1The site is located near the marina of the village of Fahrdorf opposite Schleswig on the southern banks of the Schlei at a depth of circa 2 m (fig. 1) and very close to the famous Viking-age settlement of Hedeby. The Schlei is a 42 km long inlet that stretches from the Baltic Sea to the town of Schleswig. In combination with the Rivers Treene and Eider and a land passage, the Schlei functioned as the main exchange route between the Baltic and the North Sea in the Viking Age and early medieval times (Brandt 2000, p. 319). Its importance is documented by over 40 wrecks from different periods, which have so far been found in its waters (Nakoinz 2014, p. 104).
Fig. 1: Map of the Schleswig isthmus region
1. Hamburg, 2. Lübeck, 3. Kiel, 4. Hollingstedt, 5. Schleswig, 6. Jürgensburg, 7. Möweninsel wreck, 8. Haithabu 3 wreck, 9. Fahrdorf wreck, 10. Reesholm Barrier, 11. Karschau wreck, 12. Kappeln logboat, 13. Olsburg.
(J. Enzmann)
2In August 2014 a fisherman, J. Ross, accidentally recovered a very well-preserved floor timber and reported the potential wreck site to the Archaeological State Department of Schleswig-Holstein (ALSH) (fig. 2). Two professional divers, G. and R. Lorenz, conducted a survey of the site in September 2014, under the supervision of J. Fischer from the ALSH, O. Nakoinz from the Institute of Pre- and Protohistoric Archaeology of the CAU in Kiel and S. Hartz of the Archaeological State Museum in Schleswig. The survey confirmed the potential of the site by lifting several loose timbers including a complete knee. Due to the poor visibility that day, no pictures or videos of the site could be taken. Based on the examination of the finds lifted, O. Nakoinz suggested a clinker-built vessel which he preliminary dated to the 11th or 12th century AD. Up to this point there was no precise data regarding the extent of the wreck site and due to fishing activities and erosion the site was endangered. While preparing for the field school Maritime Archaeological Programme (MAP) of the Syddansk University in Esbjerg and the ALSH on the Viking-age barrier at Reesholm in summer 2015, the company Innomar from Rostock surveyed the wreck site with a parametric sub-bottom profiling system. Furthermore, the Department of Geophysics of the Institute of Geoscience of the CAU Kiel experimented on the site with a new 3D marine reflection seismic system called PingPong (Wilken et al. 2019). The results showed an approximately 20 m-long anomaly in the area of the site. Additionally, some measurements of the site were taken with a total station during the field school in summer 2015. The measurements of fragile frame heads roughly confirmed the extent of the site.
Fig. 2: The floor timber that was raised in 2014 by a fisherman and led to the discovery of the Fahrdorf wreck
(ALSH/Jonas Enzmann)
3Due to the high archaeological potential of the Schlei, the AMLA of the CAU at Kiel, composed mostly of European certified scientific divers studying prehistorical and historical archaeology, concentrated their research on this region. In 2014, five members of the AMLA examined and documented a logboat, most likely from the medieval or early modern period, near the town Kappeln (Steffensen, Wilkes 2016, p. 16–19). Fortunately, the visibility of more than 1 m was extraordinarily good during that project. In summer 2015, the AMLA joined the aforementioned field school of the Maritime Archaeological Programme (MAP) of the Syddansk University in Esbjerg and the ALSH.
4Thus, in autumn 2015 the AMLA began working on the Fahrdorf wreck as a project on a voluntary basis. The timbers found up to this date were not suitable for dendrochronological dating. In agreement with the ALSH, AMLA returned to the site in October 2015. The survey and dendro-sampling turned out to be highly successful this time, due to unexpectedly good visibility of circa 0.30 m. Fortunately, one of the divers took a GoPro video camera on the dive and filmed parts of the site. Afterwards F. Wilkes was able to generate a 3D model and a photomosaic from these short clips using the software Agisoft Photoscan. The results turned out to be extraordinarily helpful.
5A plan was therefore established to develop this method further and test it at different sites. Because of the good results AMLA achieved with this method, the ALSH supported the idea and made it possible to conduct a systematic video survey of the Fahrdorf wreck site. Assuming that the visibility on site is strongly connected to the seasons, the AMLA returned to the site in mid-December. Because different cameras and settings were tried, two dive days were necessary to fully document 11 × 4 m of seafloor with visibility between 30 and 40 cm, while the visible parts of the wreck covered an area of 9 × 4 m. All exposed timbers were marked with bright yellow, numbered plastic discs. The georeferenced photomosaic was combined with the results of the parametric sub-bottom profiler and the PingPong 3D system data (as mentioned above) by using the four measured corners and aligned well with the recorded anomaly.
