The Arles-Rhône 3 project : From the excavation and

raising of a Gallo-Roman barge, in the southern coastal

French city of Arles, to its documentation and modelling
in 3D (2011-2012)
Sabrina Marlier, Pierre Poveda, Nicolas Ranchin

To cite this version:

Sabrina Marlier, Pierre Poveda, Nicolas Ranchin. The Arles-Rhône 3 project : From the excavation
and raising of a Gallo-Roman barge, in the southern coastal French city of Arles, to its documentation
and modelling in 3D (2011-2012). 13th International Symposium on Boat and Ship Archaeology,
Jerzy Gawronski; André van Holk; Beno van Tillburg, Oct 2012, Amsterdam, Netherlands. pp.383-
389. �halshs-01807685�

HAL Id: halshs-01807685

https://shs.hal.science/halshs-01807685
Submitted on 15 Jan 2019

HAL is a multi-disciplinary open access L’archive ouverte pluridisciplinaire HAL, est

archive for the deposit and dissemination of sci- destinée au dépôt et à la diffusion de documents
entific research documents, whether they are pub- scientifiques de niveau recherche, publiés ou non,
lished or not. The documents may come from émanant des établissements d’enseignement et de
teaching and research institutions in France or recherche français ou étrangers, des laboratoires
abroad, or from public or private research centers. publics ou privés.
 60. The Arles-Rhône 3 project (Arles, France).
From the excavation and raising of a Gallo-Roman
barge to its documentation and 3D-modelling
(2011-2012)

Sabrina Marlier, Pierre Poveda & Nicolas Ranchin

Introduction one end to the other - even in its upper part, up to

the gunwale. Moreover, all the internal wooden
The Arles-Rhône 3 (AR3) shipwreck was discovered elements related to the galley gear and tools, and to
in 2004 on the right bank of the Rhône river, in Arles, the cargo have been preserved.
during the creation of an archaeological coast survey 3 One steering oar, associated with the barge because
map of the area, directed by Luc Long, of the French of its appropriate ratio and its similar dating, was
Department of Underwater Archaeological Research also discovered close to the shipwreck in 2004 (Long
(Drassm) (Long in Long & Picard, 2009: 233-234). in Long & Picard, 2009: 240 and 242).
Situated between 4 and 9 m of depth and on a 35° slop- 4 Finally, the boat presented one inscription
ing bank, the wreck was embedded in harbour garbage (C.L.POSTV) stamped in the wood of the starboard
dump of the city of Arles from the Roman period, which side, in the aft section (Djaoui et al., 2011: 156).
represented a stratigraphical context of extremely
important archaeological material (Djaoui et al., 2011).
Between 2005 and 2010, a number of assessments, a Excavation and raising of the AR3 shipwreck
survey and finally an excavation of the shipwreck were
carried out. These operations were conducted by the Because of the special characteristics of the AR3 ship-
Arkaeos Association, with the collaboration of the Arles wreck, at the end of 2010 the General Council of the
Museum of Antiquity, the Centre Camille Jullian/CNRS Bouches-du-Rhône, to which the Arles Museum of
and the Drassm. The results showed the importance and Antiquity belongs, decided with the Drassm to com-
value of this shipwreck, which was identified as a Gallo- plete the excavation and to raise this Gallo-Roman barge
Roman barge dated from the middle of the 1st century in order to restore it and to present it within the Arles
BC (Long et al., 2009; Marlier, 2011; Marlier et al., 2012). Museum. In order to accommodate the reconstructed
shipwreck, the museum was expanded in 2012. The 800-
m² extension, which was inaugurated in October 2013,
Characteristics and state of conservation of the AR3 hosts more than 400 objects in connection with the com-
shipwreck mercial activities, shipping and port activities of the
city of Arles in Roman times. Furthermore, the French
1 The shipwreck corresponds to a barge that sank Ministry of Culture has classified this boat as a Trésor
when it was still in use: this is not an abandoned National (‘National Treasure’). The entire operation
boat. Thus, the wreck has conserved its galley gear was linked to the cultural manifestations of Marseille-
with the glazed ceramic, and one dolium, reused as Provence in 2013, the year where the city and its region
a brasero for cooking as well as other tools. Also the were named European capital of Culture. Due to this
barge’s cargo was preserved in place, consisting of 21 event, the deadlines for the end of the excavation and
to 31 tonnes of limestone blocks from neighbouring the raising of the barge, as well as the restoration and
quarries located 15 km north of Arles (Ernaginum/ reassembly, have been very short.
Tarascon), thus indicating the direction of the boat The concluding excavation and raising of the barge
during its last trip. took place in 2011 over a period of seven months (fig. 1).
2 Concerning its state of conservation, except for one This operation mobilized important means (€ 1.9M)1 and
part of the aft port side that is missing, the wood is was created within the framework of a government con-
very well preserved and the barge is complete from tract; a tender won by the French operators Ipso Facto,
384 Sabrina Marlier, Pierre Poveda & Nicolas Ranchin

