Queen Hatshepsut’s Heavy-Lift river barge

The stone mason knew best

Rev. B, dd. July 1, 2018

Abstract

Key questions: How accurate was the relief on the wall of Hatshepsut’s temple at Deir el-Bahri? How could the Heavy-Lift barge (HL-barge) on the relief be loaded and discharged, and how was the structural integrity of the barge ensured during all stages of the operation? How did the tow navigate down-stream on the river Nile? This report presents possible answers from an operational point of view, concluding that the stone mason knew best.

Note: to read a footnote please click on the reference number between brackets behind each word. There are 20 footnotes in total.

p

Introduction

The description in this document follows as closely as possible the layout and the details shown on the reliefs on the wall of Hatshepsut’s temple at Deir el-Bahri.

To understand how the HL-barge could have been constructed and how it could have been operated in about 1470 BC, it needs to be established first how it was discharged, how it navigated on the river Nile and how it was loaded. It is evident that before the barge was loaded, it had to be ensured that it could be discharged. The barge cannot be considered on its own as the transport was a “door-to-door” operation, a marine heavy lift project. There is no information available from original sources to date. Therefore, heavy lift practice of the present times is applied with only the assistance of contemporary tools and practices such as sledges, rollers, spades, levers, ropes, and manual force from a very large number of labourers. Iron was not available at the time. Different water levels of the river Nile over one year must have been used to the advantage of the operation. It would not have been possible without the help of water.

The loaded barge (1) has been depicted as a relief in a sailing condition on the river Nile on the wall of Hatshepsut’s temple at Deir el-Bahri. The relief has been shown in fig.1. The bow points to the right.

Fig. 1. Relief 1 at Deir El-Bahri
(GNU Free Documentation License 1.2)

The waterlevel of the river Nile is very high over a period of about 3 months, especially near Aswan. The high-water period runs from August till November, during which the currents are at their highest, to the advantage of the navigation on the river. The river floods over its banks during that period. During the other months, the level is low and the depth of the river is shallow especially at the banks. The difference between high water (HW) and low water (LW) could be as much as 8 m at Aswan and 7.5 m near Cairo(2),(3). The rise at Luxor is about 0.84 m above the temple floor according to F. Said(4)  and 0.50 m, or several meters more, according to of F. Monderson(5) on the basis of salt deposits in the stone at Karnak. The latter has been selected for this study assuming regular decline between Aswan and Karnak.

The document presents also a few new considerations for the operation. The numbers quoted in this document are based on basic calculations, and are meant to support the conclusions, demonstrating the feasibility of the transport method as presented in this document. A typical construction of the barge needed to be worked out for this study considering Egyptian practice at the time.

p

Conclusions

It can be concluded from this study that the layout and the details, shown on the wall reliefs of the heavy lift barge in the temple of Deir el-Bahri, are technically correct, and that the transport could have been conducted in the manner described in this document with sufficient safety to the cargo, the craft and the crew. The reliefs are realistic. It cannot be proven that the operation was actually done in this way as no records have been found up to the present day.

The Heavy-Lift river barge of queen Hatshepsut was developed from similar smaller vessels used for smaller obelisks(6). It was destined for one type of cargo only and for one loading condition (equally distributed along the centre line), possibly only for one or two voyages. After completion of the first voyage, the vessel could be dismantled for the return voyage, or stored for future use with similar cargo.

The Egyptian naval architects had an excellent understanding of Archimedes Law, 1000 years before Archimedes.

This transport operation on the river Nile can better be defined as “controlled drifting”.

p

Acknowledgements

The author likes to thank Prof. Dr Stephan J. Seidlmayer, professor for Egyptology at the Freie Universität Berlin and Director of the Deutsches Arch. Institut – Dept. Kairo, who has reviewed this article, and who has made a few very valuable comments and suggestions, which have been included in this document. Further the author likes to thank Mr John Evans, retired director of AMT Ltd of Farnham, UK, a marine heavy lift transportation company, for his review of the practicality of the operation and of the correct use of the English language.

 

Description of the cargo

The cargo is comprised of two flat laying obelisks with the following particulars, each: Length 28.50 m. Base dimensions approx. 2.50 x 2.50 m. Weight estimated 375 t. The base dimensions have been scaled down from photographs, the weight comes from various sources. Fig. 2 shows the obelisk of Hatshepsut (right) at the beginning of the twentieth century.

Fig. 2. Obelisks of Tutmosis I and Hatshepsut (right) at Karnak. From: The Study of the unfinished obelisk at Aswan(7)

Coordinates of referenced locations.

Karnak temple complex at Luxor, Google Earth—-25° 43ʼ 05.09” N     32° 39ʼ 29.82” E

Quarry at Aswan, Google Earth —————————-24° 04ʼ 36.37” N     32° 53ʼ 43.29” E

The route has been shown on the map in fig. 3. Distance about  213 km.

Fig. 3. Section of the Nile from Aswan to Luxor (Image©2018 DigitalGlobe)

 

Discharging at Karnak.

The Karnak temple complex is orientated perpendicular to the river Nile. The barge was brought still afloat in position for discharging.  Therefore, the obelisks could be moved on rollers for the last part of the track, or skidded along a prepared track straight along an unobstructed route perpendicular to the centre line of the temple complex, to their final positions. See fig. 4. From the contour lines at Karnak, obtained from Google Earth, and the rise of the water level during a period of high water, it is clear that during HW the site of the temple complex flooded every year(8). This allowed landing of heavy weights at the site during construction.

Fig.4. Track at Karnak (Image©2018 DigitalGlobe)

During the period of low river level (LW), a bed would have been prepared near the construction site at Karnak to allow mooring of the barge when the river Nile is at its highest level (HW). When the water level dropped after a few weeks, the barge settled on the prepared bed, on which it could stay for a prolonged period of time when the area dried out completely. The bed was constructed such that it supported the structure of the hull of the barge over her full length and possibly part of the bow. The structural arrangement inside the barge allowed a good load distribution, when there was no water above the bed. A 7-9 m high dam with a length of about 70 m was required between the barge and the destination of the obelisks between the fourth and the fifth pylon at Karnak. See fig. 5.

Fig. 5. Discharging at Karnak (author)

The first obelisk was moved with the top in forward position and was rolled or skidded, and parked past its final position prior to upending. The second one was rolled or skidded with its bottom end forward and was probably erected first. For that reason, the obelisks were placed on board on the centreline of the barge, one behind the other in opposite directions. The foc’sle had to allow passage over the bow. The tie-ropes of the hull were running over portals of varying heights, forward ones being the highest to give extra support to the bow. The ropes split sideways to port and starboard allowing passage of the obelisks. The bowsprit was structurally extended down to the base line of the hull, such that the contact with the prepared bed was extended forward as far as possible.

The elevated land route had to be prepared for the haulage of both obelisks. Two rows of blocks were providing the foundations for the route. The tops of the blocks had to be in one level plane. Wooden rails were laid on top of the blocks for easier rolling or skidding. They provided the surface on which wooden rollers or skids were running underneath a sledge fixed to the obelisk. As the obelisk was moving forward, rollers and rails appeared from behind and had to be carried forward. This means that, if well planned, the skid tracks on the barge and the skid tracks on land were in one horizontal level plane and only friction needed to be overcome. After the cargo had been transferred ashore, the barge could be disassembled in the dry and the materials could be shipped back up-river using normal cargo vessels. The light weight of the barge has been estimated as 1250 t. The cargo vessels could sail up-stream using sails and oars during the LW period when the currents in the river were small and most of the time northerly winds were blowing. Thus, the HL-barge could be used again for two more obelisks if required.

With the delivery of the two obelisks on the shore, the marine operation was completed.

p

Upending.

Although the upending of the obelisks is no part of the marine operation, some thoughts have been dedicated to the procedure for upending, because if the obelisks were delivered in a condition from which they could not be upended, the marine operation would be useless. The rise of the water was used also to bring the obelisks in an elevated position, necessary for upending. So, no uphill skidding or rolling was required.  The description refers to the operation for the second obelisk. See fig. 6.

Fig. 6. Upending of the second obelisk (author).

It should be noted that this procedure is an outline only. The second obelisk was moved forward until the centre of gravity was past a pivot point. The pivot point was located on a temporary retaining wall, set at a precise distance to the base. Over the distance A-B, the sledge was firmly pulled against the obelisk by tensioning ropes such that the obelisk does not slide when the sliding force is 320 t. By digging away the embankment in front of the retaining wall, the obelisk rotated in a controlled way into position 2 by gravity until it hit a stopper down below. From position 2 to position 3, the obelisk had to be pulled by significant large number of hands, with holdback from another large number of hands.

p

Navigating on the river Nile.

The relief in fig. 7 shows 27 tugboats and 3 pilot boats tied to the bow of the HL-barge at different locations operating in three rows. Sᴓlver(9) has estimated that the joint pulling power of 30 towing rowboats would be equivalent to the bollard pull of a 50-horsepower tugboat. It has been assumed that this is at zero forward speed. Such a tug can provide about 500 kg pulling force. A major part of this force may have been used for steering the bow in the right direction avoiding obstructions en route, as well as for keeping ahead of the barge when some speed is gained from the “downhill effect” of the water surface of the river (see next paragraphs for further explanation). Keeping out of each other’s way may also have taken some effort. Consequently, there will not have been much left to give the tow some forward speed. This kind of navigation on the river can be described better as controlled drifting.  This also means that there was no room for mistakes by the master of the tow or by the pilots. It would have been nearly impossible to make corrections in time.

On the bow of the leading rowboats, pilots were present to sound the water depth with long poles and to give directions accordingly. There were three escort rowboats (below on the left), which were not connected to the barge, for performing religious ceremonies. A smaller rowboat is also shown on the relief (standby function?). It is expected that the master was on the foc’sle of the HL-barge, which would give him a good overview. Three high level dignitaries were also on the foc’sle (10).

Fig.7. Relief 2 at Deir El-Bahri
(GNU Free Documentation License 1.2)

The shipment took place during a period of high water on the Nile. The current velocity was about 3.3 knots(11). Based on information from Google Earth and the rise of the river at high water, the difference in water levels between Aswan at the quarry and at Karnak (assuming 2.50 m water over the temple floor ref. Monderson(12)) could be as much as 21 m. The sailing distance is 213 km. This means that the river runs downhill with a gradient of 0.010% on the average. This results in a downhill component of the weight of about 200 kg, in assistance to the oarsmen on the row boats(13). It is estimated that this added 1.1 knots to the speed of the barge in relation to the water, and enough to give some pressure on the rudders. The presence of four big oars for steering suggests this effect to have been known at the time. The row boats must be able to move faster than 1.1 knots, otherwise they will be overrun by the barge.

The speed of the convoy in relation to the land was probably 4.4 knots or 8.1 km/hr. Sailing at night may have been considered too risky, and people needed rest and food. When sailing is restricted to 8 hours per day, excluding (un)mooring, the route was covered in 3½ days. Leaving and entering ports especially may have been nerve-racking for the pilots, as well as manoeuvring through the rapids near Aswan, not to mention negotiating bends in the river. Maybe marker buoys were installed at critical locations. It is likely that special procedures were developed for such operations, written down in a book with instructions (operations manual). Apparently, the Egyptian navigators managed to do this without damage to the cargo or the barge, an impressive achievement in its own right!

 

Loading at Aswan

The quarry was located at Aswan. An unfinished obelisk can still be found in that location confirming the orientation of the obelisks (see fig. 8). The obelisks were orientated at an angle of about 60 degrees to the river. In line with the orientation, a prepared bed was constructed a bit further into the valley, which only flooded when the waterlevel in the river was high. It took about 7 months to cut an obelisk out of the bedrock. During the same period, the HL-barge was constructed on the prepared bed when dry, assembled from prefabricated timbers.

