1 Introduction
Since 2011, the Université du Québec à Montréal (UQAM)—University of Memphis joint epigraphic mission at Karnak has been recording and studying the complex decoration on the columns of the Great Hypostyle Hall in the Temple of Amun-Re at Karnak. With generous funding from the Social Sciences and Humanities Research Council of Canada (SSHRC), the US National Endowment for the Humanities (NEH) and other agencies, we have employed emerging technologies to gain new insights into the elaborate decorative program and complex epigraphic history of the columns2
Through digital photogrammetry, we have produced unrolled orthomosaic photographs (déroulés) showing the entire surface of each column, enabling researchers to appraise their complex array of decoration in one image and to understand the spatial layout of the scenes and stereotyped texts that adorn them. These orthomosaic déroulés improve our comprehension of the interrelationship between the main ritual scenes and the various layers of stereotyped inscriptions on the column shafts, capitals, and abaci.
The great height of the columns, architraves, and clerestory roof of the Great Hypostyle Hall presents us with the daunting challenge of how to record and study them. The sheer elevation of the 12 great open-bud papyriform columns of the central axis is an insurmountable obstacle, while their wide campaniform capitals largely obstruct the view of their abaci from ground level.
During our November–December 2017 field season at Karnak, we overcame this difficulty by using an unmanned aerial vehicle (UAV), or “drone,” to fly above and around the upper reaches of the columns, abaci, architraves, and clerestory roof, along with blockyards in the Karnak precinct where hundreds of loose blocks originally from the Hall are now stored. By photographing these more remote and less well studied parts of the Hall from various angles and building 3D models of them, we have gained a powerful new method for recording these inaccessible parts of the building and gained new insights into its architecture and inscriptions.
This chapter offers a case study illustrating the advantages of using emerging technologies, especially UAV s, which provided us with an unprecedented view of the Hypostyle Hall, to record and analyze inscribed monumental structures.
2 General Overview of the Great Hypostyle Hall at Karnak and Main Project Objectives
Situated in the heart of the temple of Amun-Re at Karnak—the most important religious monument of the New Kingdom located inside Thebes, the ancient 4th nome of Upper Egypt—the Hypostyle Hall is the largest and best preserved pharaonic architectural feature of its kind in Egypt. Measuring 103 m wide and 52 m long, the Hypostyle Hall originally contained 134 sandstone columns (Figure 18.1). A double row of 12 giant columns with campaniform capitals and exceeding 20 m in height stand along the main East-West axis. The average diameter of their shafts is 3.5 m; their circumference, replete with inscriptions, measures 10 m. To the north and south of the central aisle are two wings, each with 61 closed-bud papyrus columns divided into seven rows. Their average diameter is 2 m, with a circumference, also filled with carvings, of approximately 6.5 m.
Figure 18.1
Plan of the Hypostyle Hall at Karnak (E. Feleg)
The columns inside the Hypostyle Hall are unique, both in their scale and numbers, and also in the complexity of their decorative program. It is the first building in ancient Egyptian architectural history to have columns inscribed from top to bottom with inscriptions. This unprecedentedly complex decorative program,3 with multiple layers of reliefs, many of them palimpsests, is the work of no fewer than five kings who left their mark on the columns: Sety I, Ramesses II, Ramesses IV, Ramesses VI and, (more modestly) Herihor. The inscriptions each pharaoh left vary significantly both in terms of their position on the columns and the overall quantity of decoration. As we have discussed elsewhere, the complex arrangement and chronology of these reliefs, many of them being palimpsests, follows the notion of “prime space.”4 According to this logic, kings gave priority to sections of the columns most visible from the processional axes, and these were the first to be decorated. Understanding the successive phases of decoration on the columns is challenging. Ramesses II alone engraved them in four different stages, leaving palimpsests where he erased earlier decoration of Sety I and his own initial reliefs. Some columns have been inscribed and re-inscribed in six or seven phases during the Nineteenth and Twentieth Dynasties.
There are two categories of column decoration. In the first are approximately 350 ritual scenes depicting Sety I, Ramesses II, or Ramesses IV offering to various gods and glossed with hieroglyphic captions. Most columns have three offering scenes around their circumferences, offering the richest variety of iconographic detail: the costume and regalia of the king and gods, and the offerings (Figure 18.2 area ②; Figure 18.3 area ②).
More abundant is the stereotyped decoration of these kings and of Ramesses VI and Herihor, covering every other available space on the columns. This diverse assortment includes floral and rekhyet-bird motifs engraved at the bottom and/or top of each column (the work of Sety I in the north wing and Ramesses II in the south aisle); friezes of royal cartouches at different levels, at the basis of the columns and on the upper shafts and capitals (by Sety I, Ramesses II, IV, and VI); and multiple sets of bandeaux texts containing royal titles (Ramesses II and IV, Herihor) (Figure 18.2 area ①; Figure 18.3 area ①).
Figure 18.2
Facsimile of the déroulé of a central column (E. Feleg)
Of the 134 original columns, 129 remain standing in varying degrees of preservation. Less remains of the roof. The ceiling slabs have vanished, except for a few, still in place next to the vestibule of the Third Pylon (Figure 18.4 area ①). The upper shafts and abaci of a cluster of columns are missing inside the southwest quadrant of the Hall, swept away when the bulk of the neighboring southern tower of the Second Pylon collapsed sometime before the mid-19th century (Figure 18.4 area ②). In the northern aisle, eleven columns with their abaci and architraves collapsed in 1899 in a devasting domino effect. Legrain replaced most of the original architraves in the north wing with metal and concrete pastiches when he rebuilt the northern wing in the early 20th century (Figure 18.4 area ③). Only the bases of columns 89 and 98 remain in situ (Figure 18.4 area ④), and the lower half of column 74 in the north now stand (Figure 18.4 area ⑤). Beside the north wall, columns 130, 131, and 133 are missing entirely (Figure 18.4 area ⑥). Drum sections belonging to some of these latter columns are stored just outside of the Hall, to the north (Figure 18.4 area ⑦).