Fig.1. Raising of one of the ten sections of the Arles-Rhône 3 barge with, in the background, on the left bank of the Rhone, the blue archaeo-
logical museum (Photo: © Teddy Seguin, O’Can-Ipso Facto, Mdaa/CG13).

Fig. 2. 3D recording using a C-Track (Photo: © Teddy Seguin, O’Can-Ipso Facto, Mdaa/CG13).
 60. The Arles-Rhône 3 project (Arles, France) 385

specialized in underwater archaeology, and O’ Can, spe- 3 The reflective targets: placed all around the subject
cialized in underwater public works.2 The operation to be recorded to create a reference system within
was directed by Sabrina Marlier (Mdaa), David Djaoui a digital space. It allows the C-Track to follow the
(Mdaa), Mourad El Amouri (Ipso Facto), Sandra Greck HandyPROBE within the reference system, and
(Ipso Facto) and Benoî� t Poinard (O’Can) and executed by insures precise measurements. Once the reference
a mixed team, composed of archaeologists and under- system in place, both the probe and the C-Track may
water archaeologists (Arles Museum of Antiquity, move freely within the digital space without the
DRASSM and Ipso Facto), commercial divers (O’Can) C-Track ever losing its orientation.
and conservator-restorers (Ipso Facto, Arles Museum
of Antiquity, Arc-Nucléart and A-Corros). This under- The digital recording methodology was broken down
taking has allowed to finally complete this exhaustive into two separate phases: one ‘macro’ and one ‘micro’.
excavation, which began in 2008, and led to the rais- The ‘macro’ phase consisted of recording whole seg-
ing of the barge in 2011, measuring 31 m in length, in 10 ments, with all the architectural elements composing
sections according to a method specially designed for them still assembled to one another. This unique oppor-
the occasion by the foreman, Benoî� t Poinard, in part- tunity allowed to produce accurate and precise docu-
nership with the engineering office of Ipso Facto. First mentation. This type of recording was nevertheless
documented underwater during the excavation, the difficult to accomplish due to the chain of operational
raised sections of the boat were then meticulously and processes set in place. Between the moment a segment
exhaustively documented on the ground before their was raised from the Rhone river and the moment it was
transport to the city of Grenoble, where the conserva- conditioned and packed by the conservation team to be
tion laboratory Arc-Nucléart is located. This documen- shipped to the restoration laboratory at Arc-Nucléart,
tation included: archaeologists had only one week to accomplish every
−− Precise observations; single documentation task. With the digital recording
−− Sampling fabrics and pitch for further analysis; representing just one aspect of documentation, the
−− Identification of the wood species used for the con- ‘macro’ recording phase could not exceed 48 hours in
struction of the boat; order not to block the other documentation tasks.
−− Extensive photographic recording to support the The second recording phase was the ‘micro’ docu-
study of the architectural characteristics, as well as mentation. This consisted of recording mobile elements
the dendrochronological analyses; in detail, such as the planking at the bottom of the bulk-
−− 3D drawing of all the sections and other wooden head, or elements removed after the ‘macro’ phase, such
elements of the boat was added to these classical as the side planks. It is important to note that the con-
ways of documentation. tour of the elements that were to be documented in the
‘micro’ phase were quickly drawn in the ‘macro’ phase
to facilitate their placement in 2D and 3D plans during
Documentation and modelling in 3D of the AR3 post-treatment.
shipwreck The documentation procedure itself consisted of
recording the visible facets of each architectural elem-
The device chosen for this project, from the Canadian ent by connecting points captured in 3D to form polyline
company Créaform 3D, had never been used for archaeo- assemblages. The next step was to export the unedited
logical purposes before. It consisted of a 3-dimensional files with high-precision raw results to Adobe Illustrator
recording instrument compatible with the Rhinoceros to generate technical plans that others on the project
3D software, like the now standard FaroArm, with the could immediately use for their own documentation
particularity of being wireless (fig. 2). It is composed of tasks. Each plan depicted either a whole segment or
three complimentary elements: independent architectural elements seen from differ-
1 A HandyPROBE: a wireless probe with nine reflec- ent perspectives onto which individuals could add their
tive targets on it used in the recording phase. The own notes and observations (for example, the location
tip of the probe is positioned on the point to be of iron nails taken out of the half trunk flanks during
recorded, and the digital data recorded are instantly disassembly). These sheets allowed us to create a certain
transferred to the software and computer once the homogeneity and consistency in the sharing of informa-
trigger on the probe is activated. tion between the various teams in the field.
2 C-Track: two infrared cameras set in a case mounted Following a four months period of intensive use, a
on a photography ball-head and tripod. Their func- primary report on the utilization of the C-Track for
tion is to track ‘live’ the movement of the probe and nautical archaeology purposes could be drafted. The
to record the digital data captured with the probe. major advantage of this instrument is that it is wireless.
The moment the trigger is activated, an instant tri- This granted a great freedom of movement, without ever
angulation is made between the two cameras and having to disassemble each architectural element. This
each of the nine targets on the probe. also helped to drastically shorten the recording time
without ever neglecting data quality. That being the
386 Sabrina Marlier, Pierre Poveda & Nicolas Ranchin

