Maritime Archaeology Newsletter From Denmark 22, 2007

Newsletter

Maritime Archaeology

No. 22 • Summer 2007

from Denmark

A daring voyage. Havhingsten fra Glendalough sets sail from Roskilde to Dublin, 1 July 2007 (www.havhingsten.dk). Photo: Werner Karrasch.

N0. 22 • SUMMER 2007

CONTENTS:

One step at a time

One step at a time . . . . . . . . . . . . . . . . . . 2 Timber, Trade and Tree-rings . . . . . . . . . . . 3 Viking Age Iconography and the Square Sail . . . . . . . . . . . . . . . . . . 8 Between sea and land . . . . . . . . . . . . . . . 13 Gåsehage. A threatened wreck-site of the early 18th century . . . . . . . . . . . . . 18 Modelling Fetch using GIS . . . . . . . . . . . 23 Birger Thomsen 1945-2007. . . . . . . . . . . 26

Just published:

ISBN 978-87-85180-47-6 Published by the Viking Ship Museum in Roskilde. DKK 299 / EUR 40 2

The Maritime Archaeology Programme at the University of Southern Denmark is in full development. The first students came in September. The first semester passed as if in a split-second, while we were still fleshing out the curriculum. In January, the Zea harbour-project and its director Bjørn Lovén formally joined in, and in April, our team was extended with Jens Auer, who left a post at Wessex Archaeology. Jens is a diving archaeologist with extensive experience. He worked in Mecklenburg-Vorpommern and led many operations for Wessex in British and Irish waters. The second semester passed in no less of a whirlwind, with a seminar on 18 April as the programme’s official start. And now the third semester is already imminent! Sponsorships from Bauer and MacArtney underwater technology mean great support, as our two-year programme includes professional diving qualifications for those who want that option. In all this activity, one step at a time, the production of this newsletter got delayed to the moment we go off on fieldwork: in Greece, in England and locally on the westcoast of Denmark. In fact there is more than enough to report on. As can be seen on the last page, all Danish museums with maritime archaeological activities and responsibilities have now joined in and all contributions come from different parts. Thanks to the delay, however, the cover can display the most spectacular of experimental projects: Havhingsten fra Glendalough departing for its trip to Dublin and around the British Isles. Inside, you will find young researchers and leisure archaeologists alike, who present their work or critical opinions. On a sad note, Jan Hammer Larssen commemorates Birger Thomsen, who quite suddenly passed away. Thijs J. Maarleveld

Timber, Trade and Tree-rings On 9 July 2007, a PhD-thesis entitled Timber, Trade and Tree-rings. A dendrochronological analysis of structural oak timber in Northern Europe, c. AD 1000 to c. AD 1650 was defended at the University of Southern Denmark. The study has been three years in the making and has dealt with refining the method by which the origin of oak timber can be determined. The basic dataset for the study consists of individual measurements for oak, from both living trees (covering the last c. 200 years) and from timber from historic buildings and archaeological sites. This data has been accumulated by dendrochronologists working in northern Europe. It was assembled during the 1990s, through

a EU-funded project concerned with the reconstruction of climate from tree-rings. I was very kindly given permission to use this dataset for my research. The study is divided into three parts; methodology, a series of case studies, and a discussion of the results in terms of the trade of oak timber. The majority of the case studies are of the dendrochronological analyses (dating and provenance determination) of oak from ancient shipwrecks.

Methodology

The identification of the origin of oak timber from shipwrecks had been carried out before. However, it had been limited to a very wide

The B&W1 wreck from Copenhagen during excavation. The building of the ship was dated dendrochronologically to c. 1564, while an extensive rebuilding phase took place in c. 1608. Most of the ship’s timber grew in Lower Saxony or The Netherlands, but timber from Scandinavia and the Lübeck region was also utilised. Photo: Christian Lemée.

3

Map showing the result of the provenance determination of the Kolding cog. The map shows the correlation between the tree-ring curve for the Kolding cog timbers and a network of site chronologies. The circle size is determined by the t-value achieved. Note that all the high t-values appear with site chronologies from southern Jutland. Map: Aoife Daly. regional level at it used a network of large regional master chronologies. The aim of the first part of the study was the development of a method by which the provenance of timber could be identified to a more local level. The method developed moves away from the building of large regional master chronologies. Instead, tree-ring measurements are grouped within a single site, resulting in a network of ‘site chronologies’, as an alternative to the larger master chronologies. These site chronologies theoretically consist of trees from the locality in which the building or the archaeological site is located, and is based on the assumption that timber from the local area was used in these constructions. The problem of using this structural oak as the basis for the provenance determination tool is that these timbers also have a history, and we do not know how far these 4

structural timbers might have been transported before they ended up in a historical building. The strength of keeping the treering data separate in site chronologies is that timber in the dendrochronological dataset remains isolated in small groups, allowing the identification of building timber which is not of local origin. Each site in the dataset is also connected, by its geographical coordinates, to a geographic information system, so that any analysis can be readily plotted on a map, allowing an immediate visual representation of the results. This is essential not only in the interpretation of the results, but it is also highly illustrative, allowing a wide audience to readily understand the results. The living tree portion of the dataset was used to test the methodology where the geographical parameters are already