Figure 18.3
Facsimile of the déroulé of a lateral column (E. Feleg)
Although of paramount importance to our understanding of Ancient Egyptian religion and royal ideology, the columns of the Hypostyle Hall had never been the object of systematic study until we began our study in 2011.5 Nelson and Caminos produced an unpublished archive of hand copies of the column scenes and abaci inscriptions. L.-A. Christophe published an inventory of the deities and ritual acts depicted in the scenes.6 Otherwise, previous studies and documentation efforts centered on the wall scenes and architraves.7
Figure 18.4
Aerial view of the Hypostyle Hall
©C. Zandona
The gap in the scientific investigation of the columns can be ascribed to their huge size and geometry, which long stymied traditional methods of epigraphic documentation. The apparently redundant stereotyped inscriptions, much of it engraved well out of reach, and the impossibility of recording hundreds of large scenes wrapping around the cylindrical surfaces of the columns using traditional photography all deterred scholars from recording them until the emergence of 21st century digital technologies.
3 Project Aims, Methodology, and Results
The UQAM—University of Memphis joint epigraphic mission at Karnak has endeavored to record and publish the column scenes systematically, with a focus on transcription, transliteration, translation, and commentary of the texts. Other areas of research include the architecture of the columns and roof, the paleography of hieroglyphic signs, the study of quarrymen’s marks left on the stone blocks from the Hall, the study of palimpsest inscriptions, of iconoclasm, and chromatology (color studies) of pigments artists used on the columns.8
Between 2011 to 2019, we completed eight field seasons to achieve these goals, ranging from one to three months per season. In our first seasons, we collated the stereotyped column decoration and ritual scenes with the aid of unpublished hand copies and notes assembled by H.H. Nelson, R. Caminos and W.J. Murnane. Although tedious, this process was highly rewarding, leading to several vital insights into how the Egyptians conceptualized the vast decorative program of the columns. Most significant was the notion of “prime space,” i.e., the prioritization of the most visible spaces in a three-dimensional ritual environment.9 We also gained nuanced understanding of iconographic and paleographic norms and exceptions in the arrangement of decoration executed on a vast scale on the 134 columns and of the interrelationship between quality of workmanship and location of decorative elements. Recent seasons were also devoted to sorting out errors made in the placement of inscribed blocks during George Legrain’s directorship, when he had much of the northern wing of the Hall reconstructed at the turn of the 20th century. In 2019, we developed a new method for rebuilding dismantled columns from unrolled images of individual blocks no longer in situ.
4 First Generation déroulés of Entire Columns: Objectives, Methodology, and Results
Our future publications, online and in print, will include photographic déroulés of all the columns. In 2008, before the UQAM/University of Memphis project began, Emmanuel Laroze, who was then Director of the Franco-Egyptian Centre for the Study of the Temples at Karnak, used laser scanning and photogrammetry to create a 3D model of the Hypostyle Hall in collaboration with a private French firm called ATM-3D and the National School of Geographical Sciences near Paris.10 This undertaking included a full three-dimensional lasergrammetric reconstruction of one of the great columns and the first déroulé orthomosaic image of its entire surface. This technique unwrapped a composite image of the full circumference of a column from a cylindrical 3D digital model, allowing the entire decoration of a column to be seen at a glance (Figure 18.5 and Figure 18.6).
Figure 18.5
First generation orthophotographic déroulé of a central column (ATM-3D/Y. Egels/E. Laroze)
Using photogrammetry, the French team developed a method in which georeferenced digital images of a column, using a local coordinate system, are stitched together and assembled in 3D, then unrolled (not unlike the method for projecting a three-dimensional image of the earth’s globe into a two-dimensional world map). Prior to carrying out this difficult operation, the team acquired a data set by taking some 4,000 pictures of all the standing columns. They devised an ingenious method for taking pictures of each column by mounting four digital cameras at set heights along an 8 m high pole that they moved to eight positions around each column, taking multiple, overlapping shots, section by section.
Figure 18.6
Reprocessed orthophotographic déroulé of a lateral column (Y. Egels/E. Laroze/O. Murray)
This approach produced astonishing results: déroulés presenting the entire decoration of the column in one composite image, allowing us to gain a much better grasp of the sequence of scenes and stereotyped decoration on various levels of the column shafts. The orthophotos of unwrapped columns served as templates for producing scaled diagrams of a column’s decoration, with the freedom to distinguish between the different chronological phases of decoration (Figure 18.2 and Figure 18.3).
Admittedly, there were some drawbacks to the results this pioneering method produced. In 2008, the French team used small commercial digital cameras that produced low resolution images. Moreover the project, whose aim was to document all standing 129 columns in the Hall, was completed in just one month. As time was of the essence, photographs were taken in shade, in order to avoid contrast between shadow and sunlit areas of the columns, which combined with the low resolution of the camera, made it difficult to read damaged hieroglyph signs and palimpsest inscriptions in the resulting déroulés. Furthermore, it was necessary to collect more data in later years for the post-processing treatment of the images to be truly effective. Since ATM-3D was no longer involved in the project, E. Laroze and Y. Egels took these images and measurements between 2013 and 2017. Through their meticulous radiographical treatment of the first generation of déroulés, the final images are much more legible, but the limitations of the original images in terms of lighting and resolution remain.
5 Second Generation déroulés of the Main Scenes and 3D Reconstruction of Columns
To overcome the limitations of the first generation déroulés, we decided to produce a new set of rectified orthophotos of more than 300 ritual scenes at the mid-level of the standing columns in the Hall, these having the greater historical and religious significance and variety than the stereotyped inscriptions. These second generation déroulés consist of much higher resolution orthophotos, enabling the user to zoom in on all the relevant details of the scenes in unprecedented detail. In the 2014 and 2015 field seasons, photographer Owen Murray shot a sequence of up to 150 images of each of the smaller 117 standing closed-bud papyrus column scenes. This method of documentation relied on a set of moveable scaffolding from which a series of 7–8 photographs along the vertical, Z axis of the column were taken from 16 positions around the column (X,Y axes). In the 2016 field season, O. Murray finished photographing the remaining 12 large column scenes applying a similar methodology, this time using a hydraulic scissor lift from which a series of 13–14 photographs along the vertical, Z axis of the column were taken from 21 positions around the column (X, Y axes). The result was a sequence of up to 350 images of each of the scenes on the 12 great open-bud papyriform columns.