case, the C-Track does also present noticeable disadvan- represented a basic archive that required a large amount
tages. To mention one; it is extremely sensitive to the of cleaning up and reworking of the recordings in order
work environment. Originally designed to be used in a to get the documentation needed for the study and
controlled environment protected from the elements, it publication.
was instead used in extreme conditions throughout the
project that caused the instrument to malfunction, slow-
down, and in some instances to shut-down altogether. Restoration of shapes and missing parts
Furthermore, the C-Track is a contact-based recording
tool, which is not preferable when wanting to preserve Next, considering the very good preservation state of
the physical integrity of a subject. the barge, it seemed necessary to attempt to restore the
few missing parts, as well as to try to straighten out the
various deformations of the wreck. The purpose of this
Post-processing phase task was to establish as precisely as we could the weight
and tonnage of the barge, as well as its draught and
The first step consisted of assembling under one master stability.
file every single separate segment, as well as all the loose Two more problematic issues particularly caught our
elements, such as the bulkhead, that had been recorded attention during this restoration task: on the one hand
at various times throughout the project. This task, we needed to recover the aperture angle at which the
accomplished by using the Rhinoceros 3D software, uti- side planks opened up after being seriously deformed
lized both fixed reference points recorded on each seg- over time in the water3, and on the other hand we
ment during its respective ‘macro’ documentation, and wanted to figure out how to fill in the missing elements
the description of certain architectural characteristics. from the barge’s stern. Relating to the side planks’ aper-
Thus, in order to reconnect two adjacent segments, we ture angle, the only unquestionable point of reference
based ourselves on the outward faces of the various was located at the mast-thwart, the length of which
architectural components created along the cutting allowed us to define the aperture angle at which the
line. Their perfect realignment allowed to capture the side planks opened up at that exact point. From there,
original position of each segment in relation to the adja- and based on a symmetrical opening on both sides, it
cent segment. In terms of integrating micro records into appeared that the initial opening angle of the flanks was
larger macro files, such as with the detailed recording of 92°. Regarding the restoration of missing parts at the
a side plank which needed to be reintegrated back into stern, the first step consisted of symmetrically restoring
its corresponding segment, we based ourselves on the the port side assemblage based on the preserved star-
repositioning of noticeable topographical points set up board side. At the stern, considering the replacement
beforehand on both the individual elements and on the of the steering oar and the positioning of the helms-
larger segments to which they belonged. Upon complet- man that we could deduce, we favoured the hypothesis
ing this task, we had a primary archive of the digitally of a one-piece rear transom onto which the steering
recorded barge in which each architectural element oar would rest, similar to that which has been done on
could be isolated and highlighted. This nevertheless

Fig. 3. The 3D restoration (Model: © Pierre Poveda, Ipso Facto).