Map showing the result of the correlation between the main group of timber from B&W1 and site chronologies from northern Europe. The spread of high correlations between the ship’s timbers and sites in Lower Saxony and The Netherlands possibly reflects the diversity of sources of timber that were used in the Dutch-built ship. Map: Aoife Daly. known. A method was finally arrived at, that the region of origin of oak could be tested at three levels. The first, widest level uses a network of master chronologies. The second level is where oak is tested with all the site chronologies. The third level test is where the oak tree-ring series is tested with all individual trees in the dataset. All these results are mapped, so that for each construction being tested three maps are produced. The test that is mapped at these three levels is to see how similar the tree-ring curve, from a ship for example, is to each chronology or tree in the tree-ring dataset. The similarity is measured by a correlation statistic (t-test) and the higher the result, the more similar the ship’s tree-ring curve is to the curve it is being tested against. The map of the results then shows the t-value between the ship and each chronology. The size of

the circle on the map represents the t-value achieved. Case studies

To demonstrate the applicability of the methodology a number of shipwrecks and barrel finds were selected. The earliest of these case studies are a group of barrels found reused in wells in the town of Ribe, in west Jutland, dendrochronologically dating to the early 8th century AD. The tree-ring measurements from all three barrels match well together and an average of the three barrels matches best with a chronology from up the Rhine River in the region of Mainz. Transport of goods took place in the 8th century down the Rhine River, contained in oak barrels. These products must have been shipped to Ribe from the southern North Sea coast. Medieval ‘cogs’ make up the largest 5

The timbers of the Kolding Cog were raised in March 2001, after the hull was dismantled under water. The keelson comes up upside down and is immediately rinsed by Kristiane Strætkvern of the National Museum's Conservation Department. Photo: Line Dokkedal. group of shipwrecks that are included in the case studies. Discussed chronologically, each cog wreck that has been analysed dendrochronologically is tested using the three levels developed in the methodology. The earliest is the Kollerup cog, from northern Jutland, Denmark, dating to the 1150s, while the most recent is the NZ43 wreck from The Netherlands, from the early 15th century. Very good results are achieved for some wrecks, while less strong provenance can be identified for others. The varying success of the determination can mostly be ascribed to the fact that too few samples are analysed from a ship, or that the tree-ring series derived from the ship are too short. As we move to the later medieval period however, we encounter increasingly ships that are built of timber from a variety of sources, and the provenance determination becomes more complicated. First it is necessary to look at the composition of the ship, in terms 6

of how similar the tree-ring series from each sample are to each other. This gives an indication of how homogeneous the timber for the ship is. Was the ship built of a supply of timber felled in one area, or was the ship built of timber from a variety of sources? If the tree-ring series from the timber in the ship form a single group, then we can surmise that the timber is from a single source. If on the other hand the material indicates several groups, then the groups can be kept separate, and an average tree-ring curve for each group can be made. Each group might represent a different timber source. It is possible then to test the region or area of origin of each group. Timber trade In the analysis of the Kolding cog, made of timber felled in winter 1188-89, a single group is apparent in the material analysed. An average of 13 samples was made, and

the provenance determination was carried out using the test at the three levels, as described above. A clear indication emerges that the timber had grown around the region of Lillebælt, as the ship’s tree-ring curve matched best with timber from the town of Haderslev, in southern Jutland. The analysis of a much later ship, the socalled ‘B&W1’, found in Copenhagen harbour in 1996, gives a quite different picture. Here the correlation of the tree-ring curves from the 28 samples analysed from the ship indicate that we are dealing with several timber groups. The ship was built from oak felled in around 1564 and was modified extensively with timber felled in c. 1608. One main group can be defined, while three smaller groups are also apparent. This can indicate several sources for timber for the building of the B&W1 ship. The test of the provenance of the timber for these four groups indeed indicates that several diverse sources of timber are incorporated in this ship. Details in the construction of the ship, as observed archaeologically, indicate that the ship was built in the Netherlands. The largest group identified dendrochronologically, it could be shown, grew in the Lower Saxony region or in the Netherlands. Of the three smaller groups one matched best with material from Lübeck, while the other two indicate a Scandinavian origin. It is clear with this example that, for the late 16th century, a much more complicated picture of the timber supply for shipbuilding emerges. Another ship from this period though, the Bredfjed ship, dating to c. 1600, indicates a very homogeneous timber supply, so a neat chronological difference in timber resource availability does not emerge. Rather, what we see is that specific regions suffer from a reduction in their timber resource, necessitating import of timber from timber rich regions. In other regions, it seems, timber availability is not an issue. For the building of the Bredfjed vessel, timber is available in a nearby forest. For most shipbuilding that takes place on the Dutch coast, of which

B&W1 is an example, timber from diverse and distant locations had to be imported. The picture of the timber supply for the northern European region, over the period being studied, is the topic of the third section of the thesis. It integrates the historical and the archaeological record. The fact that building timber can have been transported, from near and far, has implications for the results emerging in the case studies, and indeed for the whole development of a methodology which relies on data from historical material which has it’s own, often untraceable history of transport. This has to be taken into account when interpreting the results of a provenance determination analysis. When looking at the northern European timber trade, it is necessary to differentiate between different timber species and different specialised timber products. While an increase in the transport of conifer species for structural building uses can be seen over the period, the usual transport of oak is in the form of specialised panelling and planking. Transport of large structural oak is the exception, not the rule. Taking the analyses of shipwrecks, where many samples are dated and the provenance of the timbers are identified; a description of the transport of timber specifically for shipbuilding is possible. Of course, we are dealing with dendrochronologically dated material. This means that the analysis of the northern European medieval oak timber trade has a chronological framework. Unlike the ships, that framework is watertight! Aoife Daly