In this second generation of déroulés, O. Murray used controlled lighting to bring out the fine details and palimpsest traces of the reliefs, creating a “soft box” to diffuse natural light and a studio strobe to evenly light the column scenes. This facilitated optimal raking light photographs of the reliefs (Figure 18.7). Combining the new photos with the 3D georeferenced data compiled since 2008 to generate a virtual model of the Hall, we could now produce a new set of orthomosaic déroulés of the column scenes. With the invaluable help of Agisoft’s photogrammetry software Metashape (previously called Photoscan), we were able to streamline the complex and tricky process for fabricating the déroulés by aligning the individual pictures (Figure 18.8), producing a dense cloud, a mesh, and finally a high-resolution texture of the model. After processing the images in Photoshop, the final composite déroulés are spectacular (Figure 18.9).11
Figure 18.7
Use of softbox for picture taking (Y. Egels)
Figure 18.8
Construction of 3D model of column (P. Brand)
Figure 18.9
Déroulé of a complete column scene (Y. Egels/E. Laroze/O. Murray)
Since a 3D model must be built in order to generate the flattened déroulés, Metashape also allows a user to rotate the columns in a virtual environment and view them from different angles (Figure 18.10). In this case, since the scenes are located mid-level on the columns, one can view them virtually in a manner not physically possible at the temple. Worthy of mention is that the production of the déroulés and the 3D projections were all done by graduate students of the history department at UQAM, under the technical supervision of E. Laroze and O. Murray. This fact shows that the expertise necessary to master these sophisticated tools can be transferred to people who have no prior experience in the field. To our great pleasure, three life-size second generations déroulés of columns from the Hypostyle Hall were exhibited in Canada in 2019, at the Montreal Fine Arts Museum in Montreal (MFAM) (Figure 18.11) and at Gallery@501 in Edmonton.
Figure 18.10
Three-dimensional view of a column (O. Murray/E. Laroze/Y. Egels)
Figure 18.11
Life-size déroulé of a column scene from our mission at the MFAM (J. Revez)
6 Assembling Loose Column Drum Fragments with the Help of Orthophotography and 3D-Reconstruction
Applying this successful method of orthophotographic déroulés developed to produce the scenes carved around the entire circumference of standing columns has also opened up new and exciting avenues of research for column drum fragments scattered throughout the blockyards around the Karnak Temple precinct. Anastylosis, as an art and science, can gain tremendously from the use of photogrammetry to unroll individual loose column blocks lying far apart from one another, and in turn, enables us to understand the sequence and relationship of these fragments to one another.
In 2019, we successfully tested the method on all the loose blocks identified as being part of column 131, a column that is no longer standing in situ inside the Hall. The results have been extremely promising. The first step was to identify and record the 20 column drum sections belonging to column 131, then to draw sketches and determine the measurements for each of them. These half-drums measured about 2.5 m in diameter and 0.93 m in height.
Identifying and reconstructing the scattered fragments that make up columns formerly standing inside the Hypostyle Hall, such as column 131, is a complex operation based on several iconographic, epigraphic, and architectural criteria.12 From an iconographic and epigraphic viewpoint, determining the type of carving (raised or sunk relief) in which the artists cut the figures and inscriptions is crucial, since Sety I used raised relief exclusively in his decorative program. Ramesses II (except for the first year of his reign) and Ramesses IV always carved their inscriptions in sunk relief. The orientation of the figures inside a scene is also generally consistent. Sety I’s scenes always face the processional ways in the northern section of the Hall, while Ramesses II’s did the same in the south wing. Ramesses IV’s decoration faced away from the axis. Thus, the orientation of the reliefs is also crucial to establishing their original position. The scenes carved along the main axes show the gods, who were perceived as residing within the temple, looking out towards the different gates and greeting the king, who faces towards the interior of the temple.
Another crucial epigraphic criterion is the presence of traces of palimpsest inscriptions, which are typical of reliefs Ramesses II recarved, over both Sety I’s in the north wing, and his own early raised relief in the south. These are only found on columns adjacent to the main east-west processional way or along the secondary north-south axis in the southern half of the Hall. Traces of iconoclasm can be seen in various parts of the Hall, but the specific sections of the body of the king or the gods that were hacked, and the type of damage left by the iconoclast’s tools, can be ascribed to particular sectors inside the Hall. Last but not least, the conventional sequence of the successive registers of stereotyped decoration made up of friezes of cartouches and bandeaux texts with their royal and divine titles and epithets, makes it fairly easy to assign the position of a block on any given column. The same can be said of original stereotyped plant motifs (the stylized triangular leaf patterns and floral design at the bottom of the columns and the ‘binding motif’ on the upper portion of the shafts representing the idea of the cluster design of the papyrus bundles), used in carving the closed-bud papyriform columns of the Hall.13 Their location in space can be assigned quite clearly along the vertical axis of each column. Ramesses IV never added stereotyped decoration on most columns in the southwest quadrant of the Hall, in which case the vegetation motifs are all that remains for us to determine the respective position of each drum. Finally, since a single scene and its accompanying legend extended over several connecting drums, observing which elements of the overall decoration are carved along the edges of each individual block makes it possible to determine how separate fragments join.
On a more architectural plane, the position of the mortises on the upper surface of the blocks, where wooden dovetails were inserted to join two half-drums together, is a key element in the anastylosis of pairs of column drums. Since each column is made up of courses of half-drums set up alternately lengthways and sideways from one course to the next, it is also possible to determine how each drum is positioned in connection to the one set up just above and below it.