 60. The Arles-Rhône 3 project (Arles, France) 387

Bm1

AR
Bm1
B3
B5 B3
B4
Bm2
B7

Bb Bm3
B7

B8
B6

Tb

Bm3

B8
B5
B4
AV
Bm4
Fig. 4. Perspective view of the assemblies of the monoxylous hard-chine bilges (BM1-BM3 and BM2-BM4) with the secondary planks (B7 and
B6) located at the bow, near the mast-step frame (Drawing: © Pierre Poveda, Ipso Facto).

the Altaripa Gallo-Roman barge replica discovered in The main features of the ship structure
Bevaix, Switzerland (Arnold, 1999: 151).
In relation with the modelling and restitution works
done on the Arles-Rhône 3 barge, the architectural study
Overall weight and tonnage analysis reveal a ‘bottom-based’ construction for which the main
features of the structure are the following.
The overall restoration, resulting in a complete 3D model
achieved with Rhinoceros, has enabled us to draft the The bottom
barge’s restored line plan from which we can run a cer- The bottom of the Arles-Rhône 3 barge is composed of
tain number of tests that will enable us to refine our three main strakes and two secondary planks located
perception of the ship’s utilization. First, it is by identi- at the bow, near the mast-step frame. All the planks are
fying the type of wood used in the barge’s construction, made of oak4 and are placed edge to edge, without any
as well as the volume of its individual elements, that we form of assembly between them, except at the extrem-
were able to determine precisely the overall weight of ities of the plank joints where transversal nails were
the empty barge, which we estimate at 8.13 t. Similarly, used for their connection to the adjacent planks.
based on the bulkhead’s restoration, we were able to
determine the exact volume available for cargo. Knowing The monoxylous hard-chine bilges and the composite bilge
the density of the rocks as well as their average stowage strakes
factor, we estimate the cargo weight at 21,48 t, which To ensure the transition between the bottom and the
would represent a significant, yet possible, draught for sides of the boat, the design of the bilges is very complex.
the barge. This first analysis phase should continue in A combination has been applied of two principal mon-
order to refine our understanding of the barge and its oxylous hard-chine bilges, in the central part of the boat
usage by determining, for example, the hull resistance and in the fore part, and of composite bilge strakes in
and the barge’s manoeuvrability. the aft and also near the mast-step frame and at the end
388 Sabrina Marlier, Pierre Poveda & Nicolas Ranchin