Fig. 8. Quarry at Aswan, unfinished obelisk (Image©2016 DigitalGlobe)

Upon load-out, the track from the quarry was extended via a dam over the barge deck, after which the first obelisk was moved on board before the start of the period of high water, immediately followed by the second one. See fig. 9. The particulars of the land route were similar to the land route at Karnak, except it was running downhill at a small gradient (probably 2.4%) over a distance of about 400 m in a straight line. The arrow on fig. 10 shows the probable location of the prepared bed. Once on board, sea fastenings were installed. Water tightness of the hull could to be confirmed during the rising of the waterlevel in the Nile. The raised foc’sle had to be temporarily removed to allow access along the centreline of the barge.

Fig. 9. Cross section along quarry towards river Nile, along route A-A (author)

Fig. 10. Route down to the river, via flooded valley at HW (Image©2016 DigitalGlobe). Dotted line shows overland route of obelisk via a dam to the possible loading location of the barge.

Fig. 10 shows the sailing route from the quarry to the river through the flooded valley. The rise of the river at Aswan was about 8 m(14) under average conditions. The fleet of rowboats and assisting vessels were mobilised in time for the float-out. Prior to joining the river, the barge was turned around bow first in the flooded valley in quiet waters. Joining the river, with a current velocity of 3.3 knots, required a lot of effort from the oarsmen, and it demanded experienced navigators and pilots.

 

Hydrostatic properties

A lines plan of the barge has been made (see fig. 11), which is the basis of the hydrostatic calculations in this document. The lines of the frames only are shown below, but a full lines plan was made for the calculations. There are two planes of symmetry, centre line and mid length, hence only one quarter of the hull is shown.

Fig.11. Frames (author)

The barge required to be shallow draft, with a curved hull and with extra deck strength along its centre line. The barge had to carry the cargo and its own light weight, and it had to have sufficient stability during the river transfer. The stability during loading and discharging was no problem as the hull was be supported on a prepared bed. To judge the free-floating stability, present day criteria for heavy lift transportation have been applied for reference.

The following is a summary of the hydrostatic calculations of the hull.

Barge particulars for this study

Length (L) over hull:  81.80 m Length(wl) : 65.30 m
Beam (B) over hull:  27.30 m Beam(wl) : 25.50 m
Depth to maindeck:   5.00 m Light draft:    1.75 m
Loaded draft:   2.50 m Block coeff. δ: 0.480
Displacement:   2000 t Main frame coeff. β: 0.735
Light weight, est:   1250 t

Stability, expressed in metacentric height (GM) in meters.

Metacentre above keel                KM = BM + BK = 23.01 + 1.50 = 24.51 m

Centre of gravity above keel       KG = 4.81 m

GM = 24.51 – 4.81 = 19.70 m(15) (> 1.00 m)

The stability appears to be more than sufficient, even for present day criteria.

 

Hull planking

For the hull planking sycamore(16) wood was used, found locally in Egypt. The reason was probably cost, as wood imported from the Lebanon for shipbuilding would have made the barge very expen-sive for only one or two shipments. The type of wood used required a special hull design and building technique as it was only available in relatively short lengths. The wooden sycamore blocks were shaped to form a segmental arch similar to a segmental stone arch in old churches, but upside down. The seams were compressed by the water pressure in transverse direction and by the hogging (middle-up) condition of the loaded barge in the other direction. Papyrus sealed the seams from the inside. The blocks were “pinned” together (mortise-and-tenon joints). Traditional Egyptian boats do normally not require extensive framing for the strength of the hull considerations. Framing was required for load distribution under vertical columns shown below (see fig. 12 and fig. 14). The shape of the arch was probably elliptical, without any flat surfaces. For the drawing of the construction of the hull details of the Khufu ship(17) have been studied. The special arrangement of the hull planking caused a certain flexibility in the overall structure of the barge. This would have resulted in a good distribution of the internal loads introduced by the cargo, thus avoiding load concentrations anywhere in the hull. The water tightness was ensured by caulking with papyrus. Only the obelisks were extremely stiff against bending, they kept the barge straight under load.

 

Framing arrangement

The relief shows the three deck beams above each other protruding the hull at specific locations (see fig. 12 below). Portals for tie-ropes have been shown every two frames at equal distances. It has been assumed that underdeck the same supports are extended down to the bottom of the hull. Three deck beams are connected with blocks such that each individual beam bends with the same deflection under load. Wooden spacer blocks assure that this is the case. This means that the stiffness of three beams can be added up for combined strength. The beams were made from cedar wood from the Lebanon, as well as the longitudinal deck planking where longitudinal strength was very important. Very long beams would normally be available from the Lebanon, although figure 12 shows three sections instead. The transverse deck beams also prevented the hull from deflecting outward and keeping the hull in shape, thus adding to the hull overall strength as well as carrying the deck load.

Fig. 12. Main frame (author)
Left: grounded during load-out, right: afloat

For argument sake, the structure of the main frame of figure 12 has been analysed for strength, the results of which have been shown in figure 13. The light weight of the frames has been added to the weight of the obelisks, and the total has been factored with 10%. The beams are comprised of cedar wood with an allowable stress(18) in bending 34 N/mm2, allowable in compression 27 N/m and an allowable shear of 5.6 N/mm2. If the three beams are deflecting together in the same way, their joint section modulus can be expressed as follows:

S = 3 x 500²x 300 / 6 = 37.5 x 106 mm3

Bending stress                  σ = 89 x 107 / (37.5 x 106) = 23.7 N/mm2  (max. 34 N/mm2)

Shear stress                       τ = 1.5 x 20.6 x 104 /(3 x 500 x 300) = 0.7 N/mm2 (max. 5.6 N/mm2)

When the barge is on the prepared bed, the full weight goes down one column and the compressive stress can be calculated as follows:

Compressive stress         σ = 20.6 x 104 / (500 x 300) = 1.4 N/mm2  (max. 27 N/mm2)

It can also be noted that intermediate frames are required as the main frames alone cannot take the load. The intermediate frames are not visible on the relief in the temple of Deir el-Bahri.

Fig. 13. Strength analysis of assumed main frame structure (author)

The obelisks were supported on at least 15 main frames and 14 intermediate frames in direct load transfer. The barge was at its strongest over the centre line. Fig. 14 shows the layout of the intermediate frames.

Fig. 14. Intermediate frame (author)
Left: grounded during load-out, right: afloat

Longitudinal strength.

There are no waves on the Nile during navigation, so any wave bending moments could be dis-regarded for the design by the Egyptian naval architects. When only one loading condition needs to be considered, it is theoretically possible to design the lines of the hull such that the distribution of the displacement along the hull is exactly the same as the distribution of the weights in free floating condition, in which case there would be no bending moments in the hull structure at all. In practice this is difficult to achieve and small bending moments will occur (see fig. 14). These bending moments must be taken up by the hull, for this vessel by tie-ropes in tension and bottom in compression. For tensioning up, the ropes must each be double and they must be comprised of combined right-hand lay and left-hand lay in each strand and all running in the same direction. The tie-ropes are also required during load-out when the barge is positioned on the prepared bed, when only the bow is (partly) unsupported. Alternatively, the bow could be further supported on an extension of the prepared bed.

The ends of the barge must be shipshape to ensure water tightness and low resistance. As the deck load is more or less equally distributed over the length, and most of the displacement is concentrated over about 70% of the waterline length, the barge will resume a hogging condition (middle up) and the obelisks will not be evenly supported. This can be corrected by tensioning the tie-ropes to resume a straight line for the deck. The weight of the cargo is transferred to the water via 29 vertical main frames, directly under the load at the locations shown on the relief by the beams protruding the hull in the side, and at intermediate frames. As long as the barge is supported on one of the prepared beds, the vertical load is transferred directly into the supports on the bed at the vertical frames and no longitudinal bending of the hull takes place (except some at the bow during loading and discharging).

The wall relief shows clearly that the obelisks have been loaded with their lower ends towards the middle of the barge (fig. 1). This is where the barge has most of its strength in bending and most of its displacement to take vertical loads. Fig. 14 shows the overall distribution of the loads in the hull resulting from the weights and the displacement (Δ and G), and the bending moments (curve M). The bending moments in the hull have been calculated along the length of the hull. This results in a maximum bending moment of 720 tm at mid length. If the relief is regarded closely (fig. 1), it can be seen that the portals supporting the tie-rope forward are higher than aft. This suggests a different load case, probably for the condition of loading and discharging over the bow. For that case, the bending moment is expected to be lower than in the free-floating condition. This is not further analysed. It confirms however that the obelisks are loaded out and discharged along the centreline over the bow of the barge.

For the analysis of the requirements for the tie-ropes, 6 pre-stressed ropes have been assumed for reasons of symmetry, each rope being comprised of 2 strands. Assuming the distance between ropes and bottom of the barge is 10.00 m, the force per strand is 720/(10 x 12) = 6 t. With a safety factor of 3 to break, 12 x 40 mm ropes would be required to take up the required tension (manila rope or similar). For the bottom planking a safety factor of 2.0 would be sufficient, resulting in a compressive load of 144 t, equally spread. From the relatively light ropes can be concluded that the Egyptian naval architects expected already that the distributions of the weight and the displacement along the length of the hull were close, resulting in relatively small bending moments.

Fig. 14. Global hull loads (author)(19)

Inconsistencies in the layout shown on the relief.

The following particulars of the wall relief cannot be explained from the procedures in this document. It is felt however, that these are of a minor nature.

  1. The leading edge of the sledge of the second obelisk should be at the bottom end and not as drawn at the top end of the obelisk as it had to be discharged over the bow, unless one is shown in loading condition and the other one in discharging condition.
  2. The barge deck cannot show any sheer or camber, it must be flat, no matter how loading and discharging were conducted.
  3. It would be more logical if there were six tie-ropes for reasons of symmetry.

 

Review of earlier publications.

  1. NRC Handelsblad 5.12.1989. Onderwijs. Article by A. Wegener Sleeswijk. Title: Zwaar Transport (Dutch). Mr. Wegener Sleeswijk’s assumption that the obelisks could not be loaded behind each other on the centreline of the barge leads to various assumptions concerning the layout and the structure of the barge, which conflict with the original relief. Estimated weight 374 t each.
  2. Björn Landström: Het Schip, pages 22, 23 (Dutch). Obelisks placed next to each other with both upper ends forward under portals with tie-ropes. This does not seem possible for reasons of loading and discharging operations, and it conflicts with the relief. Some details of construction of hull are given, which have been used for this study.
  3. A Companion to Ancient Egypt. Alan B. Lloyd. Pages 379 and 380. Refers to Stephan Seidlmayer personal communications 2009: confirms that two obelisks were transported on one barge. Refers to Björn Landström and presents 3-D drawing after Björn Landström. Quotes a displacement of loaded barge 7300 t, which is believed to be excessive.
  4. Pharaonic Egypt. reshafim.org.il, Ancient Egypt, river boats. Description of loading and discharging at Roman times. Raises the question how such a transport was done and quotes Harold L. Potts stating: Björn Landström: “My reconstruction (of the barge) hardly even convinces me” (2007).
  5. NOVA online, Mysteries of the Nile, Gifts of the River, 1999. Several solutions shortly discussed, most likely solution was to float the barge under the obelisk, which was laying across a purpose-built canal. This is not in accordance with the relief.
  6. Reginald Engelbach suggests in his Study of the unfinished Obelisk at Aswan (p. 64-65) that the barge was built into an embankment and was completely covered up. The obelisk could be rolled or skidded over the barge over the embankment. By removing the sand step by step from underneath the obelisk, it would be lowered down to the deck (“sand-jacking”). By removing sand from, and out of, the barge, the barge could be lifted off its berth using the barge’s displacement. Discharging could be done in the same way, except the embankment for discharging did not need to be so high. With two obelisks, the method is more laborious, but the principle would be the same. This operation is considered feasible, but complicated.
  7. Seán McGrail, Boats of the World. Paragraph 2.9.3 Heavy Lift Vessels and Towing. Reconstruction drawing of the barge (fig. 2.26 page 44 not in accordance with relief). Quotes other writers. Herodotus describes a way of navigating downstream on the Nile when a 250-kg weight is dropped on the river bed off the stern on a rope. As it slows down the barge, pressure is built up on the rudders making steering possible. However, this effect must be very limited, because both the weight and the rudders are on the stern. The weight prevents the rudders to steer the barge. For this method to work, the barge needs to rotate with the bow into the current and the weight needs to be dropped from the bow. The relief shows, that towage is done from the bow(20), and steering from the stern, which contradicts the description by Herodotus.
  8. The Medieval Nile: Route, Navigation and Landscape in Islamic Egypt. John Cooper. Chapter 8, pages 125-127. Maximum velocity during High Water about 3.9 kn downstream Aswan, about 1.5 kn during Low Water.
  9. H. Breasted, Ancient Records of Egypt, Volume 2, 18th Dynasty, paragraph 328, page 137: barge main particulars estimated by comparison with other more documented vessels. L = 268½ ft and B = 89½ ft.
  10. Flickr, 2008 Cameron Grant. The top of Hatshepsut’s fallen obelisk – Karnak. The obelisks were “moved down the river’s edge during winter at low water and loaded onto a huge barge and they sat until the following summer when the flood came and floated the barge”. No references given. Not until the river’s edge for loading but to inundated valley, but otherwise in agreement with assumptions in this document.

