometry to explore cult centres was demonstrated by the survey of the Great Temple Enclosure from the Saite and Late periods at Tell el-Balamun. Excavations over a long period of time uncovered and identified the most important elements of the temple complex: sanctuary dedicated to Amon, temples built by Psamtik I and Nectanebo I; a fort with an annex on the north-east; a series of burials; and some structures from the Ptolemaic period (Fig. 9a). The enclosure walls of the temenos had been traced and dated (inner wall: 26 Dynasty—664-525 BC, outer wall: 30 Dynasty—380-342 BC) and surface remains clearly indicated the presence of production areas (Spencer, 1996). The objectives of a magnetometer survey at a site so extensively explored were supposed to be limited initially to explore a few areas in between the excavated structures but the results brought so much new and unexpected information that the entire enclosure was surveyed in effect. For example, the prospection uncovered the only larger stone building preserved in the complex, namely, a bark-station in front of the temple of Nectanebo (Fig. 9; 1). A previously unknown casemate-type building was mapped next to the so-called fort (Fig. 9; 2), and the approach to the temple of Psamtik turned out to be more extensive than previously supposed judging by the anomalies in front of the pylon (Fig. 9; 3). Industrial areas with pottery kilns or similar manufacturing facilities were also located (Fig. 9; 4). An analysis of the magnetic map also revealed an older sanctuary adjacent to the temple of Psamtik (Fig. 9; 5). The remains of this earlier structure were so poorly preserved that its identification would not have been possible without the magnetic image. The recorded fragments of a 30th Dynasty enclosure wall (on the northern side of the temenos) allowed for an in-depth reconstruction of the wall structure, its size and shape made up of separate, projecting and recessed panels of brickwork (Fig. 9b). At some points, the survey showed evidence of buildings on multiple levels; for example, the magnetic map of the Fort Annex revealed that the southern part of the building is completely overbuilt by a later structure (Fig. 9c). The architecture observed southeast of the temenos indicates that the outer wall had a very shallow foundation that had eroded revealing older buildings beneath, dated probably to the Saite and Third Intermediate periods (Herbich & Spencer, 2009; Herbich, 2016).

Fig. 9
An aerial map with 3 insets of a layout and 2 aerial maps. A site with outer and inner walls and B and C's extent highlighted, along with dark and light shades. A. An enclosure plan with temples, forts, and fort annexes marked. B. An outer wall map with dark shades. C. A map with a few dark patches.

Tell el-Balamun. Results of the magnetometer survey of the Great Temple Enclosure. The boxes mark extent of maps B and C. The explanation of numbers is included in the text. A: plan of the enclosure based on excavations conducted prior to the magnetometer survey (after Spencer, 1999, plate 1). The extent of the magnetometer survey is in light grey. The inner enclosure wall (26th Dynasty) is in grey, the outer enclosure wall (30th Dynasty) in black. B: detailed view of the results of the magnetometer survey of the outer wall. Dynamic −6/+12 nT. C: detailed view of the results of the magnetometer survey of the buling erected above Fort Annexe. Dynamics −6/+12 nT

Full size image

3.4 Production Centres

Evidence of production activities is practically universal at all sites. It comes in the form of features that were used to produce objects requiring high temperatures and which took on magnetic properties through the process of thermoremanence (Gaffney & Gater, 2003, pp. 37–38). These features include all kind of hearths, pottery kilns, metallurgical furnaces and traces of food processing, like breweries, for example.

The effectiveness of magnetometer surveys to reveal pottery production centres is shown by the results achieved at Buto, a known centre of the industry in Ptolemaic-Roman times. A few dozen furnaces were discovered in the northern part of the site, occurring rarely alone but in clusters of from 2 to 10 production units (Fig. 10). The furnaces were recorded in places already suggested by large amounts of slag, ash and burned soil (e.g., furnaces in trenches P1 2002–2003), as well as in areas with no observable evidence of production on the surface, e.g., trenches P3 and P4) (Hartung et al., 2003; Ballet et al., 2019). Excavations established the date of the complex in the Roman period, identifying the pottery-making technique as one known from western Roman and eastern Mediterranean workshops, but not evidenced earlier in the Near East. Traces of architecture from the Late Period and Ptolemaic times, already abandoned when the industry began, were also mapped (Hartung et al., 2007, 2009).