The reassembly process described above was greatly helped by reference to 19th century photographic archives that show the state of the columns inside the Hall just prior to its tragic collapse in 1899. Since the Hypostyle Hall at Karnak is perhaps second only to the Great Pyramids in Giza as the most photographed monument in Egypt, we can rely on numerous archival photos to assist us in the anastylosis of the columns. In the case of column 131 for instance, Figure 18.12 not only shows us that the upper section of the column was no longer in situ at the end of the 19th century. It helped us tremendously in identifying and locating four half drums that were physically set aside on concrete platforms (commonly known as mastabas) just outside of the Hall. A drawing by R. Lepsius from the mid-19th century led us to confirm the assemblage of lower sections of the column that are hidden from view in the photo. Finally, one can also turn to the invaluable study of the columns carried out in the 1950s by L.-A. Christophe who provided useful insights for reconstructing missing columns. His hypotheses have proved to be highly accurate.
Figure 18.12
Northeastern section of Hall in 1895 and in 2017
©CFEETK/O. Murray
Moving the heavy blocks so as to reassemble them is impractical. Such an operation would be cumbersome and costly (to say nothing of the fragile condition of many of these ancient blocks that have lain in contact with ground water for a century and are often in dire need of conservation). This is not a viable approach for collating their decoration. Virtual reconstruction of déroulé images of the decorated surfaces of these column drums offers an ideal, non-invasive method for reconstructing them via photogrammetry.14
To achieve this, two approaches were tested. The first was to make individual flat déroulés of each block, as we did in our approach to develop complete scenes carved on columns that are still in situ. The second method involved producing a three-dimensional reconstruction of each fragment. This method is more complex. In either case, the 2D- or 3D-model allows for digital manipulation of each decorated block from the drums, which could be moved around a standard column template in order to test fit matching blocks (Figure 18.13 and Figure 18.14). By building a gigantic two- or three-dimensional jigsaw puzzle, block joins can be identified by trial and error.
Figure 18.13
2D reconstruction of column 131 with individual déroulés of loose blocks (Y. Egels/E. Laroze/O. Murray)
We used Cumulus, an open-source software program for 3D reconstruction of archaeological sites developed by Y. Egels, a member of our team, for producing orthomosaic déroulés of the standing columns inside the Hall. Transporting blocks virtually was done through human-machine interfaces (HMI), by attempting to interact with the software in such a way as to enact real life situations as much as possible. Thus, the process of maneuvering and rotating blocks is similar to what crane operators would do if physically moving the blocks themselves.
Figure 18.14
3D reconstruction of column 131 made up of loose blocks (Y. Egels/E. Laroze/O. Murray)
This new methodological approach, based on emerging technologies, offers several advantages. The greatest one is the speed and ease with which the reconstruction of the column can be completed. Once the recording in the field was done, we could instantly and without great effort visualize whether sets of blocks matched and could be joined together, or on the contrary, if they did not fit together. The theoretical series of criteria for identifying matching blocks for anastylosis that we itemized in the previous section becomes much easier, since they can be observed first-hand in a virtual workspace. As with the déroulés of the in situ columns, we have produced a composite image of the three adjoining scenes on column 131. Assembling the small flattened déroulés of the column fragments in various combinations made it possible to determine whether the elements of a scene overlapping several blocks fit accurately.15
In addition, the ability of this system to create two- and three-dimensional images of the column drums is profitable in other ways. Matching the two halves of the mortises with the wooden dovetails (Figure 18.15 area ①) so that the circumferences of both half-drums align perfectly (Figure 18.15 area ②) allows us to confirm that the two blocks match. Another criterion for a perfect fit becomes obvious: the two intersecting lines incised on top of the two blocks (these lines guide the stone workers when positioning the row of half-drums located just above) cross each other at a right angle (Figure 18.15 area ③). Other means that contribute to the virtual reassembly of blocks, which would be practically impossible to accomplish by manipulating the blocks themselves, include the fact that the column shafts are slightly conical rather than perfectly cylindrical Thus the diameter of a layer of blocks varies according to its vertical position on the column. Such a feature can be much readily visible with the model.
Figure 18.15
2D reconstruction of two matching half-drums belonging to column 131 (Y. Egels/E. Laroze/O. Murray)
7 Using an Unmanned Aerial Vehicle (UAV) to Document the Hypostyle Hall
Early in the 2017 field season, experiments with a camera-on-pole technique to photograph the large column abaci from atop the central axis architraves of the Hall proved viable, though requiring much time, as well as attention to safety while working at such heights (Figure 18.16). The permission to use a drone (UAV) to acquire imagery greatly changed the scope of work possible for the season (it is believed this is the first time a UAV was flown at a major temple complex in Luxor, and used for archaeological research and cultural heritage documentation purposes). It was quickly determined that the best course of action would be to produce scaled photogrammetric models of the architraves and abaci atop the large columns, as well as the accompanying clerestories on either side, as rectified orthomosaic imagery from any desired perspective could be extracted from these models. The ability to scale the models was due to CFTEEK survey efforts during the initial 2008 déroulés campaign; these same measurement and control points were used with newly acquired imagery from the UAV.
Figure 18.16
Use of a camera-on-pole technique on top of an architrave to take pictures of abaci (E. Laroze)
We used a a DJI Phantom 4 (2016) UAV mounted with a 12mb digital camera, kindly provided to our project by the Epigraphic Survey (Chicago House) (Figure 18.17). Though initial firmware update issues were cause for great consternation, DJI technical support and forum boards provided solutions quickly and resolved the issue within 48 hours.