7

Viking Age Iconography and the Square Sail The interpretation of iconographical information in technical terms is not without problems. In the research traditions that lay at basis of present-day ‘reconstructions’ of ships from the Viking Age, the iconographical material plays a relatively modest role. Nevertheless, the reconstructions are based on a variety of methods and approaches. This is partly due to the technological complexity of these vessels, and partly due to factors of preservation. While the hull itself can be more or less preserved, the sail and rigging no longer exist. As a consequence, present-day reconstructions of the rigging and the Viking Age square sail are primarily based on ethnographic evidence. The Nor-wegian square rigged vessels of the 19th century are dominant in this approach. They share a number of technological features with the Viking ships, both regarding Note the difference in the sail's aspect radio: a) a warship, depicted on the Gotlandic picture stone Riddare, Hejnum parish, dated to c. 750-900; b) a Norwegian square rigged vessel from the 19th century; c) the reconstruction of the 11th century warship Skuldelev 2, launched as full scale replica in 2004. (Lindquist 1941, fig. 80; Andersen & Andersen 1989, fig. 65; S. Nielsen del. in: Vinner 2002, p.36). 8

the hull and sail and rigging. While these Norwegian vessels – 750 years younger than the Viking ships – have become the principal material for reconstruction of the Viking Age sail, the iconography of the Viking Age has ended up in a subordinate role. Several researchers have pointed out that this development is problematic and needs further investigation. The design of the sails of our modern replicas simply does not correspond visually with the iconographical evidence of the Viking Age: the depicted sails are generally low and extremely wide, while the sails reconstructed are quadratic or even upright rectangular. Erik Nylén has pointed this out repeatedly and the festschrift entitled Klink og seil that was produced in honour of Arne Emil Christensen, contains a range of articles discussing this matter. In this contribution to the Maritime Archaeology Newsletter, I would like to present some of the research that resulted in my master’s thesis at the department of Prehistoric Archaeology at the University of Copenhagen. It is an analysis of iconographical material in relationship to rigging and concentrates on the Gotlandic picture stones and the reconstruction of the Viking ship’s sail.

The material and the objectives of the study The ship iconographic material from Scandinavia covers motifs originating from a number of rune stones, graffiti, Hedeby coins, textiles and several of the well-known Gotlandic picture stones. A detailed and systematic iconographical analysis of this material’s technological potential has hitherto not been carried through. The absence of an iconographical analysis is the more notable since recent research showed that Sune Lindqvist’s classical chronology of the Gotlandic picture stones, regarding the dating of the period with detailed motifs of sailing ships (period C/D), should be abandoned. On the basis of Lisbeth Imer’s research, this period should be dated to 7501000 AD, in particular to the 9th century. The motifs depicted ought therefore to be analyzed in direct connection to the ships of the Viking Age. Dataset and analytical description The examined iconographical material covers a selection of 36 Scandinavian motifs depicting vessels with sail. 26 motifs originate from the Gotlandic picture stones, three from Swedish rune stones, four from graffiti and three motifs derive from Hedeby coins. The latter three motifs are a representative selection among the c. 23 known Hedeby coins depicting ships. All motifs are generally dated to the Viking Age – the picture stones: 750-1000; the rune stones and the graffiti: 800-1050; the coins: the first half of the 9th century. Each motif is analyzed to the background of the following aspects and attributes: •

general

•

description

of

the

ship

and

(morphology of stem/stern, sailing direction, equipment, placing of rudder, direction of waves, wind vane); rigging and sail (position of the mast, description and number of stays, shrouds, sheets, braces, bowlines, crow’s feet/network beneath the sail, ornamentation of the sail, the aspect scene

•

ratio of the sail and of the mast/yard); crew (number of visible crew members on board and their possible weapons and/or equipment).

The table on p. 10 presents the scores. Analysis and results The analysis emphasizes several interesting points. For one thing a large proportion of the motifs depict warships, which is stated by the presence of offensive weapons and/or shields. For another the sails are generally low and wide. Their rectangular shape has a height /width ratio of 1:2. Moreover, the so called crow’s feet in the sail’s under leech make technical sense, but the mysterious and much discussed network seen beneath the sail on a few picture stones, does not. The fact is that this network only appears on picture stones from Northern Gotland. It is consequently suggested that this is a local, stylistic feature instead of a technical one. The alternative would be a very local type of rigging, with parallels in neither other Viking Age iconography nor in ethnographical source material. Elements such as the direction of waves and wind vanes indicate that these vessels were generally depicted as running before the wind. The fact that the sail’s tack is never seen fastened to the hull side supports this idea. Regarding the use of perspective, it seems that no three-dimensional effects were employed. Instead a distorted perspective is in use: the sail is seen from behind, while the hull is seen from the side. A technological development in the sail design is not perceived. Medieval iconography confirms that the low and wide sails are common throughout the Viking Age. This design is probably to be seen in direct connection to the long and slender warships. An essential unknown factor though, is the question of scale. On the picture stones, it seems as if ‘horror vacui’ has induced the artists to force and/or stretch their motifs. 9

10

Gotlandic picture stones Alskog Tjängvide I Alva Bopparve Ardre VIII Bro Eriks I Buttle Änge I (ship 1) Buttle Änge I (ship 2) Fröjel Botvatte Garda Bote Halla Broa III Halla Broa IV Halla Broa IX Halla Broa XVI Hejnum Riddare Hejnum Rings Klinte Hunninge I Lärbro St. Hammars I Lärbro St. Hammars III Lärbro Tängelgårda I Lärbro Tängelgårda II När Smiss I Stenkumla Forsa I Stenkyrka Lillbjärs III Stenkyrka Smiss I Tofta Dyple Väte Gullarve Väte Mölner Rune stones Ledberg Sparlösa Stratomta Hedeby coins Spangereid 2 Flokerudstorp Birka 526 Graffiti Gauldalen (ship 1) Gauldalen (ship 2) Karlby Löddeköpinge

Name

?/14 ?/? ?/? ?/? ?/18 ?/? ?/?

warship cargo carrier? cargo carrier?

warship? warship ? ?

1:2,2 1:3,2

1:1,6

?/14 ?/26 ?/?

warship ? warship?