of the bow. The monoxylous hard-chine bilge strakes are Nail fastenings
made of oak and form an ‘L’ shape, slightly rounded and All the frames, flat floor timbers and knees are fitted
in an almost right angle with the bottom. In addition to with iron nails to the bottom, to the bilge strakes and to
ensuring the transition between bottom and sides, these the sides. For the assembly to the planks on the bottom,
pieces give a structural longitudinal rigidity to the boat. the nails are inserted from the back of the frames and
The composite bilge strakes are made of a vertical plank from the outside of the bottom in order to form a strong
assembled to the adjacent plank on the bottom by means cross assembly. All nails have a folded tip, while the tip of
of a transversal nail inserted from the outside. the nails inserted from the back of the frames is double
folded on the outside of the bottom, creating a hook.
The frames The nails ensure not only the assembly of the frames
The framing-system consists of 47 flat floor-timbers and but also the cohesion of the complete structure of the
some occasional knees (20), all also made of oak. On the boat as the planks are not connected. There are nearly
fore and on the aft parts of the boat, the spacing between 1,700 nails used for the assembly of all the structures of
the frames is very wide (40.27 to 48.70 cm). By contrast, the barge with two different sizes of nails.5 Moreover,
in the central part of the boat, due to the heavy cargo, a great number of metallic elements were used to lock
the space between the frames is much tighter (31.96 cm some scarfs between two planks or to strengthen the
on av.). For the same reason, the dimensions of the flat bow, which is filiform.
floor timbers are in the central part of the boat (L.: 22-27
to 28-34 cm; H.: 8.2 cm on av.) than in the fore and the aft The watertightness
parts (L.: 14.5 to 31 cm; H.: 6.6 cm on av.). The mast-step is The boat is made watertight by means of wool fabrics with
a great flat floor-timber of 42 cm in width for a thickness pitch placed between the planks according to the luting
of 8.5 to 12.5 cm. Concerning the lengths of the frames as technique. The pollen analyses of the pitch revealed the
they correspond to the internal width of the boat, there presence of some olive tree taxa, which points towards a
is no real main frame that can be defined but rather a local construction of the boat. Numerous traces of pitch,
wider zone of the boat: this beam, situated in the central found both inside and outside the hull, also contribute to
part, measures about 2.30 m. The floor-timbers are flat the watertightness of the boat.
with rounded inner edges, never exceeding the top edge
of the chine-strake. The internal wooden structures
Linked to the galley gear and tools, and to the cargo, two
The monoxylous side planks types of internal wooden structures were placed in the
Over the total length of the boat, the sides are composed boat: on the aft part, in relation with the cooking and the
of monoxylous side planks made of a half-trunk of fir. working activities of the boatmen, a series of planks, of
Arranged perfectly symmetrically on each side of the resinous species, were glued with pitch directly against
boat, the principal monoxylous side planks measure the bottom or on the frame. In the central part, an open
more than 26 m in length. With a maximal height of 90 hold constituted of about 140 longitudinal and transver-
cm and a maximal thickness of 21 cm, they also provide sal-stacked movable timbers, and also made of resinous
longitudinal rigidity to the barge. They are fitted to the species, allowed to protect the structure of the hull and
bilge strakes by means of iron nails inserted obliquely to contain the heavy cargo.
from the lower external face of the monoxylous pieces
through the bilge strakes where the tips are driven back Shape and dimensions
at a right angle. The upper parts of the monoxylous side The complete barge measures 31 m in length, with a
planks are also fitted to the knees by iron nails. On the width of 2.90 m, and a height of about 1 m. It is a long
fore and on the aft of the boat, some gunwales, made out and narrow barge, with a very tapered bow until today
of oak, are nailed to the upper part of the monoxylous never observed on other Gallo-Roman barges discovered
side planks. in Europe. This boat is very well built, in a very complex
fashion, with particular attention paid to the perfect
The mast thwart symmetry of the numerous arrangements of various
Located above the mast-step frame, the mast thwart lies elements and structures.
on longitudinal fittings upon which knees are mounted
and fastened. The mast thwart has rather large dimen-
sions, similar to the mast-step frame. In the centre of the Conclusion
piece, a circular opening of 16 cm in diameter is pierced
for the passage of the mast. This mast was discovered After four years of excavations and one year (2011) for
under the stone cargo. It corresponds to the towing mast the remaining fieldwork and the raising and documen-
and measures 3.70 m in length. It is made of ash. tation of the shipwreck, the Arles-Rhône 3 project has
 60. The Arles-Rhône 3 project (Arles, France) 389