Due to the large volume of tourists and the desire to acquire imagery in the best possible lighting conditions, flights in the Hypostyle Hall were conducted from approximately 6am to 8am each morning from December 6 to 13, 2017. A staging area in the first court between the first and second pylons clear of obstacles was used as a base for a launching and landing zone, and as a return-to-home location in case the data link between the UAV and the Remote Control device (RC) failed during flight. We determined that a flight radius of 500 m and a maximum altitude settings of 40 m, with a preferred 25 m ceiling, were optimal and were input in the UAV DJI Go 4 control application before launch, along with setting the camera for automatic acquisition of JPEG images at two seconds intervals during flight. Owen Murray piloted the UAV by Visual Line of Sight (VLOS) at all times, with Emmanuel Laroze co-piloting and giving instructions as to directional movements of the UAV and adjustments to the camera position as needed. Although wind patterns within the temple were of concern, this work method proved quite effective, with work in the aforementioned areas completed within the morning flight window each day.
Figure 18.17
The DJI Phantom 4 (2016) UAV mounted with a 12mb digital camera (E. Laroze)
8 Recording the Hitherto Inaccessible Upper Portions of the Hypostyle Hall
From the onset of our project, one of the important goals was to record all of the cartouches carved on all four sides of the abaci resting on top of the columns. The analysis of the spelling of the large-size royal names and the type of engraving used (raised or sunk relief, with or without traces of palimpsest) encapsulates the history of the decoration of the Hall under the reigns of Sety I and Ramesses II. After having shot points with a Laser Theodolite on predefined reference points located on the abaci faces to produce rectified images, we were able to improve on our method by pairing the georeferenced data from the initial 2008 CFEETK survey with the images taken from the drone.
Figure 18.18
Scaffolding and Scissor Lift used by the project (E. Laroze)
Flying the drone around and above the abaci located on top of the central colonnade, as well as over other sections of the Hall, enabled us to take pictures of areas hitherto inaccessible by other means, whether with the help of a scissor lift or of scaffolding (Figure 18.18), two modi operandi we had turned to successfully to reach the upper sections of the smaller lateral columns which are only 13 meters high. The sheer height—over 20 meters—was one problem; another was the fact that the abaci on top of the large east-west central columns were resting on large open-bud capitals that hid the view of the abaci from ground level. The UAV allowed us to observe and record many heretofore unfamiliar epigraphic and architectural details of the abaci (Figure 18.19), capitals, and clerestory roof (Figure 18.20).
Figure 18.19
Orthorectified abacus on top of large column (O. Murray/ E. Laroze/Y. Egels)
Figure 18.20
3D model of upper section of central colonnade (O. Murray/ E. Laroze/Y. Egels)
Pictures taken from the flying camera corroborated and reinforced some observations previously made about the ancient repairs of the westernmost column capitals and abaci under Greco-Roman rulers. V. Rondot and J.-C. Golvin16 observed that some parts of the campaniform capitals of columns 1 and 2 lacked any trace of decoration, in contrast to other sections of these capitals that were inscribed with cartouches of Ramesses II. They also pointed out that the abacus sitting atop column 2 was left blank (Figure 18.21 area ④) where one would have expected to see the cartouches of Sety I, as is the case with all the other abaci from the same row of columns running to the east. They concluded that the undecorated fragments were later additions from Greco-Roman times, made after the partial collapse of an architrave resting atop the abacus, destroying it and part of the capital when it fell.
Figure 18.21
Upper section of large columns no. 1 & 2 (O. Murray/E. Laroze)
With the assistance of our drone, we were not only able to confirm the accuracy of these remarks, but we could also clearly identify similar restoration blocks that were inserted inside the upper section of the capital of column 1 as well, a detail that is not visible in photos included in the article by Rondot and Golvin due to the angle from which the photograph was taken.17 Images from the drone show that ancient restorers neatly cut a small vertical hole into both sides of a replacement block on column 1 (Figure 18.21 area ①). These two perforations were probably inserted to allow a rope to run through them to help support the new block as it was inserted in between two adjacent blocks.
Images taken from the drone also provided insights into other architectural peculiarities. The center of the upper surface of the capital of column 2 was slightly cut out in order to allow the restored Greco-Roman abacus to fit into the gap. Notice on the picture that the stonecutters did not deem it necessary to smooth the upper surface of the capital, which is only roughly hewn (Figure 18.21 area ②). Usually, the opposite occurs: the abacus is set on top of a small flattened outgrowth that protrudes upwards from the upper surface of the capital. Once set into its place, both the sides of the abacus and the edges of the excess are cut back, smoothed, and aligned. A closer look at the images taken from the drone shows that all four corners of the elevated platform, sticking out from the upper surface of the capital in order to support the abacus, were hollowed out (Figure 18.22 area ①). The small canals that were thus carved out probably crisscrossed through the platform diagonally and were used to control the accuracy of the cutting, assuring in this case that the platform was of even height throughout, and that its top side was levelled off correctly in order to receive the lower surface of the new abacus. It is worth noting that the upper surface of the capital measures more than 30 square meters. Cutting flat and horizontal bedding joints on such a large surface was no small feat. Such a construction technique is well attested in architecture generally speaking, but evidence for this method in ancient Egypt is relatively scant. One can also detect in Figure 18.22 area ② pry-holes that were carved at the base of blocks in order to secure the position of a wooden lever that was used to set the blocks into the right position. A small patch stone was later put into place to plug in the pry-hole at the base of the architrave.
Figure 18.22
Elevated platform under abacus and pryholes (O. Murray/E. Laroze)
On a different plane, scaled orthorectified longitudinal sections of the northern row of the clerestory made it possible to record the measurements of all the different architectural features that make up the upper section of the colonnade. Anomalies or deviations from the norm in some of these measurements proved to be quite significant in meaning. Thus, one can make out that the abacus resting on top of column 1 was not set precisely on the vertical axis of the column. It is, in fact, off center from the axis by 25 cm to the west, as Figure 18.23 area ① clearly shows. One could at first conclude from this observation that the abacus, which is an original one dating back to Sety’s reign, could have been accidentally moved during the later restoration campaign of the Classical Period. Such hypothesis turns out to be very unlikely, since the abacus on top of column 7 that lays symmetrically directly across column 1, to the south, presents a similar pattern, with a 25 cm departure from the norm. However, the capital on top of column 7 did not undergo any repairs during Greco-Roman times, as opposed to the capital of column 1. In fact, the lines, that were incised on the upper surface of the capital in order to guide the workers when laying the abacus in its right position, are still clearly visible on the picture and show that the abacus in its current state rests in its original place, since the sides of the abacus run along those lines squarely (Figure 18.21 area ③).