1:2,5 1:2,5 1:3,5

9/18 3/6 6/12 ?/? 1/12 2/18 3/? 4/8 5/10 7/20 ?/? 6/12 13/26 ?/? 7/14 10/20 7/14 9/18 9/18 ?/12 ?/? 2/4 11/30 ?/? ?/? ?/?

Crew: visible /est.

warship warship warship ? warship? warship? warship? warship ? warship ? warship? warship ? warship warship warship warship warship warship ? ? warship ? ? ?

Vessel type

1:2,4 1:1,3 1:2 1:2 ? 1:2,2 1:1,9 1:2 1:2 1:1,9 1:2,3 1:2,3 1:1,8 1:2 1:2,3 1:2 1:1,6 1:1,6 1:2 1:2,2 1:2,6 1:2,2 1:2 1:2 1:2,1 1:2

Sail: aspect ratio

? ? 0 ?

0 0 0

0 0 0

4 0 5 ? ? ? 1 0 3 2 (?) ? 1 11 (net) 3+ (net) 1 8 (net) 6 (net) 6 (net) 11 (net) 0 ? 2 11 (net) ? ? ?

Crow’s feet

X X X

X

X

X

X

X

X X

Sheet

X X

Tack

2

1?

Braces

X? X

X

Bowline

3

2 2

2 2 3

2 2

X

3?

1

2? 1 1

X

X X

X X X

X X X

X X X X X X X

X X X X

X X

2

X X

6

X

X X

1

2

X

Forestay

2

Shrouds

X

X

X

X

X

X

X

Afterstay

A C B ?

C A C

A C C

C ? C ? ? C B C B C C A A ? A C C A A A C A A B ? C

Mast placing

11th cent. 11th cent. after 800 9th cent.

c. 825 c. 825 c. 825

11th cent. 9th cent. 11th cent.

10th cent. 9th cent. 10th cent. c. 750-900 9/10th cent. 9/10th cent. 9th cent. 9th cent. 9/10th cent. 9th cent. c. 750-900 c. 750-900 c. 750-900 9th cent. 9/10th cent. 10th cent. 9th cent. 10th cent. 10th cent. c. 750-900 c. 750-900 after 750 10th cent. c. 750-900 c. 750-900 9th cent.

Dating

In general, I conclude that the source material is reliable. One reason is the very specific maritime details shown in the motifs, another is the fact that the low and wide sail design is confirmed by independent sources – monuments, coins and graffiti – geographically covering a large part of Scandinavia and covering the period 750-1100. Implications The analysis shows that a specific class of ships from the Viking period is consistently depicted in a way that is not reflected in most present-day reconstructions of Viking ships, or in the ethnographic reference material that was used for them: Norwegian square-rigged vessels from the 19th century. The Norwegian boats share a number of technical similarities with the Viking Ships, but also visually differ from the iconography, due to their high, narrow and trapezoid sails. Left: A schematic overview of the results, based on Kastholm Hansen 2006 and in prep. Abbreviations, mast placing: C=centre; A=ahead of centre; B=behind centre.

The problem may be that the major presentday tradition to reconstruct the Viking Ship’s sail, has its starting point in wreck 3 from Skuldelev. The ethno-archaeological method that had been successful in that instance has subsequently been applied to the construction of a large number of Viking Ship replicas of widely different age, vessel type, provenance and state of preservation. Skuldelev no. 3 is of course unique, regarding traces of the rigging – the precise position of mast, tack and sheet is known. This high degree of preservation is not known from any other Viking Age vessel. In most wrecks, the only trace of rigging is the mast step, and thus it is in fact not possible to determine the length of the sail. However, Skuldelev 3 is also different from other Viking age ships we know, ‘Imme Sleipner’ – the hull is not an exact replica, though, but a combination of the Kvalsund Boat and Oseberg Ship, rigged with a square sail reconstructed on the basis of the Gotlandic picture stone Riddare, Hejnum parish, dated to c. 750-900 (Vadstrup 1993, p. 76).

11

in being a small cargo carrier from the 11th century. As such, it is certainly different from the class of ships most prominently depicted in the iconographical material. Alternative ‘reconstructions’ have been built and rigged at the start of the 1980's. These are the Krampmacken in Sweden and the Imme Sleipner in Denmark. In many ways, these are experiments to test the iconography. Whereas a number of successful experimental archaeological projects and trials have supported the dominant tradition of present-day reconstruction, these two ships show that it is also possible to use a square sail with an aspect ratio of 1:2 (h/w) on a Viking Ship. It is quite effective when running before the wind, but not effective at all beating against the wind. It may also be assumed, as I do in my thesis, that the low and wide sail has an important advantage over a higher and narrower ‘Norwegian’ sail. On the long and slender warships, it is essential to have a low centre of effort, i.e. a theoretical point, where the wind force hits the sail. Otherwise the ship might become unstable and risk to capsize. The low sail’s primary disadvantage is lack of effectiveness when beating against the wind. Trials with warship replicas suggest though, that also with a higher sail, this particular vessel type is not very effective when tacking. This applies for instance to the Skuldelev 5 replica Helge Ask. Rowing, or waiting for fair wind, seems to be the plausible solution when heading directly against the wind. Conclusions The analysis shows that the majority of the ship motifs examined should be considered as fairly reliable source-material regarding parts of Viking Age ship technology. Most of the vessels depicted are warships with low and wide square sails. This should be taken into consideration when trying to reconstruct. It is quite clear that any such attempt needs to consider both primary and secondary sources and analogies. Any reconstruction bears the risk of confusion between evidence and interpretation. Primary source