 

Attachment

An artist impression of Hatshepsut’s Heavy-Lift river barge following of the particulars discussed in this document is shown also in fig. 15.

p

References

  1. NRC Handelsblad. Onderwijs. Article by A. Wegener Sleeswijk. Title: Zwaar Transport (Dutch). 12.1989.
  2. Björn Landström: Het Schip (Dutch). 1961.
  3. A Companion to Ancient Egypt. Alan B. Lloyd. 2008
  4. Stephan Seidlmayer. Die Vermessung des Nils im Alten Ägypten. 2011
  5. Encyclopedia debinnenvaart.nl : stevelen. (Dutch)
  6. Pharaonic Egypt. reshafim.org.il, Ancient Egypt, river boats.
  7. Monderson. Temple of Karnak: The Majestic Architecture of Ancient Kemet. 2007
  8. Sad. The River Nile, Geology, Hydrology and Utilisation.
  9. NOVA online, Mysteries of the Nile, Gifts of the River.
  10. Study of the unfinished Obelisk at Aswan, Reginald Engelbach. 1923.
  11. Seán McGrail, Boats of the World. 2001.
  12. The Medieval Nile: Route, Navigation and Landscape in Islamic Egypt. John Cooper. 2014.
  13. H. Breasted, Ancient Records of Egypt. 1906.
  14. Flickr, 2008 Cameron Grant. The top of Hatshepsut’s fallen obelisk – Karnak. 2008.
  15. Google Earth, Coordinates of reference locations and location maps.
  16. The royal ship of Khufu, nautarch.tamu.edu
  17. GL Noble Denton, Guidelines for Marine Transportations, rev. 5
  18. realcedar.com. Groei, eigenschappen en toepassingen van WRC.pdf

 

Footnotes

[1] Stephan Seidlmayer. Personal communication 2009: ”an official of Hatshepsut called Hapusenet states that he was responsible for the transportation of two obelisks on one barge” from: A Companion to Ancient Egypt.

[2] NRC Handelsblad 5.12.1989. Onderwijs. Article by A. Wegener Sleeswijk.

[3] Stephan Seidlmayer. Die Vermessung des Nils im Alten Ägypten. “28 ellen in Elephantine, 14 ellen in Kaïro”.

(28 ellen = approx. 14.60 m, 14 ellen = approx. 7.30 m). Elephantine is an island at Aswan. Ref. comment by Prof. Dr Seidlmayer: “it is true that the rise of the river Nile could reach (in certain periods) 28 cubits on the nilometer scale; however, for not entirely clear reasons, the 0-point of this nilometer scale happened to be far below low water. Actually, the difference in height between high and low water at Aswan used to be under natural conditions around 8 m only”.

[4] F. Said. The River Nile, Geology, Hydrology and Utilisation. P.152, rise above temple floor 0.84 m.

[5] F. Monderson. Temple of Karnak: The Majestic Architecture of Ancient Kemet

[6] Ref. comment by Prof. Dr Seidlmayer: I would like to add the following information: on your Fig 2 also the obelisk of Thutmosis I is visible. Actually, there exists an ancient text of the person who was responsible for building the boat which was used to ship this obelisk and its sibling. Comparing the length of this boat, which is stated in the text, and the length of the existing obelisk it seems clear that the obelisks of Thutmosis Ist were transported in exactly the same way – both on a single boat and arranged in one line along its centre. This seems to have been standard practice. The unusual accomplishment in the case of Hatshepsut’s obelisks seems to have been the unparalleled size and weight of her obelisks rather than the fact that both were loaded on a single vessel.

[7] Study of the unfinished Obelisk at Aswan, Reginald Engelbach. 1923

[8] F. Monderson. Temple of Karnak: The Majestic Architecture of Ancient Kemet. Quote from 1925 max. flood level: “… a zone extending from the floor level to a height varying from about half a meter to several meters….”. p. 160

[9] NRC Handelsblad 5.12.1989. Onderwijs. Article by A. Wegener Sleeswijk

[10] J. H. Breasted, Ancient Records of Egypt, Volume 2, par. 328, 329, pages 137, 138.

[11] The Medieval Nile: Route, Navigation and Landscape in Islamic Egypt. John Cooper. Chapter 8, page 125-127. A maximum of 3.9 knots has been quoted in the 19th century.

[12] F. Monderson. Temple of Karnak: The Majestic Architecture of Ancient Kemet. P.160

[13] Up to about the year 1900 cargo vessels navigated down the river Rhine without using engines or sails making sometimes a speed of 4 to 5 km/hr in relation to the water surface using the same principle. Ref. www.debinnenvaart.nl : stevelen (Dutch). For steering they used oars and rudders. Rafts with sometimes 51.000 m3 wood were floated down the Rhine in the 17th and 18th centuries using the same principle. It is known that during Roman times this was done as well. The down-hill gradient of the Nile is less than of the Rhine, but it is considered enough to have a similar effect during periods of high water.

[14] Ref. comment by Prof. Dr Seidlmayer: “it is true that the rise of the river Nile could reach (in certain periods) 28 cubits on the nilometer scale; however, for not entirely clear reasons, the 0-point of this nilometer scale happened to be far below low water. Actually, the difference in height between high and low water at Aswan used to be under natural conditions around 8 m only”.

[15] GL Noble Denton, Guidelines for Marine Transportations, Section 10, Clause 10.2.4. Normally GMo>1.00 m

[16] J. H. Breasted, Ancient Records of Egypt, par 329, p.137

[17] The royal ship of Khufu, www.nautarch.tamu.edu

[18] www.realcedar.com. Groei, eigenschappen en toepassingen van WRC.pdf

[19] Δ is displacement and G is weight. The difference Δ – G = Q (ton) depicts the longitudinally distributed load on the hull. From this curve, the distribution of the bending moments can be calculated by two times integration, resulting in the maximum bending moment of 720 tm at ½L in this case

[20] In The Netherlands, sailing river barges without engines used to navigate backwards through narrow bridges and under low railway bridges by deploying an anchor chain from the bow up-steam and steering with the rudder after they lowered their sails and the mast. This was still done during the first half of the 20th century.

T.  Hoogeveen, naval architect and marine heavy lift specialist.

Drawings prepared by author, except where noted.

e-mail: marine-consultant@kpnmail.nl

October 1, 2016

Fig. 15. 3-D impression of queen Hatshepsut’s Heavy-Lift river barge (author)

Queen Hatshepsut’s Heavy-Lift river barge

The stone mason knew best

Rev. B,dd. July 1, 2018

Abstract

Key questions: How accurate was the relief on the wall of Hatshepsut’s temple at Deir el-Bahri? How could the Heavy-Lift barge (HL-barge) on the relief be loaded and discharged, and how was the structural integrity of the barge ensured during all stages of the operation? How did the tow navigate down-stream on the river Nile? This report presents possible answers from an operational point of view, concluding that the stone mason knew best.

Note: to read a footnote please click on the reference number between brackets behind each word. There are 20 footnotes in total.

p

Introduction

The description in this document follows as closely as possible the layout and the details shown on the reliefs on the wall of Hatshepsut’s temple at Deir el-Bahri.

To understand how the HL-barge could have been constructed and how it could have been operated in about 1470 BC, it needs to be established first how it was discharged, how it navigated on the river Nile and how it was loaded. It is evident that before the barge was loaded, it had to be ensured that it could be discharged. The barge cannot be considered on its own as the transport was a “door-to-door” operation, a marine heavy lift project. There is no information available from original sources to date. Therefore, heavy lift practice of the present times is applied with only the assistance of contemporary tools and practices such as sledges, rollers, spades, levers, ropes, and manual force from a very large number of labourers. Iron was not available at the time. Different water levels of the river Nile over one year must have been used to the advantage of the operation. It would not have been possible without the help of water.

The loaded barge (1) has been depicted as a relief in a sailing condition on the river Nile on the wall of Hatshepsut’s temple at Deir el-Bahri. The relief has been shown in fig.1. The bow points to the right.

Fig. 1. Relief 1 at Deir El-Bahri
(GNU Free Documentation License 1.2)

The waterlevel of the river Nile is very high over a period of about 3 months, especially near Aswan. The high-water period runs from August till November, during which the currents are at their highest, to the advantage of the navigation on the river. The river floods over its banks during that period. During the other months, the level is low and the depth of the river is shallow especially at the banks. The difference between high water (HW) and low water (LW) could be as much as 8 m at Aswan and 7.5 m near Cairo(2),(3). The rise at Luxor is about 0.84 m above the temple floor according to F. Said(4)  and 0.50 m, or several meters more, according to of F. Monderson(5) on the basis of salt deposits in the stone at Karnak. The latter has been selected for this study assuming regular decline between Aswan and Karnak.

The document presents also a few new considerations for the operation. The numbers quoted in this document are based on basic calculations, and are meant to support the conclusions, demonstrating the feasibility of the transport method as presented in this document. A typical construction of the barge needed to be worked out for this study considering Egyptian practice at the time.

p

Conclusions

It can be concluded from this study that the layout and the details, shown on the wall reliefs of the heavy lift barge in the temple of Deir el-Bahri, are technically correct, and that the transport could have been conducted in the manner described in this document with sufficient safety to the cargo, the craft and the crew. The reliefs are realistic. It cannot be proven that the operation was actually done in this way as no records have been found up to the present day.

The Heavy-Lift river barge of queen Hatshepsut was developed from similar smaller vessels used for smaller obelisks(6). It was destined for one type of cargo only and for one loading condition (equally distributed along the centre line), possibly only for one or two voyages. After completion of the first voyage, the vessel could be dismantled for the return voyage, or stored for future use with similar cargo.

The Egyptian naval architects had an excellent understanding of Archimedes Law, 1000 years before Archimedes.

This transport operation on the river Nile can better be defined as “controlled drifting”.

p

Acknowledgements

The author likes to thank Prof. Dr Stephan J. Seidlmayer, professor for Egyptology at the Freie Universität Berlin and Director of the Deutsches Arch. Institut – Dept. Kairo, who has reviewed this article, and who has made a few very valuable comments and suggestions, which have been included in this document. Further the author likes to thank Mr John Evans, retired director of AMT Ltd of Farnham, UK, a marine heavy lift transportation company, for his review of the practicality of the operation and of the correct use of the English language.