Fig. 10
An aerial map of the site and 4 box layouts. A map in a medium-lighter shade and a few patches of darker and lighter shades for survey results. P 1 2004, P 1 2002 to 2003, P 3, and P 4 are highlighted. 4 box layouts of P 1 2004, P 1 2002 to 2003, P 3, and P 4 with Roman kilns and Ptolemaic walls.

Buto/Tell el-Fara’in. Results of the magnetometer survey of the Roman pottery production centre in the north-eastern part of the site. The boxes show the location of kilns and other architectural remains. (After Hartung et al., 2007, pp. 126–143)

Full size image

Traces of metallurgical production were noted during the investigation of the site of Marea on the southern banks of Lake Mareotis. The workshops were located on a peninsula situated about 100 m from the city itself. The magnetometer results revealed a large workshop quarter extending over much of the surveyed area (Fig. 11). The walls correspond to negative anomalies, indicating the use of stone of weak magnetic properties, limestone in this case. The outlines of some units were established by anomalies of high magnetic values (reaching −30/+100 nT), interpreted as remains of a metallurgical workshop. Excavations confirmed the interpretation, uncovering a workshop which processes iron and bronze (Pichot, 2010). The high values were caused by slag, small fragments of iron and iron oxides, and ash covering the entire surface of the room. The metallurgical centre is Ptolemaic-Roman, earlier than the city, which is dated to the fifth to seventh centuries AD. The magnetometer exploration of the city area, with stone-made structural remains, did not produce an equally distinct result (Derda et al., 2020). Soils in this area are characterised by a low magnetic susceptibility in the range of 0.2*10−3SI—0.3*10−3 SI. The enhanced contract at the production area between the soil and walls of stone in terms of the magnetic susceptibility was due to waste, chiefly ash, metal fragments and slag, being present in the surface layers. Therefore, the soil and other deposits had a high magnetic susceptibility which strongly contrasted with the weak magnetic susceptibility of walls, something that was not the case in the city area.

Fig. 11
An aerial map of the Marea, an aerial layout plan of the excavated area, and a top-view photo. A. A map in gradient shades for a magnetometer survey with patches of light and dark shades. B. A plan with a horizontal scale ranging from 0 to 10 meters. C. A land with scattered stones and an ash cover.

Marea. Left: fragment of the results of a magnetometer survey of the industrial area on the peninsula. The yellow box marks extent of the excavated area; the arrow marks the anomalies of high magnetic values corresponding to a metal workshop (Space 10). Right, above: map of the excavated area (courtesy Valerie Pichot, CEA). The grey arrow marks Space 10. Right, below: Space 10 seen from the north-west (phot. Valerie Pichot)

Full size image

Magnetometer prospection in Hierakonpolis led to the discovery of a series of breweries. Excavation of one of the clusters of anomalies interpreted as possible production centres uncovered clay vats for making beer as well as pottery kilns and evidence of food processing, all from the Predynastic period (Baba et al., 2017; Herbich, 2019, pp. 235–238). The centre was presumably a place of activities connected with the neighbouring cemetery, and it constitutes unique evidence of food processing directly connected with pottery production. Breweries like those from Hierakonpolis, discovered in the Predynastic layers at Tell el-Farkha, were not observed there because they were too deep below the surface (Ciałowicz, 2012).

3.5 Landscape Research

The introduction of handled tape recorders and field computers to geophysical research in the 1980s made it possible dense sampling of large areas, enabling measurements of areas outside sites, the objective being a reconstruction of the landscape around the ancient settlements. Tell el-Dabca and Qantir are both examples of magnetometer prospection of this kind. The total area covered by measurements exceeded 350 ha. The cities were situated on the now defunct Pelusiac branch of the Nile. Drillings carried out in this area in the 1990s traced the river course and established the extent of higher ground during the annual Nile flood, which as a matter of course invited permanent settlement and was therefore a potential area for archaeological exploration (Dorner, 1999) (Fig. 12). A hypsometric map of the area as it would have been 4000 years ago was verified with geophysical prospection applying both the magnetometry and electrical resistivity (in this case using vertical electrical soundings or VES) (Forstner-Müller, 2009; Herbich, 2024). Measurements of the total Earth’s magnetic field with a caesium magnetometer gave good results in terms of tracing the course of the Nile in this area and the extent of the floodplain. The results revealed the location of the riverbank and the orientation of deposits in the riverbed, aligned with the flow of the main current (Fig. 12, box). Measurements with a fluxgate magnetometer were not as clear regarding the riverbank but turned out useful in reconstructing the waterfront and pinpointing mooring areas and places for loading and unloading goods. The VES performed in several places (every 10 m or 20 m, along lines of combined length equal to 8.6 km) helped to trace the interface between the riverbanks and floodplains thanks to a strong contrast between the low-resistivity deposits in the river bed and the higher-resistivity surface layers representing human occupation (Herbich & Forstner-Müller, 2013). A juxtaposition of the geophysical results with the geomorphological map changed the reconstruction produced as an effect of the drillings. The F3 branch was hardly a dead bay accessible only from F2 in the north but was connected with the F1 branch flowing around the southern part of the city (Herbich, 2024) (Fig. 12). Considering the bays on this branch, of which the northern one (F4) was a harbour, the mooring places on the left bank at Ezzawin and Mehesin, and the presence of a Hyksos palace in the southern bay (F5), one should view the F3 branch as a year-round passage rather than a temporary branch.