12

material, however, should be leading. The iconographical material is primary source material and should not be rejected in favour of ethnographical analogies. My research underlines the thesis that Nylén and others hold that perhaps we should perceive of the sails of Viking Age ships in a more complex way than has been usual so far. I suggest that the long and slender warships were equipped with low square sails, and that these specialized vessels were not primarily constructed to tack against the wind. Ole Thirup Kastholm References Andersen, B. & E. Andersen 1989: Råsejlet – Dragens Vinge. Roskilde. Arisholm, T., K. Paasche & T.L. Wahl (red.) 2006: Klink og seil – Festskrift til Arne Emil Christensen. Oslo. Imer, L.M. 2004: Gotlandske billedsten – dateringen af Lindqvist gruppe C og D. Aarbøger for nordisk Oldkyndighed og Historie 2001, pp. 47-111. Kastholm, O.T. 2006: De gotlandske billedsten og rekonstruktionen af vikingeskibenes sejl. Unpublished Cand. Mag. Thesis in Prehistoric Archaeology, University of Copenhagen. Kastholm, O.T. in prep.: De gotlandske billedsten og rekonstruktionen uktionen af vikingeskibenes sejl. Aarbøger for Nordisk Oldkyndighed og Historie 2005. Lindqvist, S. 1941: Gotlands Bildsteine I. Kungl. Vitterhets Historie och Antikvitets Akademien. Stockholm. Nylén, E. 1986: Thee “Krampmacken” Project. I: O. Crumlin-Pedersen & M. Vinner (eds.): Sailing into the past. Proceedings of the International Seminar on Replicas of Ancient and Medieval Vessels, Roskilde 1984, pp. 104-113. Roskilde. Vadstrup, S. 1993: I vikingernes kølvand. Erfaringer og forsøg med danske, svenske og norske kopier af vikingeskibe 18921992. Roskilde. Vinner, M. 2002: Vikingeskibsmuseets både. Roskilde.

Between sea and land Marine-Arkæologisk Gruppe in Fredericia is a group of diving recreational archaeologists that has contributed to the development of maritime archaeology in Denmark in a very special way. It is best known from the Tybrind Vig project, the excavation of a submerged Ertebølle site on the western shore of the island Funen, that it undertook under the supervision of the Prehistory department of the University of Århus. The discovery of that site and the excavation campaigns in the late seventies and eighties may have been the reason for the foundation of the group, but it has certainly not been the group’s only activity. More than thirty years later the same Marine-Arkæologisk-Gruppe Fredericia surveying on Brandsø. Photo: Hans Dal.

group members are still active. The system of heritage management has evolved during that time. Procedures for reporting, decision-making and cooperation that include the maritime areas have been developed. This meant new relations between museum and heritage authorities and professional and recreational stakeholders. Roles may have changed, but the Marine-Arkæologisk Gruppe continues its valuable contribution: survey of unexplored areas and monitoring of changes in the environment. Unlike most other recreational groups they concentrate on Prehistoric remains rather than shipwrecks. This contribution reports on survey and discovery around the island of Brandsø. Brandsø is a little island in the Lillebælt.

13

It has a surface of just about 206 ha. Some of its Prehistoric monuments have been known since time immemorial. Other sites are discovered just now. The largest monument on the island is ‘Svenskehøj’. It is a 20 m longbarrow with two chambers that lies at an elevation of 8,4 meter above sea-level. Thanks to a painting by Dankvart Dreyer another monument, however, is much better known. It is the socalled ‘kæmpehøj on Brandsø’, a dolmen with a topstone of around 3 meters and a barrow with a diameter of 14 meters. In historical documents the island was first mentioned in king Valdemars inventory of around 1200 AD, where it is listed as Øbranzø. It belongs to the county Frijsenborg & Wedellsborg and it has been used for agriculture, forestry and the nursery of Christmas-trees. In 2004 a reconstruction project was started with the aim to safeguard the island’s wildlife and ecology. Ecological The trunk well on Brandsø. Photo: Svend Amlund.

14

hedges and demarcations between plots are established for that reason. The island has undulating barrier beaches in the west and a raised bog of up to 4 m in thickness in the central part. In the east there are deciduous forests on clayey topsoil. In 1656 Swedish troops invaded the island in order to use it as a way station during the campaign towards Funen over the iceridden Little Belt. When cholera pestered Europe in the 18th century the island was declared a quarantine station for travelers. They had to stay on the island for twenty days and nights before they could travel on. The island is presently uninhabited, but previously it had a small community with a school farmhouses and homesteads. A visit and a find Marine-Arkæologisk-Gruppe visited the island during a sun-drenched weekend in September 2006. They were asked to carry through a survey of the shallow seabed surrounding the island looking for oak tree-

stumps. In this task the group cooperates with the tree-ring laboratory in Moesgård. Submerged tree-stumps from documented locations are much sought after for the creation and refinement of tree-ring curves, especially in order to fill a gap in the master chronology for the period around 4000 - 3600 BC. During this visit the group only found one trunk along the western side of the island and besides that a few scattered worked flints. It was not a great success. When all the diving gear was packed and ready it would still be another 1,5 hour before Wedellsborgs ferry could collect it. What can a diver do if he cannot survey under water? Well, at least he can look around on land... Two members of the group went to the beach barriers to examine the place where the submerged trunk had been found. Two others went southwards along the beach The dolmen (runddysse) on Brandsø. Photo: Hans Dal.

towards the south point of the island. At the waterside some worked, but water worn flint was found, including flakes and a coreaxe. Then suddenly, around the area where the shore turns into a sand beach and is just 1-1.5 m high, a dark round circle was observed. Was it a tyre? A bicycle wheel? No, it was wood! It is a well made of an oak trunk. It measures around 72-74 cm in diameter and sticks out of the beach-plain with just a few cm. It therefore shows up as a dark ring of 5-7 cm width and it is flooded by occasional waves, bladder-weed, sand and gravel. The find was immediately documented in photographs and reported to Odense Bys Museer, as Brandsø is under their heritage competence. Origin and use of the trunk well