now been completed and the scientific publication is Batellerie gallo-romaine: pratiques régionales et influences mar-
finished (Marlier, 2014). Parallel with the research, the itimes méditerranéennes, Actes de la table-ronde internation-
barge was restored and reassembled and can be seen on ale, 27-28 oct. 2008 (Aix-en-Provence, MMSH). Bibliothèque
display in the Arles Museum of Antiquity. The success- d’Archéologie Méditerranéenne et Africaine 9. Centre
ful completion of this undertaking within a very short Camille Jullian, Editions Errance: 153-164.
time is firstly due to the high degree of competence Long, L. & Picard, P. (eds), 2009. César. Le Rhône pour mémoire.
of all of the excellent professionals who intervened at Vingt de fouilles dans le fleuve à Arles. Catalogue d’exposi-
all levels for this exceptional project, secondly to the tion, Musée départemental Arles antique, Arles. Actes Sud,
important means provided by the politicians and one Arles.
great patron, and finally to the generosity of the Rhône Long, L., Rival, M. & Marlier, S., 2009. The Gallo-Roman
river that allowed us to dive every scheduled day (except Wreck Arles-Rhône 3. In: R. Bockius (ed.), Between
for one week) in 2011, from May to November, in order the Seas. Transfer and Exchange in Nautical Technology.
to achieve our goal. Proceedings of the Eleventh International Symposium on Boat
and Ship Archaeology, Mainz 2006. Römisch-Germanisches
Zentralmuseum Tagungen 3, Mainz: 303-311.
Notes Marlier, S., 2009. Le Rhône, autoroute fluviale. In: Long &
Picard, 2009: 28-35.
1 The whole operation Arles-Rhône 3 (excavation, raising, Marlier, S., 2011. L’épave Arles-Rhône 3: étude préliminaire d’un
restoration, extension of the Museum and exhibit de- chaland gallo-romain. In: G. Boetto, P. Pomey & A. Tchernia
sign) costed € 10M: € 7M were funded by the Collectivity (eds), Batellerie gallo-romaine: pratiques régionales et influ-
(Conseil Général des Bouches-du-Rhône), € 2.5M by a pa- ences maritimes méditerranéennes, Actes de la table-ronde
tron (the Compagnie Nationale du Rhône), € 400.000 by the internationale, 27-28 oct. 2008 (Aix-en-Provence, MMSH).
French Ministry of Culture and € 80.000 by the Collectivity Bibliothèque d’Archéologie Méditerranéenne et Africaine
Provence Alpes Côte d’Azur Region. 9. Centre Camille Jullian, Éditions Errance, Paris: 131-151.
2 http://www.ips-o.fr/; http://www.ocan.fr/. Marlier, S. (ed.), 2014. Arles-Rhône 3. Un chaland gallo-romain
3 The disparities observed around the preserved angles on the du Ier siècle après Jésus-Christ. Archaeonautica 18. Éditions
transversal knees has quickly led us to decide not to include du CNRS, Paris, en co-édition avec le Musée départemental
them in the restitution due to important deformations. Arles antique, Arles.
4 All wood determinations were done by Sandra Greck (Ipso Marlier, S., Greck, S., Guibal, F. & Andrieu-Ponel, V., 2012.
Facto). Arles-Rhône 3: Architectural and Paleobotanical Study of a
5 The smaller size is of 16 cm on av. (L) with a diameter of the Gallo-Roman Barge from the 1st Century in the Rhône river.
head of 23-26 mm and a section of the stem of 8-9 mm; the In: N. Günsenin (ed.), Between Continents. Proceedings of the
largest size is of 20 cm on av. (L) with a diameter of the head Twelfth Symposium on Boat and Ship Archaeology, Istanbul
of 27-29 mm and a section of the stem of 8-9 mm. 2009. Ege Yayinlari, Istanbul: 203-210.
Marlier, S., Djaoui, D., El Amouri, M., Greck, S. & Poinard,
B., 2013. Arles-Rhône 3 – Ausgrabung und Bergung eines
References Römischen Lastkahns (Arles, Frankreich). In: M. Reinfeld
(ed.), Archäologie im Mittelmeer. Verlag Philipp von Zabern,
Arnold, B., 1999. Altaripa. Archéologie expérimentale et architec- Darmstadt/Mainz: 177-184.
ture navale gallo-romaine. Archéologie Neuchâteloise 25. Marlier, S. & Rieth, E., (in press). Le chaland Arles-Rhône 3:
Djaoui, D., Greck, S. & Marlier, S. (eds), 2011. Arles-Rhône 3. Le présentation architecturale et particularités au sein de la
naufrage d’un chaland antique dans le Rhône, enquête pluridis- construction navale gallo-romaine, actes du colloque. La
ciplinaire. Actes Sud, Arles. construction navale et ses objets: nouvelles approches, nouveaux
Greck, S. & Guibal, F., 2011. Étude dendrologique de l’épave outils (Nantes, 20-21 septembre 2012).
Arles-Rhône 3. In: G. Boetto, P. Pomey & A. Tchernia (eds),