Figure 18.23
Longitudinal section of central row of large columns (O. Murray/ E. Laroze/Y. Egels)
Thus, the reason for this 25 cm discrepancy must be found elsewhere. It probably has to do with the relative position of both abaci on top of the two columns (no. 1 and no. 7) that were erected right next to the gateway of the Second Pylon at the western end of the Hypostyle Hall. Because of the outward slope (or batter) of the eastern face of the pylon, there is a greater distance between the top of the column (which stands about 20 m above ground level, lest we forget) and the top section of the gate structure, than between the base of the column and the base of the gate. By placing the abacus 25 cm closer to the inner wall of the second pylon, the ancient architects of the Hypostyle Hall most likely wanted to reduce the size of the architrave that spans the distance between the top of the column and the gate of the pylon. This distance is indeed shortened from 6.31 m to 6.06 m.
Figure 18.24
Upper section of large colonnade with remains of roof-slabs (O. Murray/ E. Laroze /Y. Egels)
During past seasons at Karnak, two types of New Kingdom quarry marks (also called masons’ marks) were newly identified in no fewer than 17 spots throughout the Hall. The first type of mark is a circle connected to a small cross (similar to a nfr hieroglyphic sign and not unlike the modern female sign); the other is an ‘hourglass’-shaped symbol. These marks were carved presumably on rough-hewn faces of blocks in order to identify the team responsible for their extraction in the quarry and transportation to the temple. Up to now, these marks were found on both the northern and southern walls of the Hall, in the masonry of the east wall, and of more direct concern to us, on the drums and abaci of the small columns only. The conspicuous absence of these marks on the large central columns led us to believe at first that their erection could perhaps be of a different date than the smaller ones. However, pictures taken from the drone during the 2017 field season made it possible to detect nfr-shaped quarrymen’s marks on four different blocks supporting the north and south clerestories of the Hall and another one, on the eastern face of the abacus sitting on the large column no. 7.
Other normally inaccessible areas of the temple were also surveyed with the help of the drone. In the highest part of the Hall, the UAV flew over the only remains of the roof slabs that covered the building. From the clear traces left by the action of cutting tools on the edge of the remaining slabs to the north of the Hall (Figure 18.24), one can gather that the roof slabs connecting the north clerestory to the large northern colonnade did not cave in by accident, but were intentionally broken off before collapsing, probably in order to reuse the fallen material for reconstruction purposes elsewhere.
Figure 18.25
Overhead view of south-west quadrant of the Hypostyle Hall (O. Murray/E. Laroze)
The top view of the southern part of the Hypostyle Hall offers a wealth of technical architectural features. One can see the system of clamping of six blocks on top of a column (Figure 18.25 area ①). Notches facing each other and intended to support wooden beams are also visible (Figure 18.25 area ②), as well as pry-holes used to move roof-slabs (Figure 18.25 area ③). Sockets for wooden cramps (or dovetails) are visible; their emplacement is random (Figure 18.25 area ④). From the extra thickness of stone and rough surface discernible on top of the architraves, one can deduce that the roof-slabs did not cover the entire surface of the architrave but rested solely on the smoothed parts (Figure 18.25 area ⑤). Incised lines at the edge of some architraves border the area of stone that was due to be cut back (Figure 18.25 area ⑥).
The most inaccessible zone of the Hypostyle Hall in terms of elevation is the clerestory, where the highest inscriptions were carved some 22 m above ground level. Reliefs and texts carved on the vertical piers and architraves encasing the huge window grilles of the clerestory roof are of great interest but are largely unpublished. On the northern clerestory for instance, just underneath the roof slabs, Ramesses II engraved a line containing his royal titles over those of his father, Sety I (Figure 18.26). Though the hieroglyphic signs are large in size and can thus be observed with binoculars from below, the fine details of the palimpsest inscription can best be examined fully with close-up shots taken from the drone. In the picture shown, traces of the original raised relief text of Sety I (Figure 18.26 area ②) can be quite easily deciphered under the recarved sunk relief inscription of Ramesses II (Figure 18.26 area ①). Noteworthy of mention here is the use of the same 
m-owl sign (Gardiner sign-list G17) in both texts (for the word 
wḥm in the suppressed raised relief text of Sety I and the verb 
mk in the later sunk relief of Ramesses II) (Figure 18.26 area ③). In this case, the sculptor reworking the inscription on Ramesses II’s orders had only to convert the carving of the glyph from raised to sunk relief to the other, without the need to erase the whole sign, as it is the case with the other glyphs in the inscription that could not be “recycled” in this manner.
Figure 18.26
Palimpsest inscription on North clerestory (O. Murray/E. Laroze)
Finally, flying the UAV over the Hall gives one the ability to look at the state of the building from the perspective of its conservation. For instance, it is easier to determine from up close whether fragments of roof blocks that had fallen onto the clerestory level need to be removed or if cracks and fissures within the architraves require urgent restoration or consolidation.
9 Recording and Building 3D Models of the Loose Blocks Stored Outside of the Hall
In 1899, the fragile state of the column foundations led to the collapse of 11 of them in the northern half of the Hall.18 Georges Legrain, the then maître d’œuvre at Karnak, proceeded to rebuild most of them, along with other columns that had fallen previously. This operation, carried out over many years, was a truly amazing feat, considering the size of the columns and the extent of their damage. Some columns inside the Hall that were either truncated or in poor shape at the time of the catastrophe, were not repaired, while others, such as columns 130 and 131, were dismantled, but never reconstructed. These latter two columns stood immediately next to the northern gate, in the first row inside the Hall lining the eastern section of the northern wall, where there is now a large empty space. For decades, the loose drums belonging to these hitherto disassembled columns, and to other columns belonging to the northern half of the Hall, were stashed away just north of the monument, before being partly removed further east, next to the Osiris-Heqa-Djet chapel. More blocks, belonging to the upper sections of columns originally in the southern half of the Hall—themselves fallen casualties of blocks toppled down from the upper portions of the Second Pylon when it caved in—now rest on mastabas lined up in the large open air storage area between the Khonsu temple and the Hypostyle Hall.