In order to say anything useful about the well, we should know its age. So, one of the group assisted the keeper of the museum, Karsten Kjær Michaelsen, who visited the find in October and took a sample for radiocarbon

15

16

dating. At that occasion it was once more established that the well was not stave-built as had been suggested, but that indeed it was a hollowed oak tree trunk. Nobody puts a well on the waterside, as it should collect freshwater. The aerial photograph seems to support the idea that the well was made at a reasonable distance from the coastline, as it suggests that the shore is presently eroding. The vegetation behind the embankment consists of bushes and little trees alternating with meadows and stops short at the bank. If the well was made before erosion, then we should add at least 1 m to its present height. The expectation was that possibly it could be an Iron Age or a Bronze Age well. The results, however, showed that the well is much younger. It dates from in between 1400 and 1600, so the case was referred to the museum’s medieval archaeologist. Other such wells in Denmark Unlike the Brandsø well, most similar wells that have been documented, date from the Bronze Age. One example was even found in a similar environment. It is actually a little spring in the beach of the western coast of Samsø, close to the hamlet Pillemark. The spring is confined by a hollowed out oak trunk of ca. 80 cm diameter, that only raises a few cm from the beach. The site is called Ilsemade and it is subject to many folk-tales about a sacred spring. In 1967, the well and spring were excavated by J. Troels-Smith of the National Museum’s scientific department, who proved that the trunk had no roots, as he could pass his hand under its sides at a depth of around 90 cm. In the well’s filling several modern coins were found, some bone and shards of drainage piping. Clearly, it had recently been cleaned. The more surprising was the C14 date of a sample of the tree: it had grown more than a thousand years before our era, at the end of

the early Bronze Age. Another well-known trunk well was found in 1903 in the neighbourhood of Budsene on Møn. But in this case we are dealing with an alder trunk that was dug into a low-lying part of the terrain. Remarkably it contained the bones of cattle, sheep, horse, pig and dog, as well as large Bronze bracelets and belt decorations, all dating from the younger Bronze Age. A similar find was made in 1942 in Smederup south of Odder in Jutland. This well again was made in a wet depression. In this case the round well was constructed of solid oak staves. It was partly filled with stones and the remains of at least 14 pots. Like the Budsene well, it is interpreted as a sacrificial well. It dates from the early Iron Age, a couple of centuries later than the Budsene well. Although we had set our hopes that perhaps the Brandsø well would also have had such a sacred or sacrificial function, this seems to be unlikely in view of its late date. Completely different questions arise now: for which inhabitants would it have provided drinking water? Can that be reconstructed? Let us wait and see what theories will develop. Anyway, Marine – Arkæologisk - Gruppe has offered to assist in an excavation if that is decided to, for once as rubber boot archaeologists, rather than rubber suit archaeologists. Svend Amlund Marine-Arkæologisk-Gruppe Fredericia

Left: Aerial photograph of Brandsø. Photo: Frijsenborg & Wedellsborg Skov og Landbrug. 17

Gåsehage. A threatened wreck-site of the early 18th century Repeated investigations of the wreck-site off Gåsehage during the years 2001-2006 have shown that some 30 cm of the upperparts have been lost during that period. The site as a whole is threatened by strong erosion. When discovered, the wreck was very wellpreserved. The site lies southeast of Ebeltoft. The sea in that area is notorious for its strong currents and many a ship has come to grief. The ’Gåsehage’-wreck was actually the first archaeological wreck-site to be reported in the area. This was in 1989 when B. Nielsen found the wreck while diving. Since then many other wrecks have been recorded. This is mostly the result of the tenacious surveys of the sportsdiver and recreational archaeologist M. Vendelbjerg and his prolonged ‘Århus Bugt’-project. In 1990, the National Forest and Nature Agency investigated the site in cooperation with the Danish Sportsdiving Association, the museum in Ebeltoft and Moesgård Museum. An auger, sacks with grain and matting were among the most notable finds. The ship was assessed as an oak-built vessel with flush-laid planking from the start of the 18th century. It had been around 20 m long and 5 m wide. The cargo defined it as a freighter. During the yearly field-course that was organized by the Agency in 1991, the site was investigated again. This time limited excavation took place. Wooden ladles with markings, leather seaman’s boots, clothing and clay pipes added to the find material of grainsacks and matting. The presence of such personal belongings indicates that the ship was abandoned in a hurry. The dynamic currents cause intermittent deposition and erosion. In some years the wreck has therefore lain exposed, whereas in other years it remained covered. When it was exposed, considerable decay has 18

occurred as a result both of macro- and of micro-organisms. In order to understand that process, Moesgård Museum has carried through a monitoring program since 2001. Investigations 2001-2006

The aim of research from 2001 through 2006 has been to document and access the condition of the wreck by monitoring exposure and decay. The picture that emerges is one of rapidly accelerating decay as soon as wooden remains are uncovered, whereby the anaerobic condition in the sediments with low oxygen content is exchanged for the aerobic flow of oxygen rich water in the strong currents around Gåsehage. In order Changes in sediment level 2001-2006. Level Timber