 

Description of the cargo

The cargo is comprised of two flat laying obelisks with the following particulars, each: Length 28.50 m. Base dimensions approx. 2.50 x 2.50 m. Weight estimated 375 t. The base dimensions have been scaled down from photographs, the weight comes from various sources. Fig. 2 shows the obelisk of Hatshepsut (right) at the beginning of the twentieth century.

Fig. 2. Obelisks of Tutmosis I and Hatshepsut (right) at Karnak. From: The Study of the unfinished obelisk at Aswan(7)

Coordinates of referenced locations.

Karnak temple complex at Luxor, Google Earth       25° 42ʼ 01.16” N     32° 38ʼ 23.28” E

Quarry at Aswan, Google Earth                                  24° 04ʼ 36.37” N     32° 53ʼ 43.29” E

The route has been shown on the map in fig. 3. Distance 213 km.

Fig. 3. Section of the Nile from Aswan to Luxor (Image©2016 DigitalGlobe)

 

Discharging at Karnak.

The Karnak temple complex is orientated almost parallel to the river Nile. Therefore, the obelisks could be moved on rollers, or skidded along a prepared track straight from the river along an unobstructed route perpendicular to the centre line of the temple complex, to their final positions. See fig. 4. From the contour lines at Karnak, obtained from Google Earth, and the rise of the water level during a period of high water, it is clear that during HW the site of the temple complex flooded every year(8). This allowed landing of heavy weights at the site during construction.

Fig.4. Track at Karnak (Image©2016 DigitalGlobe)

During the period of low river level (LW), a bed would have been prepared near the construction site at Karnak to allow mooring of the barge when the river Nile is at its highest level (HW). When the water level dropped after a few weeks, the barge settled on the prepared bed, on which it could stay for a prolonged period of time when the area dried out completely. The bed was constructed such that it supported the structure of the hull of the barge over her full length and possibly part of the bow. The structural arrangement inside the barge allowed a good load distribution, when there was no water above the bed. A 7-9 m high dam with a length of about 70 m was required between the barge and the destination of the obelisks between the fourth and the fifth pylon at Karnak. See fig. 5.

Fig. 5. Discharging at Karnak (author)

The first obelisk was moved with the top in forward position and was rolled or skidded, and parked past its final position prior to upending. The second one was rolled or skidded with its bottom end forward and was probably erected first. For that reason, the obelisks were placed on board on the centreline of the barge, one behind the other in opposite directions. The foc’sle had to allow passage over the bow. The tie-ropes of the hull were running over portals of varying heights, forward ones being the highest to give extra support to the bow. The ropes split sideways to port and starboard allowing passage of the obelisks. The bowsprit was structurally extended down to the base line of the hull, such that the contact with the prepared bed was extended forward as far as possible.

The elevated land route had to be prepared for the haulage of both obelisks. Two rows of blocks were providing the foundations for the route. The tops of the blocks had to be in one level plane. Wooden rails were laid on top of the blocks for easier rolling or skidding. They provided the surface on which wooden rollers or skids were running underneath a sledge fixed to the obelisk. As the obelisk was moving forward, rollers and rails appeared from behind and had to be carried forward. This means that, if well planned, the skid tracks on the barge and the skid tracks on land were in one horizontal level plane and only friction needed to be overcome. After the cargo had been transferred ashore, the barge could be disassembled in the dry and the materials could be shipped back up-river using normal cargo vessels. The light weight of the barge has been estimated as 1250 t. The cargo vessels could sail up-stream using sails and oars during the LW period when the currents in the river were small and most of the time northerly winds were blowing. Thus, the HL-barge could be used again for two more obelisks if required.

With the delivery of the two obelisks on the shore, the marine operation was completed.

p

Upending.

Although the upending of the obelisks is no part of the marine operation, some thoughts have been dedicated to the procedure for upending, because if the obelisks were delivered in a condition from which they could not be upended, the marine operation would be useless. The rise of the water was used also to bring the obelisks in an elevated position, necessary for upending. So, no uphill skidding or rolling was required.  The description refers to the operation for the second obelisk. See fig. 6.

Fig. 6. Upending of the second obelisk (author).

It should be noted that this procedure is an outline only. The second obelisk was moved forward until the centre of gravity was past a pivot point. The pivot point was located on a temporary retaining wall, set at a precise distance to the base. Over the distance A-B, the sledge was firmly pulled against the obelisk by tensioning ropes such that the obelisk does not slide when the sliding force is 320 t. By digging away the embankment in front of the retaining wall, the obelisk rotated in a controlled way into position 2 by gravity until it hit a stopper down below. From position 2 to position 3, the obelisk had to be pulled by significant large number of hands, with holdback from another large number of hands.

p

Navigating on the river Nile.

The relief in fig. 7 shows 27 tugboats and 3 pilot boats tied to the bow of the HL-barge at different locations operating in three rows. Sᴓlver(9) has estimated that the joint pulling power of 30 towing rowboats would be equivalent to the bollard pull of a 50-horsepower tugboat. It has been assumed that this is at zero forward speed. Such a tug can provide about 500 kg pulling force. A major part of this force may have been used for steering the bow in the right direction avoiding obstructions en route, as well as for keeping ahead of the barge when some speed is gained from the “downhill effect” of the water surface of the river (see next paragraphs for further explanation). Keeping out of each other’s way may also have taken some effort. Consequently, there will not have been much left to give the tow some forward speed. This kind of navigation on the river can be described better as controlled drifting.  This also means that there was no room for mistakes by the master of the tow or by the pilots. It would have been nearly impossible to make corrections in time.

On the bow of the leading rowboats, pilots were present to sound the water depth with long poles and to give directions accordingly. There were three escort rowboats (below on the left), which were not connected to the barge, for performing religious ceremonies. A smaller rowboat is also shown on the relief (standby function?). It is expected that the master was on the foc’sle of the HL-barge, which would give him a good overview. Three high level dignitaries were also on the foc’sle (10).

Fig.7. Relief 2 at Deir El-Bahri
(GNU Free Documentation License 1.2)

The shipment took place during a period of high water on the Nile. The current velocity was about 3.3 knots(11). Based on information from Google Earth and the rise of the river at high water, the difference in water levels between Aswan at the quarry and at Karnak (assuming 2.50 m water over the temple floor ref. Monderson(12)) could be as much as 21 m. The sailing distance is 213 km. This means that the river runs downhill with a gradient of 0.010% on the average. This results in a downhill component of the weight of about 200 kg, in assistance to the oarsmen on the row boats(13). It is estimated that this added 1.1 knots to the speed of the barge in relation to the water, and enough to give some pressure on the rudders. The presence of four big oars for steering suggests this effect to have been known at the time. The row boats must be able to move faster than 1.1 knots, otherwise they will be overrun by the barge.

The speed of the convoy in relation to the land was probably 4.4 knots or 8.1 km/hr. Sailing at night may have been considered too risky, and people needed rest and food. When sailing is restricted to 8 hours per day, excluding (un)mooring, the route was covered in 3½ days. Leaving and entering ports especially may have been nerve-racking for the pilots, as well as manoeuvring through the rapids near Aswan, not to mention negotiating bends in the river. Maybe marker buoys were installed at critical locations. It is likely that special procedures were developed for such operations, written down in a book with instructions (operations manual). Apparently, the Egyptian navigators managed to do this without damage to the cargo or the barge, an impressive achievement in its own right!

 

Loading at Aswan

The quarry was located at Aswan. An unfinished obelisk can still be found in that location confirming the orientation of the obelisks (see fig. 8). The obelisks were orientated at an angle of about 60 degrees to the river. In line with the orientation, a prepared bed was constructed a bit further into the valley, which only flooded when the waterlevel in the river was high. It took about 7 months to cut an obelisk out of the bedrock. During the same period, the HL-barge was constructed on the prepared bed when dry, assembled from prefabricated timbers.

Fig. 8. Quarry at Aswan, unfinished obelisk (Image©2016 DigitalGlobe)

Upon load-out, the track from the quarry was extended via a dam over the barge deck, after which the first obelisk was moved on board before the start of the period of high water, immediately followed by the second one. See fig. 9. The particulars of the land route were similar to the land route at Karnak, except it was running downhill at a small gradient (probably 2.4%) over a distance of about 400 m in a straight line. The arrow on fig. 10 shows the probable location of the prepared bed. Once on board, sea fastenings were installed. Water tightness of the hull could to be confirmed during the rising of the waterlevel in the Nile. The raised foc’sle had to be temporarily removed to allow access along the centreline of the barge.

Fig. 9. Cross section along quarry towards river Nile, along route A-A (author)

Fig. 10. Route down to the river, via flooded valley at HW (Image©2016 DigitalGlobe). Dotted line shows overland route of obelisk via a dam to the possible loading location of the barge.

Fig. 10 shows the sailing route from the quarry to the river through the flooded valley. The rise of the river at Aswan was about 8 m(14) under average conditions. The fleet of rowboats and assisting vessels were mobilised in time for the float-out. Prior to joining the river, the barge was turned around bow first in the flooded valley in quiet waters. Joining the river, with a current velocity of 3.3 knots, required a lot of effort from the oarsmen, and it demanded experienced navigators and pilots.

 

Hydrostatic properties

A lines plan of the barge has been made (see fig. 11), which is the basis of the hydrostatic calculations in this document. The lines of the frames only are shown below, but a full lines plan was made for the calculations. There are two planes of symmetry, centre line and mid length, hence only one quarter of the hull is shown.

Fig.11. Frames (author)

The barge required to be shallow draft, with a curved hull and with extra deck strength along its centre line. The barge had to carry the cargo and its own light weight, and it had to have sufficient stability during the river transfer. The stability during loading and discharging was no problem as the hull was be supported on a prepared bed. To judge the free-floating stability, present day criteria for heavy lift transportation have been applied for reference.

The following is a summary of the hydrostatic calculations of the hull.

Barge particulars for this study

Length (L) over hull:  81.80 m Length(wl) : 65.30 m
Beam (B) over hull:  27.30 m Beam(wl) : 25.50 m
Depth to maindeck:   5.00 m Light draft:    1.75 m
Loaded draft:   2.50 m Block coeff. δ: 0.480
Displacement:   2000 t Main frame coeff. β: 0.735
Light weight, est:   1250 t

Stability, expressed in metacentric height (GM) in meters.

Metacentre above keel                KM = BM + BK = 23.01 + 1.50 = 24.51 m

Centre of gravity above keel       KG = 4.81 m

GM = 24.51 – 4.81 = 19.70 m(15) (> 1.00 m)

The stability appears to be more than sufficient, even for present day criteria.

 

Hull planking

For the hull planking sycamore(16) wood was used, found locally in Egypt. The reason was probably cost, as wood imported from the Lebanon for shipbuilding would have made the barge very expen-sive for only one or two shipments. The type of wood used required a special hull design and building technique as it was only available in relatively short lengths. The wooden sycamore blocks were shaped to form a segmental arch similar to a segmental stone arch in old churches, but upside down. The seams were compressed by the water pressure in transverse direction and by the hogging (middle-up) condition of the loaded barge in the other direction. Papyrus sealed the seams from the inside. The blocks were “pinned” together (mortise-and-tenon joints). Traditional Egyptian boats do normally not require extensive framing for the strength of the hull considerations. Framing was required for load distribution under vertical columns shown below (see fig. 12 and fig. 14). The shape of the arch was probably elliptical, without any flat surfaces. For the drawing of the construction of the hull details of the Khufu ship(17) have been studied. The special arrangement of the hull planking caused a certain flexibility in the overall structure of the barge. This would have resulted in a good distribution of the internal loads introduced by the cargo, thus avoiding load concentrations anywhere in the hull. The water tightness was ensured by caulking with papyrus. Only the obelisks were extremely stiff against bending, they kept the barge straight under load.