6From February 15 to 16, 2017 the AMLA started a new campaign on the wreck. The aim was to check the preservation and sedimentation conditions of the wreck by filming it again and comparing the generated pictures and models to the old ones. In addition, more dendrochronological samples were taken because no sample had so far provided a date. To validate some ideas about the construction and position of the wreck, a small sondage on the southern end of the wreck site was dug.
7The first objective was to establish a reference system for the diver. Luckily, only one of the markers from the previous investigation was lost, and it was possible to georeference the new video material with the other markers and the positions taken by total station in 2015. In addition to the already marked timbers, some new timbers between the old ones but further to the east could be documented. Subsequently, two divers started to excavate the sondage area, which was then filmed the next day when the water was clear of the disturbed sediment. On the basis of the video material, new 3D models and orthophotos of the visible timbers of the whole wreck site and the test trench were produced (fig. 3). Additionally, two new samples for dendrochronology were obtained from eroded planks at the southern end of the wreck.
Fig. 3: 3D model of the area excavated in 2017 generated from videos
(F. Wilkes)
8The Fahrdorf wreck is largely covered with sandy sediment and soil. Only some timbers are visible and rise up to 15 cm above the sea floor. The known length from the excavated and visible timbers is nearly 9 m. Up to the 2017 survey, eight parallel floor timbers had been discovered on the seabed. They are orientated east/west, which means that the wreck’s orientation is exactly north/south (fig. 4). A few plank fragments and both preserved ends of the southern two timbers are visible on the surface of the sediment. This area was excavated during the 2017 campaign and thus the southern quarter of the wreck could be documented. The exposing of a floor timber revealed that the ship heeled nearly 20 degrees to the east (fig. 4). It was not possible to evaluate if the excavated end is the bow or the stern. To accommodate the absence of bow and stern, port and starboard, the strakes were numbered from the keel to east (E) and west (W) and the frames from south to north. The regular interval of the frames revealed that two timbers are missing, so the last timber to the north is numbered 10, although there are only eight visible on the seafloor. Most likely one of the missing floor timbers (04 or 08) is the one that was lifted during the exploration of the wreck in 2014 (fig. 2). This lifted floor timber extended over six strakes on the right-hand side of the photo whereas on the left-hand, broken side only four strakes are visible. The central part shows a W-shaped cross section. At the position of the second and third strakes of the right-hand side two treenails remain in place. The georeferenced 3D model (fig. 3) shows that floor timber 02 had a visible length of 150 cm and is 13 cm wide at its thickest part.
Fig. 4: Overview of the currently documented features of the wreck based on computer vision photogrammetry
(F. Jürgens)
9As the previously recovered parts had already shown, the vessel was clinker-built. The boards, with dimensions of 20 × 2.5 cm, were made from radially cleft oak and finished with axes, as is indicated by tool marks. Only the fifth strake (E05) is smaller, with a width of 10 cm: it is possibly a repair plank. Iron rivets and rectangular roves were used to clench the individual planks at a constant interval of 18 cm. Because of the poor preservation it has not yet been possible to take detailed measurements of the rivets. The planking is luted with animal hair. Because of the wreck’s list, only two strakes (W01, W02) are preserved on the west side, whereas on the east seven (E02-E08) could be excavated. E01, the first eastern plank on the keel, is missing. Additionally, the extent of the surface frame heads indicates that twice as many strakes could still be concealed in the sediment.
10The floor timbers were axe-cut out of compass timbers and were furnished with notches to house the planks. They are fixed with treenails, which have a diameter of approximately 3 cm. The timbers span symmetrically across the keel and are notched for the first six (01 and 04/08) or seven (02) strakes. The distance between the frames is consistently 70 cm. Floor timber 02 has a preserved length of 157 cm. Its cross section has the dimensions of 12 by 12 cm. A previously salvaged knee and a special biti-stanchion already indicated the typical Nordic frame construction of floor timbers and above positioned lower beams, known as bitis. The test trench of the 2017 survey revealed two treenails in the planking east of, or rather above floor timber 02, which must have held such a biti construction in position. It is conspicuous that there is no hole in the floor timber, which could house a biti-stanchion. Maybe these stanchions are only used further forward, or respectively aft. Two timbers in the east of the wreck, in line with frames 04 and 05, are interpreted as futtocks, but one cannot say yet if they are used as part of a biti construction in the midships area. There are two limber holes on the bottom of the frames, which allowed the bilge water to run above the keel. The keel itself has not yet been found. It might lie deeper in the sediment, but more likely it is missing in the southern part of the wreck, as indicated by the absent strake E01. As is customary in Nordic shipbuilding, the keel was not fixed to the frames, so no nail holes are visible in timber 02.