Sediment 1300 cm

1990-1991

14-12-2001 18-12-2003 1400 cm 10-10-2006 Deposited sand 2002-2003 10-10-2006

Sediment level forward 14-12-2001

18-12-2002 Scour

1500 cm

1990-1991

Sediment level aft 14-12-2001

to assess wood-decay, a controlled penetration method is deployed, in which a specific steel probe is pressed into the wood under a given and preset pressure, after which the penetration depth can be measured. A detailed description of the method is available in report no. 6 in the series Konservering og naturvidenskabelig afdeling of 2004. Assessment of sedimentary conditions around the wreck At inspection in December 2006, the Gåsehage wreck was partly covered in sediment. Only the tops of vertical timbers and some horizontal planking and ceiling were exposed. Timbers and planking that had been recorded previously were missing. The surface of the surrounding sand bottom consisted of white sand with a few small stones close to the wreck. The stern part protruded most from the surroundings, to a maximum height of 50 cm, whereas part of the bow had been ripped of the wreck in such a way that the remains of the stem were fully covered in sediment. The sedimentation level amidships showed a slight deepening of around 10 cm. The estimated sedimentation level on the sketch shows a rise of about 10cm in the 2006 investigation. It can be assumed that the sediment level is slightly above average, considering the observations at the measuring points and the progressing decay. In general, the sediment level around the ship as measured in 20012006 seems to be unchanged as compared to the level of 1990-1991. Assessment of the condition of wooden parts The investigation of 2006 results in a division of two vaguely distinguishable zones of decay of the exposed vertical timbers, whereas the investigation of 2001 resulted in three sharply defined and distinct zones. All the surfaces of wooden parts have now degraded through vigorous macroscopic decay by marine borers. It should be remarked that this destruction strongly stimulates total degradation as the many

holes and tunnels exponentially increase the exposed surface of wood structure. This means that the negative influence of microscopic organisms such as fungi and bacteria is also given a boost. There are fewer barnacles since the investigation of 2001. The exposed wood is both soft and perforated and provides less solid footing as a substrate. There are no areas where wooden parts have been freshly exposed. Just like in 2001, the condition of the stern part is best. Along the ship’s sides one can observe advanced degradation. Both sides look tattered and ragged. Decay progresses quickly. The wood structure is highly degraded. It has a low solid matter content. All exposed and accessible parts are visibly ruined. There are no accessible parts with original tool marks. The Mast amidships is clearly eroded. During the period 2001 through 2006, some 8-10% of the upperparts of wreck have been lost. Unlike in 2001, no previously unseen parts had been uncovered and so no ’fresh’ wood had been exposed. Probing The probing series to test resistance and penetration of the wood was planned to be executed uniformly. It was attempted to get four to seven probings of the original wood surface at a preset distance of 10 cm. In 2001, an in-between distance of 6 cm was used. Position of the probing and the wood species were recorded. In 2006, the number of probings was limited by the advanced degradation of the wood. Only eight of the 15 series that had been recorded as A through O in 2001 could be repeated as a result of increased decomposition. At positions D, E, F, I, J, L and O the planks and timbers had been teared to such extent that the measurements had become impossible. The penetration depth at the 8 predefined measuring areas was generally very deep. The wood had decayed so much that the penetration depth regularly exceeded the instrument’s measuring range. This is in stark 19

20

Conclusion

C D

A B

N Section AA E I

J H

G Section BB F

K L

N M O

Visible wreck parts and the position of the probings. Drawing: Claus Skriver. contrast to the 2001 results. Penetration was then generally very low in the vertical sides. That indicated that the wood was still in good condition. The tables show where the strongest decay occurred: the sides and the upper parts. If we compare the penetration at the investigations in 2001, 2002 and 2006 we can see that the surface-degradation has significantly changed over a relatively short period of five years. Stable surfaces with a few millimeters penetration changed into strongly degraded wood with a soft surface. Top left: Timbers and planking. Photo: Claus Skriver. Bottom left: The mast stump is attached by organisms. Photo: Claus Skriver.

The observations and measurements provide us with a consistent picture of the conservation condition of the Gåsehage-wreck. It indicates a range of environmetal factors that further degradation. Since 2001, some 30cm of the upperparts have been lost. Moreover, the state of preservation of the exposed wood surfaces has changed from being well-preserved to strongly degraded with a high level of probe penetration. The documented process confirms degradation or formation processes as we know them. Wooden wreckage that protrudes from the sand bottom breaks down from top and sides after which it fractures. As a consequence wreckage becomes dispersed over a larger area and it will disappear after a winterstorm. The registered decay helps to predict that large pieces of wreckage will break loose. The exposed parts will disappear as a result of the dynamic currents in the area. We may hope that the present sea-bottom conditions are relatively stable, so that the sediments will continue to be at the present level. The wreckage at this level will then be covered by some sand. All exposed wood will disappear in a few years. Any research of the wreck should therefore be undertaken as soon as possible. Protective covering of the wreck

During the excavation in 1991 part of the wreck was uncovered. Sediments where removed from the wreck and its surroundings. After the campaign the area within the wreck was covered with sandbags. At the 2001 inspection some destroyed and worn sandbags were observed (presumably these were of polythylene). At later inspections no traces of sandbags were seen. Presumably some are still buried. If only a future covering of wreck-sites in current ridden channels would be possible! Experience in Denmark and abroad to trap sediment with plastic netting or artificial sea-grass shows that direct protection of degradable organic materials is possible. Experiments with different systems 21

mm penetration H I J K

Cm above groundlevel 0

A

B

C

D

E

F

G

L

M

N

O

-

-

-

-

-

-

-

-

-

-

-

-

10

10

10

12

20

10

20

7

5

20

20

10

15

20

17

20

10

20

30

20

20

12

40

100

80

15

30

40

20

10

12

80

15

>100

50

30

20

>100

30

60

>100

-

-

-

Fig. 7. Sedimentniveauet er ved 2006 undersøgelsen hævet ca.10-15 cm over bundniveauet for 2001 og 2002 undersøgelsen. 2006

Top table: Summary of measurements in 2006. The sedimentation level in the 2006 investigation is raised appr. 10-15 cm above the ground-level for the investigations in 2001 and 2002. Thus, the 2006 measurement points start at the 10 cm measurement points. Measurements for D, E, F, I, J, L, and O, could not be carried out as timbers and planks were missing at these positions on the Gåsehage wreck.