Fig. 12
An aerial layout of the site with contour lines and an inset of the aerial map of the survey site. The layout has main river channels, inundated areas, inundated area extent, and excavated areas marked, along with the survey site highlighted. A survey site with a few dark patches.

Tell el-Dabca/southern Qantir. Reconstruction of the water channels of the Pelusiac branch of the Nile and the inundated area. The box marks the extent of the survey using a caesium magnetometer and the results below (by Christian Schweitzer). 1—main river channels F1 and F2; 2—inundated areas/F3 channel according to the survey results; 3—extent of inundated area according to Dorner (1999), not verified by the geophysical survey; 4—excavated areas. Contour lines after Dorner (1999)

Full size image

4 Concluding Remarks: What the Future Holds

The number of archaeological sites in Egypt tested by geophysical methods in the past 30 years is probably close to 150–180, much less than in most European countries, but if one looks at how many of these have been verified archaeologically, then Egypt clearly ranks among the highest on the list. Personal observation by the author suggests that broad-scale excavations have been conducted at almost all of the sites where he has carried out geophysical research, sometimes covering areas measured in hectares (e.g., Fig. 6). It clearly sets down the objectives for geophysical research in Egypt: it is considered as a tool to identify settlement or cemetery site layout and search for features to be excavated. The key function of archaeological geophysics in some European countries, which is to determine areas for heritage protection, is practically unknown as such in Egypt.

It does not seem that magnetometry will lose its primacy in the investigations of sites in Egypt to be the superior method of geophysical prospection. As discussed above, it is the result of a combination of feature/pedological/geological characteristics and logistic-related aspects. The share of projects where GPR will be applied will probably increase in the near future. The method has shown potential to establish the layout of settlement structures at desert sites, where so far most of the work was done exclusively with magnetometry. GPR has the further advantage of being capable of tracing changes in urban layout at different depths. Its application would be particularly effective on sites located on the fringes of oases and the seacoast where the silt used for brickmaking has a poor content of iron oxides. It could also be a method of choice for sites with an abundance of pottery in the surface layers, where magnetometry cannot be applied due to the high magnetic disturbance.

In term if instrumentation, mobile multi-probe systems used for magnetometry do not appear to have a future in Egypt. The size of individual plots of land in the Nile Valley (including the Delta) is very small and there is an extensive system of irrigation channels. Also, the ground surface at desert sites is usually too uneven due to extensive illicit digging in the past. In the author’s experience, a system of this kind could be applied only to a handful of sites of this kind.

Another specificity of Egypt is that most of the geophysical prospection is carried out within the frame of projects run by archaeological centres located outside of Egypt. Egyptian geophysicists working on archaeological sites are for the most part from NIARG, an institution for which archaeological geophysics is definitely a sideline. Their publications seldom include examples of archaeologically verified results, while the interpretations presented are based usually on earlier determinations regarding given sites (e.g. Abdallatif et al., 2019). It is only in recent years that the presence of geophysicists in Egyptian archaeological projects has started to be noticeable, showing a change in the relation between the two specialties. This will probably be a growing trend because the number of strictly Egyptian archaeological projects is steadily on the rise.

There is also another factor in favour of archaeological geophysics in Egypt—ultimately, virtually every survey in the Nile Valley and Delta, at least with magnetometry, will reveal archaeological features owing to the specificity of sun-dried mud brick used as the principal building material. This is hardly the case in European archaeology. Therefore, the results of research in Egypt provide examples that I believe convincingly support the use of geophysical methods in archaeology, understood as a discipline that studies the entire cultural heritage, regardless of region.