 

Framing arrangement

The relief shows the three deck beams above each other protruding the hull at specific locations (see fig. 12 below). Portals for tie-ropes have been shown every two frames at equal distances. It has been assumed that underdeck the same supports are extended down to the bottom of the hull. Three deck beams are connected with blocks such that each individual beam bends with the same deflection under load. Wooden spacer blocks assure that this is the case. This means that the stiffness of three beams can be added up for combined strength. The beams were made from cedar wood from the Lebanon, as well as the longitudinal deck planking where longitudinal strength was very important. Very long beams would normally be available from the Lebanon, although figure 12 shows three sections instead. The transverse deck beams also prevented the hull from deflecting outward and keeping the hull in shape, thus adding to the hull overall strength as well as carrying the deck load.

Fig. 12. Main frame (author)
Left: grounded during load-out, right: afloat

For argument sake, the structure of the main frame of figure 12 has been analysed for strength, the results of which have been shown in figure 13. The light weight of the frames has been added to the weight of the obelisks, and the total has been factored with 10%. The beams are comprised of cedar wood with an allowable stress(18) in bending 34 N/mm2, allowable in compression 27 N/m and an allowable shear of 5.6 N/mm2. If the three beams are deflecting together in the same way, their joint section modulus can be expressed as follows:

S = 3 x 500²x 300 / 6 = 37.5 x 106 mm3

Bending stress                  σ = 89 x 107 / (37.5 x 106) = 23.7 N/mm2  (max. 34 N/mm2)

Shear stress                       τ = 1.5 x 20.6 x 104 /(3 x 500 x 300) = 0.7 N/mm2 (max. 5.6 N/mm2)

When the barge is on the prepared bed, the full weight goes down one column and the compressive stress can be calculated as follows:

Compressive stress         σ = 20.6 x 104 / (500 x 300) = 1.4 N/mm2  (max. 27 N/mm2)

It can also be noted that intermediate frames are required as the main frames alone cannot take the load. The intermediate frames are not visible on the relief in the temple of Deir el-Bahri.

Fig. 13. Strength analysis of assumed main frame structure (author)

The obelisks were supported on at least 15 main frames and 14 intermediate frames in direct load transfer. The barge was at its strongest over the centre line. Fig. 14 shows the layout of the intermediate frames.

Fig. 14. Intermediate frame (author)
Left: grounded during load-out, right: afloat

Longitudinal strength.

There are no waves on the Nile during navigation, so any wave bending moments could be dis-regarded for the design by the Egyptian naval architects. When only one loading condition needs to be considered, it is theoretically possible to design the lines of the hull such that the distribution of the displacement along the hull is exactly the same as the distribution of the weights in free floating condition, in which case there would be no bending moments in the hull structure at all. In practice this is difficult to achieve and small bending moments will occur (see fig. 14). These bending moments must be taken up by the hull, for this vessel by tie-ropes in tension and bottom in compression. For tensioning up, the ropes must each be double and they must be comprised of combined right-hand lay and left-hand lay in each strand and all running in the same direction. The tie-ropes are also required during load-out when the barge is positioned on the prepared bed, when only the bow is (partly) unsupported. Alternatively, the bow could be further supported on an extension of the prepared bed.

The ends of the barge must be shipshape to ensure water tightness and low resistance. As the deck load is more or less equally distributed over the length, and most of the displacement is concentrated over about 70% of the waterline length, the barge will resume a hogging condition (middle up) and the obelisks will not be evenly supported. This can be corrected by tensioning the tie-ropes to resume a straight line for the deck. The weight of the cargo is transferred to the water via 29 vertical main frames, directly under the load at the locations shown on the relief by the beams protruding the hull in the side, and at intermediate frames. As long as the barge is supported on one of the prepared beds, the vertical load is transferred directly into the supports on the bed at the vertical frames and no longitudinal bending of the hull takes place (except some at the bow during loading and discharging).

The wall relief shows clearly that the obelisks have been loaded with their lower ends towards the middle of the barge (fig. 1). This is where the barge has most of its strength in bending and most of its displacement to take vertical loads. Fig. 14 shows the overall distribution of the loads in the hull resulting from the weights and the displacement (Δ and G), and the bending moments (curve M). The bending moments in the hull have been calculated along the length of the hull. This results in a maximum bending moment of 720 tm at mid length. If the relief is regarded closely (fig. 1), it can be seen that the portals supporting the tie-rope forward are higher than aft. This suggests a different load case, probably for the condition of loading and discharging over the bow. For that case, the bending moment is expected to be lower than in the free-floating condition. This is not further analysed. It confirms however that the obelisks are loaded out and discharged along the centreline over the bow of the barge.

For the analysis of the requirements for the tie-ropes, 6 pre-stressed ropes have been assumed for reasons of symmetry, each rope being comprised of 2 strands. Assuming the distance between ropes and bottom of the barge is 10.00 m, the force per strand is 720/(10 x 12) = 6 t. With a safety factor of 3 to break, 12 x 40 mm ropes would be required to take up the required tension (manila rope or similar). For the bottom planking a safety factor of 2.0 would be sufficient, resulting in a compressive load of 144 t, equally spread. From the relatively light ropes can be concluded that the Egyptian naval architects expected already that the distributions of the weight and the displacement along the length of the hull were close, resulting in relatively small bending moments.

Fig. 14. Global hull loads (author)(19)

Inconsistencies in the layout shown on the relief.

The following particulars of the wall relief cannot be explained from the procedures in this document. It is felt however, that these are of a minor nature.

  1. The leading edge of the sledge of the second obelisk should be at the bottom end and not as drawn at the top end of the obelisk as it had to be discharged over the bow, unless one is shown in loading condition and the other one in discharging condition.
  2. The barge deck cannot show any sheer or camber, it must be flat, no matter how loading and discharging were conducted.
  3. It would be more logical if there were six tie-ropes for reasons of symmetry.

 

Review of earlier publications.

  1. NRC Handelsblad 5.12.1989. Onderwijs. Article by A. Wegener Sleeswijk. Title: Zwaar Transport (Dutch). Mr. Wegener Sleeswijk’s assumption that the obelisks could not be loaded behind each other on the centreline of the barge leads to various assumptions concerning the layout and the structure of the barge, which conflict with the original relief. Estimated weight 374 t each.
  2. Björn Landström: Het Schip, pages 22, 23 (Dutch). Obelisks placed next to each other with both upper ends forward under portals with tie-ropes. This does not seem possible for reasons of loading and discharging operations, and it conflicts with the relief. Some details of construction of hull are given, which have been used for this study.
  3. A Companion to Ancient Egypt. Alan B. Lloyd. Pages 379 and 380. Refers to Stephan Seidlmayer personal communications 2009: confirms that two obelisks were transported on one barge. Refers to Björn Landström and presents 3-D drawing after Björn Landström. Quotes a displacement of loaded barge 7300 t, which is believed to be excessive.
  4. Pharaonic Egypt. reshafim.org.il, Ancient Egypt, river boats. Description of loading and discharging at Roman times. Raises the question how such a transport was done and quotes Harold L. Potts stating: Björn Landström: “My reconstruction (of the barge) hardly even convinces me” (2007).
  5. NOVA online, Mysteries of the Nile, Gifts of the River, 1999. Several solutions shortly discussed, most likely solution was to float the barge under the obelisk, which was laying across a purpose-built canal. This is not in accordance with the relief.
  6. Reginald Engelbach suggests in his Study of the unfinished Obelisk at Aswan (p. 64-65) that the barge was built into an embankment and was completely covered up. The obelisk could be rolled or skidded over the barge over the embankment. By removing the sand step by step from underneath the obelisk, it would be lowered down to the deck (“sand-jacking”). By removing sand from, and out of, the barge, the barge could be lifted off its berth using the barge’s displacement. Discharging could be done in the same way, except the embankment for discharging did not need to be so high. With two obelisks, the method is more laborious, but the principle would be the same. This operation is considered feasible, but complicated.
  7. Seán McGrail, Boats of the World. Paragraph 2.9.3 Heavy Lift Vessels and Towing. Reconstruction drawing of the barge (fig. 2.26 page 44 not in accordance with relief). Quotes other writers. Herodotus describes a way of navigating downstream on the Nile when a 250-kg weight is dropped on the river bed off the stern on a rope. As it slows down the barge, pressure is built up on the rudders making steering possible. However, this effect must be very limited, because both the weight and the rudders are on the stern. The weight prevents the rudders to steer the barge. For this method to work, the barge needs to rotate with the bow into the current and the weight needs to be dropped from the bow. The relief shows, that towage is done from the bow(20), and steering from the stern, which contradicts the description by Herodotus.
  8. The Medieval Nile: Route, Navigation and Landscape in Islamic Egypt. John Cooper. Chapter 8, pages 125-127. Maximum velocity during High Water about 3.9 kn downstream Aswan, about 1.5 kn during Low Water.
  9. H. Breasted, Ancient Records of Egypt, Volume 2, 18th Dynasty, paragraph 328, page 137: barge main particulars estimated by comparison with other more documented vessels. L = 268½ ft and B = 89½ ft.
  10. Flickr, 2008 Cameron Grant. The top of Hatshepsut’s fallen obelisk – Karnak. The obelisks were “moved down the river’s edge during winter at low water and loaded onto a huge barge and they sat until the following summer when the flood came and floated the barge”. No references given. Not until the river’s edge for loading but to inundated valley, but otherwise in agreement with assumptions in this document.

 

Attachment

An artist impression of Hatshepsut’s Heavy-Lift river barge following of the particulars discussed in this document is shown also in fig. 15.

p

References

  1. NRC Handelsblad. Onderwijs. Article by A. Wegener Sleeswijk. Title: Zwaar Transport (Dutch). 12.1989.
  2. Björn Landström: Het Schip (Dutch). 1961.
  3. A Companion to Ancient Egypt. Alan B. Lloyd. 2008
  4. Stephan Seidlmayer. Die Vermessung des Nils im Alten Ägypten. 2011
  5. Encyclopedia debinnenvaart.nl : stevelen. (Dutch)
  6. Pharaonic Egypt. reshafim.org.il, Ancient Egypt, river boats.
  7. Monderson. Temple of Karnak: The Majestic Architecture of Ancient Kemet. 2007
  8. Sad. The River Nile, Geology, Hydrology and Utilisation.
  9. NOVA online, Mysteries of the Nile, Gifts of the River.
  10. Study of the unfinished Obelisk at Aswan, Reginald Engelbach. 1923.
  11. Seán McGrail, Boats of the World. 2001.
  12. The Medieval Nile: Route, Navigation and Landscape in Islamic Egypt. John Cooper. 2014.
  13. H. Breasted, Ancient Records of Egypt. 1906.
  14. Flickr, 2008 Cameron Grant. The top of Hatshepsut’s fallen obelisk – Karnak. 2008.
  15. Google Earth, Coordinates of reference locations and location maps.
  16. The royal ship of Khufu, nautarch.tamu.edu
  17. GL Noble Denton, Guidelines for Marine Transportations, rev. 5
  18. realcedar.com. Groei, eigenschappen en toepassingen van WRC.pdf

 

Footnotes

[1] Stephan Seidlmayer. Personal communication 2009: ”an official of Hatshepsut called Hapusenet states that he was responsible for the transportation of two obelisks on one barge” from: A Companion to Ancient Egypt.

[2] NRC Handelsblad 5.12.1989. Onderwijs. Article by A. Wegener Sleeswijk.

[3] Stephan Seidlmayer. Die Vermessung des Nils im Alten Ägypten. “28 ellen in Elephantine, 14 ellen in Kaïro”.