All in all, no fewer than eight large areas spread across the Karnak Temple precinct contain hundreds of pieces of various sizes belonging to drums and abaci that are no longer in situ inside the Hall. Consequently, a major (and exciting) challenge of the project—and one that has been the focus of our efforts for the past three seasons—is the anastylosis of the dismantled, broken down or decayed sections of the monument. The virtual, and hopefully in the future the physical, reassembly of all the fallen portions of the columns is one of the project’s goals. For this purpose, several related objectives had to be met: first, identifying, recording, and tagging all the loose drums and abaci that were part of the Hall; second, determining their original position along the vertical axis that typically make up the eleven layers of a column; and third, determining the original location of these column fragments within the Hall.
Using a general plan of the site made by the Centre Franco-Égyptien d’Étude des Temples de Karnak (CFEETK), we first identified the areas in the Karnak Temple precinct where the loose blocks were located. We then flew the drone over those areas in order to generate detailed high-resolution sketch maps of each zone, onto which we plotted and labelled each and every loose column fragment from the Hall (Figure 18.27). This undertaking proved extremely useful when, chart in hand, we needed to locate individual blocks in the field among thousands of sandstone masonry blocks. Perfectly square-shaped abaci and semi-circular drums have characteristic contours, so the pictures taken from above allowed us to identify fragments that had hitherto escaped our attention.
Figure 18.27
Sketch-map of zone 1 comprising loose blocks belonging to the Hall (O. Murray/E. Laroze)
Another interesting result of our experience working with the drone was our ability to use the image data to build 3D models of the mastabas and their blocks in each zone where the loose fragments lay. O. Murray produced such a model for zone 2, located in the northeast quadrant of the temple, consisting of three mastabas where the greatest concentration of loose fragments and column drum sections can be found (Figure 18.28). Off site, we could use these models to navigate virtually over the mastabas, moving around the blocks as one might in person in order to examine their inscriptions, or whatever other details we deemed of interest, more closely. For missions like ours that can only stay in Egypt for relatively short field seasons each year, the ability to navigate and explore the site virtually and experiment with the 3D model to test fit pieces together while working in our office back home, saves valuable time in the field in future seasons.
Figure 18.28
3D model of mastabas in zone 2 comprising loose blocks (O. Murray)
The high-resolution maps of each zone were generated by flying multiple passes of the drone over the areas in question in a linear fashion at two heights (approximately 8 m and 15 m) with the camera facing down, but moving parallel to the ground (Figure 18.29). In addition to these topographic views, the 3D model of zone 2 used multiple passes of the drone over the mastabas with the camera at an oblique angle to the ground (approx. 45°), as well as a perimeter pass around each mastaba, with the camera facing inwards towards the fragments, from a height of 3–5 m. The camera was then rotated to a 90° angle so as to face the block fragments, and another perimeter pass was flown around each mastaba at the lowest height possible from the ground; 1–1.5 m, and as close to fragments on the mastabas as possible without triggering the close proximity warning feature of the drone: 2.5 m.
In order to augment the quality of the model, in addition to the photographs captured using the drone camera, a hand-held perimeter pass around each mastaba was also completed using a Nikon D800 with a Nikkor 28 mm lens. These additional facing photos were captured at a distance of 1–1.5 m from the fragments and allowed a greater degree of precision in shooting pieces with more complex and detailed features. All of the photographs were post processed using Adobe Lightroom and exported as high-resolution jpegs to Agisoft Metashape (then Photoscan) and aligned, retaining the drone metadata for accurate position and scale information. A high-quality dense cloud of approximately 66.8 million points was generated from the aligned photos, which in turn was run on a high setting to produce a model with approximately 13.3 million polygons. This was textured at 16,492 pixels and the resulting model allowed researchers and students the ability to virtually visit zone 2 from their offices back home using the same software (Agisoft Metashape).
Figure 18.29
O. Murray piloting the drone over the mastabas in zone 2 (E. Laroze)
3D Models mastabas with their column block fragments and virtual anastylosis of columns like no. 131 can also prove interesting and invaluable ways of conveying information to both professionals and laymen. On site temple QR code signage that allows anyone with a mobile phone to visit these models, hosted on the relevant research institutions websites, provides both guides and tourists alike with a more in-depth understanding of the temple structure, and a glimpse into what in many cases is no longer physically standing.
10 Conclusion
The use of an unmanned aerial vehicle (UAV) and other emerging techniques has been invaluable to the Université du Québec à Montréal (UQAM)—University of Memphis joint epigraphic mission at Karnak. Producing series of orthophotographic déroulés of entire columns, of their most meaningful sections (the middle registers containing the ritual scenes) and of individual loose blocks has made it possible for us to overcome successfully the obstacles associated with producing 2D and 3D representations of the non-planar surfaces of columns. As for the UAV, it has given us access to the upper portions of the Hall hitherto out-of-reach by conventional means. Emerging techniques, such as drones, photogrammetry and photo rectification, represent a great scientific asset to generate high-quality images of the columns of the Great Hypostyle Hall at Karnak. But even more importantly, these sophisticated technologies have led to phenomenal progress in terms of procedure and methodology, especially in the field of anastylosis, where emerging techniques serve as an inestimable testing ground to validate hypotheses of reconstruction.