11The length of the wreckage, known through the excavated and surface visible timbers, is nearly 9 m, but the curvature of the uncovered two floor timbers of the south end of the hull reveals that the position of the missing bow or stern must be a minimum of 3 m further to the south. A comparison of the dimensions and the shape of these floor timbers with those of the Karschau wreck, which is presented below, indicates that the original ship was approximately 20 m long and 6 m wide: the proportions of a large cargo sailing vessel. The cargo-carrying capacity of such ships is an estimated 50 tonnes (Kühn 2009, p. 82).
12Until now there is no evidence of the nature of the ship’s load. A brick found near the wreck seems to be a later addition. The position of the wreck in shallow water near the village of Fahrdorf and the absence of small finds, as with the (partly) missing keel, would indicate that the ship broke up near this place. A plank fragment containing a repaired leak, salvaged in 2015, suggests that the ship was used over a long period.
13The first three wood samples obtained in 2014 were found to contain too few tree rings, making it impossible to date them (Daly 2015). However, by adding the new samples taken in 2017, all samples could be dated due to the cross matching of some tree rings. Even sapwood was preserved on one sample. The result is a cutting date between AD 1106 and 1121 (Daly 2017). This dating fits well with the first typological classification based on the lifted timbers in 2014 by O. Nakoinz. All samples are from planks made of oak (Quercus spec.). Based on an average curve of all samples, a length of 110 years could be compared to provenance curves. The highest correlations are achieved with curves from Schleswig-Holstein and Southern Jutland, indicating a local provenance for the Fahrdorf wreck.
14Three additional wrecks from the 11th and 12th centuries are known in the Schlei inlet, giving an excellent impression of the medieval cargo shipping in this region.
15The first is the Karschau wreck, which was discovered during extreme low water in 2000 near Arnis (fig. 1). In 2001, the partly preserved portside and the keel were fully excavated, lifted and documented on land. The wreckage was 12.5 m long and had several features – floor timbers, biti-stanchion and frame spacing – similar to the Fahrdorf wreck (Englert 2015). Even the proportions and size are analogous. Only the surface ornamentation with grooves present on the wreck could not be found on the Fahrdorf wreck. The Karschau wreck was dated by A. Daly to a cutting date around 1145 AD. In addition, the Haithabu 3 wreck is similar in size at 22.1 m long and 6.25 m wide: it is the oldest of this ship type and was built around 1025 (Rösch 2018, p. 219). It was found in Haddebyer Noor, near the former Viking centre of Hedeby (fig. 1).
16Another wreck was found even closer to Schleswig and Fahrdorf at the Möweninsel (fig. 1). The wreck’s condition was very poor: the longest preserved part was the keel at 9.5 m, next to a small part (4 × 1.8 m) of the hull. This ship, with a reconstructed length of 15 m and a width of 4 m (Belasus 2009, p. 96), belongs to a medium-sized class of cargo vessels (Crumlin-Pedersen 1999, p. 18). It was built after 1163 AD, as indicated by dendrochronological dating. Similar to the Karschau wreck, it is ornamented with grooves on some planks and timbers. Particularly noticeable is that the distance between the frames on the Möweninsel wreck is only 51 cm, while it is 70 cm on the Fahrdorf example. This indicates a typological development in Nordic shipbuilding, since the older ships, relating to the Viking period, had a frame space of 0.9-1 m (Belasus 2009, p. 93–94). This change in construction might be due to the greater carrying capacity of these cargo vessels.
17All in all, the wrecks of Karschau, Möweninsel, Haithabu 3 and now Fahrdorf are a singular inventory that testifies to the rise of medieval cargo shipping in the Baltic. These ships represent the latest development in the Nordic shipbuilding tradition, starting in the Viking Age. Specific features of Nordic shipbuilding, like the bitis, vanished shortly afterwards (Nakoinz 2014, p. 109–113).
18Discovered by chance, pursued by the interests of students, the Fahrdorf wreck has grown to become an extensive research subject for the AMLA. What makes the present project so unique is that only a few have so far applied computer vision photogrammetry in waters with such poor visibility and with affordable camera equipment (Van Damme 2015). The success of the method enables underwater archaeologists to document features invisible to the diver’s eye. This possibility will undoubtedly form a future focal point in the works of the AMLA.
19The wreck itself fits right into an already well-outlined historical context. It is a large cargo ship from the end of the development process of the Nordic shipbuilding tradition. The preservation of nearly one complete side of the ship gives hope that there are plenty of discoveries left on the seafloor and that these discoveries might eventually lead to answering questions such as why and how shipbuilding changed in this region.