(28 ellen = approx. 14.60 m, 14 ellen = approx. 7.30 m). Elephantine is an island at Aswan. Ref. comment by Prof. Dr Seidlmayer: “it is true that the rise of the river Nile could reach (in certain periods) 28 cubits on the nilometer scale; however, for not entirely clear reasons, the 0-point of this nilometer scale happened to be far below low water. Actually, the difference in height between high and low water at Aswan used to be under natural conditions around 8 m only”.

[4] F. Said. The River Nile, Geology, Hydrology and Utilisation. P.152, rise above temple floor 0.84 m.

[5] F. Monderson. Temple of Karnak: The Majestic Architecture of Ancient Kemet

[6] Ref. comment by Prof. Dr Seidlmayer: I would like to add the following information: on your Fig 2 also the obelisk of Thutmosis I is visible. Actually, there exists an ancient text of the person who was responsible for building the boat which was used to ship this obelisk and its sibling. Comparing the length of this boat, which is stated in the text, and the length of the existing obelisk it seems clear that the obelisks of Thutmosis Ist were transported in exactly the same way – both on a single boat and arranged in one line along its centre. This seems to have been standard practice. The unusual accomplishment in the case of Hatshepsut’s obelisks seems to have been the unparalleled size and weight of her obelisks rather than the fact that both were loaded on a single vessel.

[7] Study of the unfinished Obelisk at Aswan, Reginald Engelbach. 1923

[8] F. Monderson. Temple of Karnak: The Majestic Architecture of Ancient Kemet. Quote from 1925 max. flood level: “… a zone extending from the floor level to a height varying from about half a meter to several meters….”. p. 160

[9] NRC Handelsblad 5.12.1989. Onderwijs. Article by A. Wegener Sleeswijk

[10] J. H. Breasted, Ancient Records of Egypt, Volume 2, par. 328, 329, pages 137, 138.

[11] The Medieval Nile: Route, Navigation and Landscape in Islamic Egypt. John Cooper. Chapter 8, page 125-127. A maximum of 3.9 knots has been quoted in the 19th century.

[12] F. Monderson. Temple of Karnak: The Majestic Architecture of Ancient Kemet. P.160

[13] Up to about the year 1900 cargo vessels navigated down the river Rhine without using engines or sails making sometimes a speed of 4 to 5 km/hr in relation to the water surface using the same principle. Ref. www.debinnenvaart.nl : stevelen (Dutch). For steering they used oars and rudders. Rafts with sometimes 51.000 m3 wood were floated down the Rhine in the 17th and 18th centuries using the same principle. It is known that during Roman times this was done as well. The down-hill gradient of the Nile is less than of the Rhine, but it is considered enough to have a similar effect during periods of high water.

[14] Ref. comment by Prof. Dr Seidlmayer: “it is true that the rise of the river Nile could reach (in certain periods) 28 cubits on the nilometer scale; however, for not entirely clear reasons, the 0-point of this nilometer scale happened to be far below low water. Actually, the difference in height between high and low water at Aswan used to be under natural conditions around 8 m only”.

[15] GL Noble Denton, Guidelines for Marine Transportations, Section 10, Clause 10.2.4. Normally GMo>1.00 m

[16] J. H. Breasted, Ancient Records of Egypt, par 329, p.137

[17] The royal ship of Khufu, www.nautarch.tamu.edu

[18] www.realcedar.com. Groei, eigenschappen en toepassingen van WRC.pdf

[19] Δ is displacement and G is weight. The difference Δ – G = Q (ton) depicts the longitudinally distributed load on the hull. From this curve, the distribution of the bending moments can be calculated by two times integration, resulting in the maximum bending moment of 720 tm at ½L in this case

[20] In The Netherlands, sailing river barges without engines used to navigate backwards through narrow bridges and under low railway bridges by deploying an anchor chain from the bow up-steam and steering with the rudder after they lowered their sails and the mast. This was still done during the first half of the 20th century.

T.  Hoogeveen, naval architect and marine heavy lift specialist.

Drawings prepared by author, except where noted.

e-mail: marine-consultant@kpnmail.nl

October 1, 2016

Fig. 15. 3-D impression of queen Hatshepsut’s Heavy-Lift river barge (author)

Queen Hatshepsut’s Heavy-Lift river barge

The stone mason knew best

Rev. B,dd. July 1, 2018

Abstract

Key questions: How accurate was the relief on the wall of Hatshepsut’s temple at Deir el-Bahri? How could the Heavy-Lift barge (HL-barge) on the relief be loaded and discharged, and how was the structural integrity of the barge ensured during all stages of the operation? How did the tow navigate down-stream on the river Nile? This report presents possible answers from an operational point of view, concluding that the stone mason knew best.

Note: to read a footnote please click on the reference number between brackets behind each word. There are 20 footnotes in total.

p

Introduction

The description in this document follows as closely as possible the layout and the details shown on the reliefs on the wall of Hatshepsut’s temple at Deir el-Bahri.

To understand how the HL-barge could have been constructed and how it could have been operated in about 1470 BC, it needs to be established first how it was discharged, how it navigated on the river Nile and how it was loaded. It is evident that before the barge was loaded, it had to be ensured that it could be discharged. The barge cannot be considered on its own as the transport was a “door-to-door” operation, a marine heavy lift project. There is no information available from original sources to date. Therefore, heavy lift practice of the present times is applied with only the assistance of contemporary tools and practices such as sledges, rollers, spades, levers, ropes, and manual force from a very large number of labourers. Iron was not available at the time. Different water levels of the river Nile over one year must have been used to the advantage of the operation. It would not have been possible without the help of water.

The loaded barge (1) has been depicted as a relief in a sailing condition on the river Nile on the wall of Hatshepsut’s temple at Deir el-Bahri. The relief has been shown in fig.1. The bow points to the right.

Fig. 1. Relief 1 at Deir El-Bahri
(GNU Free Documentation License 1.2)

The waterlevel of the river Nile is very high over a period of about 3 months, especially near Aswan. The high-water period runs from August till November, during which the currents are at their highest, to the advantage of the navigation on the river. The river floods over its banks during that period. During the other months, the level is low and the depth of the river is shallow especially at the banks. The difference between high water (HW) and low water (LW) could be as much as 8 m at Aswan and 7.5 m near Cairo(2),(3). The rise at Luxor is about 0.84 m above the temple floor according to F. Said(4)  and 0.50 m, or several meters more, according to of F. Monderson(5) on the basis of salt deposits in the stone at Karnak. The latter has been selected for this study assuming regular decline between Aswan and Karnak.

The document presents also a few new considerations for the operation. The numbers quoted in this document are based on basic calculations, and are meant to support the conclusions, demonstrating the feasibility of the transport method as presented in this document. A typical construction of the barge needed to be worked out for this study considering Egyptian practice at the time.

p

Conclusions

It can be concluded from this study that the layout and the details, shown on the wall reliefs of the heavy lift barge in the temple of Deir el-Bahri, are technically correct, and that the transport could have been conducted in the manner described in this document with sufficient safety to the cargo, the craft and the crew. The reliefs are realistic. It cannot be proven that the operation was actually done in this way as no records have been found up to the present day.

The Heavy-Lift river barge of queen Hatshepsut was developed from similar smaller vessels used for smaller obelisks(6). It was destined for one type of cargo only and for one loading condition (equally distributed along the centre line), possibly only for one or two voyages. After completion of the first voyage, the vessel could be dismantled for the return voyage, or stored for future use with similar cargo.

The Egyptian naval architects had an excellent understanding of Archimedes Law, 1000 years before Archimedes.

This transport operation on the river Nile can better be defined as “controlled drifting”.

p

Acknowledgements

The author likes to thank Prof. Dr Stephan J. Seidlmayer, professor for Egyptology at the Freie Universität Berlin and Director of the Deutsches Arch. Institut – Dept. Kairo, who has reviewed this article, and who has made a few very valuable comments and suggestions, which have been included in this document. Further the author likes to thank Mr John Evans, retired director of AMT Ltd of Farnham, UK, a marine heavy lift transportation company, for his review of the practicality of the operation and of the correct use of the English language.

 

Description of the cargo

The cargo is comprised of two flat laying obelisks with the following particulars, each: Length 28.50 m. Base dimensions approx. 2.50 x 2.50 m. Weight estimated 375 t. The base dimensions have been scaled down from photographs, the weight comes from various sources. Fig. 2 shows the obelisk of Hatshepsut (right) at the beginning of the twentieth century.

Fig. 2. Obelisks of Tutmosis I and Hatshepsut (right) at Karnak. From: The Study of the unfinished obelisk at Aswan(7)

Coordinates of referenced locations.

Karnak temple complex at Luxor, Google Earth       25° 42ʼ 01.16” N     32° 38ʼ 23.28” E

Quarry at Aswan, Google Earth                                  24° 04ʼ 36.37” N     32° 53ʼ 43.29” E

The route has been shown on the map in fig. 3. Distance 213 km.

Fig. 3. Section of the Nile from Aswan to Luxor (Image©2016 DigitalGlobe)

 

Discharging at Karnak.

The Karnak temple complex is orientated almost parallel to the river Nile. Therefore, the obelisks could be moved on rollers, or skidded along a prepared track straight from the river along an unobstructed route perpendicular to the centre line of the temple complex, to their final positions. See fig. 4. From the contour lines at Karnak, obtained from Google Earth, and the rise of the water level during a period of high water, it is clear that during HW the site of the temple complex flooded every year(8). This allowed landing of heavy weights at the site during construction.

Fig.4. Track at Karnak (Image©2016 DigitalGlobe)

During the period of low river level (LW), a bed would have been prepared near the construction site at Karnak to allow mooring of the barge when the river Nile is at its highest level (HW). When the water level dropped after a few weeks, the barge settled on the prepared bed, on which it could stay for a prolonged period of time when the area dried out completely. The bed was constructed such that it supported the structure of the hull of the barge over her full length and possibly part of the bow. The structural arrangement inside the barge allowed a good load distribution, when there was no water above the bed. A 7-9 m high dam with a length of about 70 m was required between the barge and the destination of the obelisks between the fourth and the fifth pylon at Karnak. See fig. 5.

Fig. 5. Discharging at Karnak (author)

The first obelisk was moved with the top in forward position and was rolled or skidded, and parked past its final position prior to upending. The second one was rolled or skidded with its bottom end forward and was probably erected first. For that reason, the obelisks were placed on board on the centreline of the barge, one behind the other in opposite directions. The foc’sle had to allow passage over the bow. The tie-ropes of the hull were running over portals of varying heights, forward ones being the highest to give extra support to the bow. The ropes split sideways to port and starboard allowing passage of the obelisks. The bowsprit was structurally extended down to the base line of the hull, such that the contact with the prepared bed was extended forward as far as possible.

The elevated land route had to be prepared for the haulage of both obelisks. Two rows of blocks were providing the foundations for the route. The tops of the blocks had to be in one level plane. Wooden rails were laid on top of the blocks for easier rolling or skidding. They provided the surface on which wooden rollers or skids were running underneath a sledge fixed to the obelisk. As the obelisk was moving forward, rollers and rails appeared from behind and had to be carried forward. This means that, if well planned, the skid tracks on the barge and the skid tracks on land were in one horizontal level plane and only friction needed to be overcome. After the cargo had been transferred ashore, the barge could be disassembled in the dry and the materials could be shipped back up-river using normal cargo vessels. The light weight of the barge has been estimated as 1250 t. The cargo vessels could sail up-stream using sails and oars during the LW period when the currents in the river were small and most of the time northerly winds were blowing. Thus, the HL-barge could be used again for two more obelisks if required.

With the delivery of the two obelisks on the shore, the marine operation was completed.

p

Upending.

Although the upending of the obelisks is no part of the marine operation, some thoughts have been dedicated to the procedure for upending, because if the obelisks were delivered in a condition from which they could not be upended, the marine operation would be useless. The rise of the water was used also to bring the obelisks in an elevated position, necessary for upending. So, no uphill skidding or rolling was required.  The description refers to the operation for the second obelisk. See fig. 6.