The project discussed in this chapter is a joint University of Quebec in Montreal—University of Memphis epigraphic mission that has been ongoing since 2011. For current and past financial support, we thank the Social Sciences and Humanities Research Council of Canada, the National Endowment for the Humanities, The American Research Center in Egypt, The University of Memphis, The University of Quebec in Montreal, the Tandy Institute at Southwestern Baptist Theological Seminary, and the Centre national de la recherche scientifique (CNRS). For their invaluable collaboration and assistance during our field seasons in Karnak, we are very grateful to the Centre franco-égyptien d’étude des temples de Karnak (CFEETK. MAE-CNRS USR 3172), the Egyptian Ministry of Antiquities, and the Chicago House in Luxor for having allowed us to use their flying camera (UAV) in December 2017.
Except for Figure 18.4 and the archival photo shown in Figure 18.12 (picture on the left), all the pictures and illustrations included in the present article were produced specifically for the joint University of Quebec in Montreal—University of Memphis epigraphic mission. Owen Murray, our photographer, took most of the pictures on site.
Note that the in-text citations are kept to a minimum and the reference list at the end of the article comprises entries that are only closely related to our project.
Revez 2020; Egels et al. 2020.
Brand 2000, 192–218.
Revez and Brand 2015.
Revez and Brand 2012, 11–13 for a short historiography on the Hall; see also now Biston-Moulin 2016.
Christophe 1955.
Epigraphic Survey 1986; Rondot 1997; Brand, Feleg, and Murnane† 2018.
Brand et al. 2013.
Revez and Brand 2015.
Chandelier et al. 2009; Laroze and Chazaly 2009.
Revez 2020, 394–397.
Egels et al. 2020, 47–49.
Brand 2018.
Egels et al. 2020.
Egels et al. 2020, 48–50.
Rondot and Golvin 1989.
Rondot and Golvin 1989, pl. 31.
Azim and Réveillac 2004, vol. I, 130–178; vol. II, 39–83.
Bibliography
Azim, Michel, and Gérard Réveillac. 2004. Karnak dans l’objectif de Georges Legrain: Catalogue raisonné des archives photographiques du premier directeur des travaux de Karnak de 1895 à 1917. 2 vols. Paris: CNRS editions.
Biston-Moulin, Sébastien. 2016. Inventaire des monuments, objets, scènes et inscriptions des temples de Karnak. Montpellier.
Brand, Peter J. 2000. The Monuments of Seti I: Epigraphic, Historical and Art Historical Analysis. PdÄ 16. Leiden: Brill.
Brand, Peter J. 2018. “Patterns of innovation in the monumental art of Ramesside Thebes: the example of the Great Hypostyle Hall of Karnak.” In The Ramesside period in Egypt: studies into cultural and historical processes of the 19thand 20thdynasties, edited by Sabine Kubisch and Ute Rummel, 45–61. Berlin; Boston: Walter de Gruyter.
Brand, Peter J., Emmanuel Laroze, and Jean Revez. 2011–2012. “Le projet américano-canadien des colonnes de la salle hypostyle de Karnak. Objectifs globaux et méthodologie.” Journal of the Society of the Study of Egyptian Antiquities 38: 17–34.
Brand, Peter J., Jean Revez, Janusz Karkowski, Emmanuel Laroze, and Cedric Gobeil. 2013. “Karnak Hypostyle Hall Project. Report on the 2011 field season for the University of Memphis & the Université du Québec à Montréal.” Les Cahiers de Karnak 14: 193–229.
Brand, Peter J., Rosa Erika Feleg, and William J. Murnane†. 2018. The Great Hypostyle Hall in the temple of Amun at Karnak. 2 vols. Oriental Institute Publications 142. Chicago: Oriental Institute of the University of Chicago.
Chandelier, Laure, Bertrand Chazaly, Yves Egels, Emmanuel Laroze, and Daniel Schelstraete. 2009. “Numérisation 3D et déroulé photographique des 134 colonnes de la Grande Salle Hypostyle de Karnak.” Revue XYZ, Revue de l’association française de topographie 120: 33–39.
Christophe, Louis-André. 1955. Temple d’Amon à Karnak: Les divinités des colonnes de la grande salle hypostyle et leurs épithètes. Bibliothèque d’Étude 21. Cairo: Institut français d’archéologie orientale.
Egels, Yves, Emmanuel Laroze, Jean Revez, Owen Murray, and Peter Brand. 2020. “Une nouvelle colonne dans la salle hypostyle de Karnak. Récit d’une méthode de restitution.” Revue XYZ. Revue de l’association française de topographie 164: 45–50.
Epigraphic Survey, The. 1986. Reliefs and Inscriptions at Karnak, Vol. 4: The Battle Reliefs of King Sety I. Oriental Institute Publications 107. Chicago: Oriental Institute of the University of Chicago.
Laroze, Emmanuel, and Bertrand Chazaly. 2009. “Relevés des colonnes de la grande salle hypostyle de Karnak.” Comptes rendus de l’Académie des Inscriptions et Belles-lettres 153: 669–685.
Revez, Jean. 2020. “3D scanning, photogrammetry, and photo rectification of columns in the Karnak hypostyle hall.” In The Oxford Handbook of Egyptian epigraphy and palaeography, edited by Vanessa Davies and Dimitri Laboury, 387–404. New York: Oxford University Press.
Revez, Jean, and Peter J. Brand. 2012. “Le programme décoratif des colonnes de la grande salle hypostyle de Karnak: Bilan de la mission canado-américaine de 2011.” Bulletin de la Société française d’Égyptologie 183:10–38.
Revez, Jean, and Peter J. Brand. 2015. “The Notion of Prime Space in the Layout of the Column Decoration in the Great Hypostyle Hall at Karnak.” Les Cahiers de Karnak 15: 253–310.
Rondot, Vincent. 1997. La Grande Salle Hypostyle de Karnak: Les architraves. 2 vols. Paris: Éditions Recherche sur les civilisations.
Rondot, Vincent, and Jean-Claude Golvin. 1989. “Restaurations antiques à l’entrée de la salle hypostyle ramesside du temple d’Amon-Rê à Karnak.” Mitteilungen des Deutschen Archäologischen Instituts, Abteilung Kairo 45: 249–259.