Fig. 6. Upending of the second obelisk (author).

It should be noted that this procedure is an outline only. The second obelisk was moved forward until the centre of gravity was past a pivot point. The pivot point was located on a temporary retaining wall, set at a precise distance to the base. Over the distance A-B, the sledge was firmly pulled against the obelisk by tensioning ropes such that the obelisk does not slide when the sliding force is 320 t. By digging away the embankment in front of the retaining wall, the obelisk rotated in a controlled way into position 2 by gravity until it hit a stopper down below. From position 2 to position 3, the obelisk had to be pulled by significant large number of hands, with holdback from another large number of hands.

p

Navigating on the river Nile.

The relief in fig. 7 shows 27 tugboats and 3 pilot boats tied to the bow of the HL-barge at different locations operating in three rows. Sᴓlver(9) has estimated that the joint pulling power of 30 towing rowboats would be equivalent to the bollard pull of a 50-horsepower tugboat. It has been assumed that this is at zero forward speed. Such a tug can provide about 500 kg pulling force. A major part of this force may have been used for steering the bow in the right direction avoiding obstructions en route, as well as for keeping ahead of the barge when some speed is gained from the “downhill effect” of the water surface of the river (see next paragraphs for further explanation). Keeping out of each other’s way may also have taken some effort. Consequently, there will not have been much left to give the tow some forward speed. This kind of navigation on the river can be described better as controlled drifting.  This also means that there was no room for mistakes by the master of the tow or by the pilots. It would have been nearly impossible to make corrections in time.

On the bow of the leading rowboats, pilots were present to sound the water depth with long poles and to give directions accordingly. There were three escort rowboats (below on the left), which were not connected to the barge, for performing religious ceremonies. A smaller rowboat is also shown on the relief (standby function?). It is expected that the master was on the foc’sle of the HL-barge, which would give him a good overview. Three high level dignitaries were also on the foc’sle (10).

Fig.7. Relief 2 at Deir El-Bahri
(GNU Free Documentation License 1.2)

The shipment took place during a period of high water on the Nile. The current velocity was about 3.3 knots(11). Based on information from Google Earth and the rise of the river at high water, the difference in water levels between Aswan at the quarry and at Karnak (assuming 2.50 m water over the temple floor ref. Monderson(12)) could be as much as 21 m. The sailing distance is 213 km. This means that the river runs downhill with a gradient of 0.010% on the average. This results in a downhill component of the weight of about 200 kg, in assistance to the oarsmen on the row boats(13). It is estimated that this added 1.1 knots to the speed of the barge in relation to the water, and enough to give some pressure on the rudders. The presence of four big oars for steering suggests this effect to have been known at the time. The row boats must be able to move faster than 1.1 knots, otherwise they will be overrun by the barge.

The speed of the convoy in relation to the land was probably 4.4 knots or 8.1 km/hr. Sailing at night may have been considered too risky, and people needed rest and food. When sailing is restricted to 8 hours per day, excluding (un)mooring, the route was covered in 3½ days. Leaving and entering ports especially may have been nerve-racking for the pilots, as well as manoeuvring through the rapids near Aswan, not to mention negotiating bends in the river. Maybe marker buoys were installed at critical locations. It is likely that special procedures were developed for such operations, written down in a book with instructions (operations manual). Apparently, the Egyptian navigators managed to do this without damage to the cargo or the barge, an impressive achievement in its own right!

 

Loading at Aswan

The quarry was located at Aswan. An unfinished obelisk can still be found in that location confirming the orientation of the obelisks (see fig. 8). The obelisks were orientated at an angle of about 60 degrees to the river. In line with the orientation, a prepared bed was constructed a bit further into the valley, which only flooded when the waterlevel in the river was high. It took about 7 months to cut an obelisk out of the bedrock. During the same period, the HL-barge was constructed on the prepared bed when dry, assembled from prefabricated timbers.

Fig. 8. Quarry at Aswan, unfinished obelisk (Image©2016 DigitalGlobe)

Upon load-out, the track from the quarry was extended via a dam over the barge deck, after which the first obelisk was moved on board before the start of the period of high water, immediately followed by the second one. See fig. 9. The particulars of the land route were similar to the land route at Karnak, except it was running downhill at a small gradient (probably 2.4%) over a distance of about 400 m in a straight line. The arrow on fig. 10 shows the probable location of the prepared bed. Once on board, sea fastenings were installed. Water tightness of the hull could to be confirmed during the rising of the waterlevel in the Nile. The raised foc’sle had to be temporarily removed to allow access along the centreline of the barge.

Fig. 9. Cross section along quarry towards river Nile, along route A-A (author)

Fig. 10. Route down to the river, via flooded valley at HW (Image©2016 DigitalGlobe). Dotted line shows overland route of obelisk via a dam to the possible loading location of the barge.

Fig. 10 shows the sailing route from the quarry to the river through the flooded valley. The rise of the river at Aswan was about 8 m(14) under average conditions. The fleet of rowboats and assisting vessels were mobilised in time for the float-out. Prior to joining the river, the barge was turned around bow first in the flooded valley in quiet waters. Joining the river, with a current velocity of 3.3 knots, required a lot of effort from the oarsmen, and it demanded experienced navigators and pilots.

 

Hydrostatic properties

A lines plan of the barge has been made (see fig. 11), which is the basis of the hydrostatic calculations in this document. The lines of the frames only are shown below, but a full lines plan was made for the calculations. There are two planes of symmetry, centre line and mid length, hence only one quarter of the hull is shown.

Fig.11. Frames (author)

The barge required to be shallow draft, with a curved hull and with extra deck strength along its centre line. The barge had to carry the cargo and its own light weight, and it had to have sufficient stability during the river transfer. The stability during loading and discharging was no problem as the hull was be supported on a prepared bed. To judge the free-floating stability, present day criteria for heavy lift transportation have been applied for reference.

The following is a summary of the hydrostatic calculations of the hull.

Barge particulars for this study

Length (L) over hull:  81.80 m Length(wl) : 65.30 m
Beam (B) over hull:  27.30 m Beam(wl) : 25.50 m
Depth to maindeck:   5.00 m Light draft:    1.75 m
Loaded draft:   2.50 m Block coeff. δ: 0.480
Displacement:   2000 t Main frame coeff. β: 0.735
Light weight, est:   1250 t

Stability, expressed in metacentric height (GM) in meters.

Metacentre above keel                KM = BM + BK = 23.01 + 1.50 = 24.51 m

Centre of gravity above keel       KG = 4.81 m

GM = 24.51 – 4.81 = 19.70 m(15) (> 1.00 m)

The stability appears to be more than sufficient, even for present day criteria.

 

Hull planking

For the hull planking sycamore(16) wood was used, found locally in Egypt. The reason was probably cost, as wood imported from the Lebanon for shipbuilding would have made the barge very expen-sive for only one or two shipments. The type of wood used required a special hull design and building technique as it was only available in relatively short lengths. The wooden sycamore blocks were shaped to form a segmental arch similar to a segmental stone arch in old churches, but upside down. The seams were compressed by the water pressure in transverse direction and by the hogging (middle-up) condition of the loaded barge in the other direction. Papyrus sealed the seams from the inside. The blocks were “pinned” together (mortise-and-tenon joints). Traditional Egyptian boats do normally not require extensive framing for the strength of the hull considerations. Framing was required for load distribution under vertical columns shown below (see fig. 12 and fig. 14). The shape of the arch was probably elliptical, without any flat surfaces. For the drawing of the construction of the hull details of the Khufu ship(17) have been studied. The special arrangement of the hull planking caused a certain flexibility in the overall structure of the barge. This would have resulted in a good distribution of the internal loads introduced by the cargo, thus avoiding load concentrations anywhere in the hull. The water tightness was ensured by caulking with papyrus. Only the obelisks were extremely stiff against bending, they kept the barge straight under load.

 

Framing arrangement

The relief shows the three deck beams above each other protruding the hull at specific locations (see fig. 12 below). Portals for tie-ropes have been shown every two frames at equal distances. It has been assumed that underdeck the same supports are extended down to the bottom of the hull. Three deck beams are connected with blocks such that each individual beam bends with the same deflection under load. Wooden spacer blocks assure that this is the case. This means that the stiffness of three beams can be added up for combined strength. The beams were made from cedar wood from the Lebanon, as well as the longitudinal deck planking where longitudinal strength was very important. Very long beams would normally be available from the Lebanon, although figure 12 shows three sections instead. The transverse deck beams also prevented the hull from deflecting outward and keeping the hull in shape, thus adding to the hull overall strength as well as carrying the deck load.

Fig. 12. Main frame (author)
Left: grounded during load-out, right: afloat

For argument sake, the structure of the main frame of figure 12 has been analysed for strength, the results of which have been shown in figure 13. The light weight of the frames has been added to the weight of the obelisks, and the total has been factored with 10%. The beams are comprised of cedar wood with an allowable stress(18) in bending 34 N/mm2, allowable in compression 27 N/m and an allowable shear of 5.6 N/mm2. If the three beams are deflecting together in the same way, their joint section modulus can be expressed as follows:

S = 3 x 500²x 300 / 6 = 37.5 x 106 mm3

Bending stress                  σ = 89 x 107 / (37.5 x 106) = 23.7 N/mm2  (max. 34 N/mm2)

Shear stress                       τ = 1.5 x 20.6 x 104 /(3 x 500 x 300) = 0.7 N/mm2 (max. 5.6 N/mm2)

When the barge is on the prepared bed, the full weight goes down one column and the compressive stress can be calculated as follows:

Compressive stress         σ = 20.6 x 104 / (500 x 300) = 1.4 N/mm2  (max. 27 N/mm2)

It can also be noted that intermediate frames are required as the main frames alone cannot take the load. The intermediate frames are not visible on the relief in the temple of Deir el-Bahri.

Fig. 13. Strength analysis of assumed main frame structure (author)

The obelisks were supported on at least 15 main frames and 14 intermediate frames in direct load transfer. The barge was at its strongest over the centre line. Fig. 14 shows the layout of the intermediate frames.

Fig. 14. Intermediate frame (author)
Left: grounded during load-out, right: afloat

Longitudinal strength.

There are no waves on the Nile during navigation, so any wave bending moments could be dis-regarded for the design by the Egyptian naval architects. When only one loading condition needs to be considered, it is theoretically possible to design the lines of the hull such that the distribution of the displacement along the hull is exactly the same as the distribution of the weights in free floating condition, in which case there would be no bending moments in the hull structure at all. In practice this is difficult to achieve and small bending moments will occur (see fig. 14). These bending moments must be taken up by the hull, for this vessel by tie-ropes in tension and bottom in compression. For tensioning up, the ropes must each be double and they must be comprised of combined right-hand lay and left-hand lay in each strand and all running in the same direction. The tie-ropes are also required during load-out when the barge is positioned on the prepared bed, when only the bow is (partly) unsupported. Alternatively, the bow could be further supported on an extension of the prepared bed.

The ends of the barge must be shipshape to ensure water tightness and low resistance. As the deck load is more or less equally distributed over the length, and most of the displacement is concentrated over about 70% of the waterline length, the barge will resume a hogging condition (middle up) and the obelisks will not be evenly supported. This can be corrected by tensioning the tie-ropes to resume a straight line for the deck. The weight of the cargo is transferred to the water via 29 vertical main frames, directly under the load at the locations shown on the relief by the beams protruding the hull in the side, and at intermediate frames. As long as the barge is supported on one of the prepared beds, the vertical load is transferred directly into the supports on the bed at the vertical frames and no longitudinal bending of the hull takes place (except some at the bow during loading and discharging).

The wall relief shows clearly that the obelisks have been loaded with their lower ends towards the middle of the barge (fig. 1). This is where the barge has most of its strength in bending and most of its displacement to take vertical loads. Fig. 14 shows the overall distribution of the loads in the hull resulting from the weights and the displacement (Δ and G), and the bending moments (curve M). The bending moments in the hull have been calculated along the length of the hull. This results in a maximum bending moment of 720 tm at mid length. If the relief is regarded closely (fig. 1), it can be seen that the portals supporting the tie-rope forward are higher than aft. This suggests a different load case, probably for the condition of loading and discharging over the bow. For that case, the bending moment is expected to be lower than in the free-floating condition. This is not further analysed. It confirms however that the obelisks are loaded out and discharged along the centreline over the bow of the barge.

For the analysis of the requirements for the tie-ropes, 6 pre-stressed ropes have been assumed for reasons of symmetry, each rope being comprised of 2 strands. Assuming the distance between ropes and bottom of the barge is 10.00 m, the force per strand is 720/(10 x 12) = 6 t. With a safety factor of 3 to break, 12 x 40 mm ropes would be required to take up the required tension (manila rope or similar). For the bottom planking a safety factor of 2.0 would be sufficient, resulting in a compressive load of 144 t, equally spread. From the relatively light ropes can be concluded that the Egyptian naval architects expected already that the distributions of the weight and the displacement along the length of the hull were close, resulting in relatively small bending moments.

Fig. 14. Global hull loads (author)(19)

Inconsistencies in the layout shown on the relief.

The following particulars of the wall relief cannot be explained from the procedures in this document. It is felt however, that these are of a minor nature.

  1. The leading edge of the sledge of the second obelisk should be at the bottom end and not as drawn at the top end of the obelisk as it had to be discharged over the bow, unless one is shown in loading condition and the other one in discharging condition.
  2. The barge deck cannot show any sheer or camber, it must be flat, no matter how loading and discharging were conducted.
  3. It would be more logical if there were six tie-ropes for reasons of symmetry.

 

Review of earlier publications.

  1. NRC Handelsblad 5.12.1989. Onderwijs. Article by A. Wegener Sleeswijk. Title: Zwaar Transport (Dutch). Mr. Wegener Sleeswijk’s assumption that the obelisks could not be loaded behind each other on the centreline of the barge leads to various assumptions concerning the layout and the structure of the barge, which conflict with the original relief. Estimated weight 374 t each.
  2. Björn Landström: Het Schip, pages 22, 23 (Dutch). Obelisks placed next to each other with both upper ends forward under portals with tie-ropes. This does not seem possible for reasons of loading and discharging operations, and it conflicts with the relief. Some details of construction of hull are given, which have been used for this study.
  3. A Companion to Ancient Egypt. Alan B. Lloyd. Pages 379 and 380. Refers to Stephan Seidlmayer personal communications 2009: confirms that two obelisks were transported on one barge. Refers to Björn Landström and presents 3-D drawing after Björn Landström. Quotes a displacement of loaded barge 7300 t, which is believed to be excessive.
  4. Pharaonic Egypt. reshafim.org.il, Ancient Egypt, river boats. Description of loading and discharging at Roman times. Raises the question how such a transport was done and quotes Harold L. Potts stating: Björn Landström: “My reconstruction (of the barge) hardly even convinces me” (2007).
  5. NOVA online, Mysteries of the Nile, Gifts of the River, 1999. Several solutions shortly discussed, most likely solution was to float the barge under the obelisk, which was laying across a purpose-built canal. This is not in accordance with the relief.
  6. Reginald Engelbach suggests in his Study of the unfinished Obelisk at Aswan (p. 64-65) that the barge was built into an embankment and was completely covered up. The obelisk could be rolled or skidded over the barge over the embankment. By removing the sand step by step from underneath the obelisk, it would be lowered down to the deck (“sand-jacking”). By removing sand from, and out of, the barge, the barge could be lifted off its berth using the barge’s displacement. Discharging could be done in the same way, except the embankment for discharging did not need to be so high. With two obelisks, the method is more laborious, but the principle would be the same. This operation is considered feasible, but complicated.
  7. Seán McGrail, Boats of the World. Paragraph 2.9.3 Heavy Lift Vessels and Towing. Reconstruction drawing of the barge (fig. 2.26 page 44 not in accordance with relief). Quotes other writers. Herodotus describes a way of navigating downstream on the Nile when a 250-kg weight is dropped on the river bed off the stern on a rope. As it slows down the barge, pressure is built up on the rudders making steering possible. However, this effect must be very limited, because both the weight and the rudders are on the stern. The weight prevents the rudders to steer the barge. For this method to work, the barge needs to rotate with the bow into the current and the weight needs to be dropped from the bow. The relief shows, that towage is done from the bow(20), and steering from the stern, which contradicts the description by Herodotus.
  8. The Medieval Nile: Route, Navigation and Landscape in Islamic Egypt. John Cooper. Chapter 8, pages 125-127. Maximum velocity during High Water about 3.9 kn downstream Aswan, about 1.5 kn during Low Water.
  9. H. Breasted, Ancient Records of Egypt, Volume 2, 18th Dynasty, paragraph 328, page 137: barge main particulars estimated by comparison with other more documented vessels. L = 268½ ft and B = 89½ ft.
  10. Flickr, 2008 Cameron Grant. The top of Hatshepsut’s fallen obelisk – Karnak. The obelisks were “moved down the river’s edge during winter at low water and loaded onto a huge barge and they sat until the following summer when the flood came and floated the barge”. No references given. Not until the river’s edge for loading but to inundated valley, but otherwise in agreement with assumptions in this document.

 

Attachment

An artist impression of Hatshepsut’s Heavy-Lift river barge following of the particulars discussed in this document is shown also in fig. 15.

p

References

  1. NRC Handelsblad. Onderwijs. Article by A. Wegener Sleeswijk. Title: Zwaar Transport (Dutch). 12.1989.
  2. Björn Landström: Het Schip (Dutch). 1961.
  3. A Companion to Ancient Egypt. Alan B. Lloyd. 2008
  4. Stephan Seidlmayer. Die Vermessung des Nils im Alten Ägypten. 2011
  5. Encyclopedia debinnenvaart.nl : stevelen. (Dutch)
  6. Pharaonic Egypt. reshafim.org.il, Ancient Egypt, river boats.
  7. Monderson. Temple of Karnak: The Majestic Architecture of Ancient Kemet. 2007
  8. Sad. The River Nile, Geology, Hydrology and Utilisation.
  9. NOVA online, Mysteries of the Nile, Gifts of the River.
  10. Study of the unfinished Obelisk at Aswan, Reginald Engelbach. 1923.
  11. Seán McGrail, Boats of the World. 2001.
  12. The Medieval Nile: Route, Navigation and Landscape in Islamic Egypt. John Cooper. 2014.
  13. H. Breasted, Ancient Records of Egypt. 1906.
  14. Flickr, 2008 Cameron Grant. The top of Hatshepsut’s fallen obelisk – Karnak. 2008.
  15. Google Earth, Coordinates of reference locations and location maps.
  16. The royal ship of Khufu, nautarch.tamu.edu
  17. GL Noble Denton, Guidelines for Marine Transportations, rev. 5
  18. realcedar.com. Groei, eigenschappen en toepassingen van WRC.pdf

 

Footnotes

[1] Stephan Seidlmayer. Personal communication 2009: ”an official of Hatshepsut called Hapusenet states that he was responsible for the transportation of two obelisks on one barge” from: A Companion to Ancient Egypt.

[2] NRC Handelsblad 5.12.1989. Onderwijs. Article by A. Wegener Sleeswijk.

[3] Stephan Seidlmayer. Die Vermessung des Nils im Alten Ägypten. “28 ellen in Elephantine, 14 ellen in Kaïro”.

(28 ellen = approx. 14.60 m, 14 ellen = approx. 7.30 m). Elephantine is an island at Aswan. Ref. comment by Prof. Dr Seidlmayer: “it is true that the rise of the river Nile could reach (in certain periods) 28 cubits on the nilometer scale; however, for not entirely clear reasons, the 0-point of this nilometer scale happened to be far below low water. Actually, the difference in height between high and low water at Aswan used to be under natural conditions around 8 m only”.

[4] F. Said. The River Nile, Geology, Hydrology and Utilisation. P.152, rise above temple floor 0.84 m.

[5] F. Monderson. Temple of Karnak: The Majestic Architecture of Ancient Kemet

[6] Ref. comment by Prof. Dr Seidlmayer: I would like to add the following information: on your Fig 2 also the obelisk of Thutmosis I is visible. Actually, there exists an ancient text of the person who was responsible for building the boat which was used to ship this obelisk and its sibling. Comparing the length of this boat, which is stated in the text, and the length of the existing obelisk it seems clear that the obelisks of Thutmosis Ist were transported in exactly the same way – both on a single boat and arranged in one line along its centre. This seems to have been standard practice. The unusual accomplishment in the case of Hatshepsut’s obelisks seems to have been the unparalleled size and weight of her obelisks rather than the fact that both were loaded on a single vessel.

[7] Study of the unfinished Obelisk at Aswan, Reginald Engelbach. 1923

[8] F. Monderson. Temple of Karnak: The Majestic Architecture of Ancient Kemet. Quote from 1925 max. flood level: “… a zone extending from the floor level to a height varying from about half a meter to several meters….”. p. 160

[9] NRC Handelsblad 5.12.1989. Onderwijs. Article by A. Wegener Sleeswijk

[10] J. H. Breasted, Ancient Records of Egypt, Volume 2, par. 328, 329, pages 137, 138.

[11] The Medieval Nile: Route, Navigation and Landscape in Islamic Egypt. John Cooper. Chapter 8, page 125-127. A maximum of 3.9 knots has been quoted in the 19th century.

[12] F. Monderson. Temple of Karnak: The Majestic Architecture of Ancient Kemet. P.160

[13] Up to about the year 1900 cargo vessels navigated down the river Rhine without using engines or sails making sometimes a speed of 4 to 5 km/hr in relation to the water surface using the same principle. Ref. www.debinnenvaart.nl : stevelen (Dutch). For steering they used oars and rudders. Rafts with sometimes 51.000 m3 wood were floated down the Rhine in the 17th and 18th centuries using the same principle. It is known that during Roman times this was done as well. The down-hill gradient of the Nile is less than of the Rhine, but it is considered enough to have a similar effect during periods of high water.

[14] Ref. comment by Prof. Dr Seidlmayer: “it is true that the rise of the river Nile could reach (in certain periods) 28 cubits on the nilometer scale; however, for not entirely clear reasons, the 0-point of this nilometer scale happened to be far below low water. Actually, the difference in height between high and low water at Aswan used to be under natural conditions around 8 m only”.

[15] GL Noble Denton, Guidelines for Marine Transportations, Section 10, Clause 10.2.4. Normally GMo>1.00 m

[16] J. H. Breasted, Ancient Records of Egypt, par 329, p.137

[17] The royal ship of Khufu, www.nautarch.tamu.edu

[18] www.realcedar.com. Groei, eigenschappen en toepassingen van WRC.pdf

[19] Δ is displacement and G is weight. The difference Δ – G = Q (ton) depicts the longitudinally distributed load on the hull. From this curve, the distribution of the bending moments can be calculated by two times integration, resulting in the maximum bending moment of 720 tm at ½L in this case

[20] In The Netherlands, sailing river barges without engines used to navigate backwards through narrow bridges and under low railway bridges by deploying an anchor chain from the bow up-steam and steering with the rudder after they lowered their sails and the mast. This was still done during the first half of the 20th century.

T.  Hoogeveen, naval architect and marine heavy lift specialist.

Drawings prepared by author, except where noted.

e-mail: marine-consultant@kpnmail.nl

July 1, 2018

Fig. 15. 3-D impression of queen Hatshepsut’s Heavy-Lift river barge (author)