Right table: Comparison of measurements of resistance and penetration in succesive years. In 2001, the distance between each penetration was 6 cm. This was changed, and from 2002 the distance deployed is 10 cm. in the area around Gåsehage would give a solid basis to decide on what works best. Such experience can than be applied to other wreck-sites under conditions that are similar to those at Gåsehage. Jesper Frederiksen Claus Skriver

References

Manders, M.R. & Th.J. Maarleveld, 2006: Managing the Maritime Heritage under Water. The Choices We Face. Proceedings of the National Service for Archaeological Heritage in the Netherlands 46, pp. 127139. 22

Series cm above groundlevel 0

A 2001

6

2

2

mm penetration A A 2002 2006 4

-

6

10

5

20

5

20

40

6

20

50

5

30

60

5

>100

70

12

10 12

5

18

5

20 24

6

30

5

36

Fig. 6. Indstiksmålinger for måleserier A er i tabel samlet for 2001, 2002

Moesgård Museum. Indstiksmåling af nedbrydning i marint arkæologisk træ. Konservering og naturvidenskabelig afdeling, rapport nr. 6, 2004. Moesgård Museum. Gåsehageskibsvragets tilstandsrapporter 2002-2007.

Modelling Fetch using GIS The ocean is a highly dynamic environment, a simple fact which has tangible impacts, both on the maritime communities which lived at the sea and by the sea, and on the maritime archaeologists who try to gain knowledge about these communities, their lives and implements. As such it is important to develop and have access to basic data on the dynamics of the oceans. The following is a presentation of the development of GIS based calculation of wind-wave fetch in the Danish waters. Data and methods Fetch is the length of open water over which the wind can generate waves. It is an important factor in understanding wave height, coastal erosion and longshore drift. These basic factors are important to maritime life A section of the map showing the layout of points.

and maritime material culture, both in the past and in the way they are handed over to us. Having data on fetch is therefore useful. Producing a GIS layer of fetch is a matter of measuring the distance from any given point on the sea surface to the coastlines. To define the coastlines the Digital Chart of the World (DCW) was used for all countries surrounding Denmark. This map collection is in a nominal 1:1,000,000 scale, and is the best available public domain GIS coverage of the world. To better model the details of the Danish coastline however, this area was replaced with a coastline in scale 1:25,000, digitized by Danmarks Miljøundersøgelser (The Danish National Environmental Research Institute). Points were laid out in a regular hexagonal grid, the hexagonal format chosen to ensure that each point had an equal distance to all neighbours. On the open waters the distance between points was 10 km, while it was set to 1 km within the nearest 5 km of

23

the coastline. This gave a total of more than 21600 points, all set by a small MapBasic macro to automate the process. To measure the distance from each point to the coastlines, lines were drawn from each point in eight directions with 45° increments. The length of each line was set to 1000 km. Using only eight lines to cover the entire compass, the result is admittedly only a proxy, but it was estimated that the marginal effect of adding more lines was relatively limited, and that the chosen resolution would give a sufficiently close estimate to the real values. An important consideration here was the sheer amount of data. This resolution gave almost 173,000 individual lines to start with, and during the process of calculation, the numbers exceeded 1 million, making processing very slow.

24

Finally the ‘compass’ lines radiating from each point were cut along the coastlines. As coastlines tend to be complicated across the entire 1000 km of each line, this gave many large and small pieces of lines, which is why the number of lines rose dramatically during this process. In the end only those sections of the lines which were in contact with the original points were retained, leaving the original number of c. 173,000 lines. Summary statistics on the length of these lines were updated to the points, and the results for maximum fetch interpolated to a continuous surface with a resolution of 100 m. The model of wind-wave-fetch. Bright red eqauls > 1000 km.

Results The resulting GIS-model immediately shows the impact of fetch in shaping the Danish archipelago. Where fetch is high, the shorelines are linear, while the meandering coastlines and many smaller islands are found where fetch is reduced. The potential impact on maritime culture from this relation is perceptible, and indeed the known medieval harbour sites and towns of the area are all found in sheltered areas with low fetch. Using the model The model of wind-wave fetch has bearings on at least two analytical directions in Maritime Archaeology. First, it is a basic variable in describing the environmental dynamics of the ocean, and as such it is important for understanding and modelling processes of wreck preservation and discovery. This research path was also touched upon by David Gregory in the previous issue of this newsletter. Secondly it can be used as a variable in understanding the Maritime Cultural Landscape, as the map shows important aspects of conditions on the shore, and indeed basic conditions for navigation and the location of safe anchoring and landing sites. As an example of the first use, a preliminary attempt to model ship preservation

was made as part of the course in IT and remote sensing on the Master programme in Maritime Archaeology at the University of Southern Denmark. Continuing the thoughts of Muckelroy of shipwrecks and environment, and applying more modern techniques of predictive modelling, the students compared four environmental variables (fetch, salinity, sediment type and sea-bed profile) to the presently known wooden wrecks older than 1850 in Danish waters. Using wreck sites, discovery is used as a proxy for preservation, and acknowledging that there is not a linear relation between the two, a model was made using Dempster-Shafer theory, as the purpose of the exercise was more to train the technique than to complete a full study. Nonetheless the model did seem to give a general indication of the preservation of shipwrecks, as the observed/expected ratio of wrecks rose exponentially with higher values on the model. On the highest values on the predictive map, wrecks were overrepresented by a ratio of 5:1, while the lowest values had a ratio around 1:10. This gives an inkling on the feasibility of such studies, and that more detailed GIS-based studies on the relation between environmental conditions and wreck preservation are probably worth while. Bo Ejstrud Left: A predictive model of wrecks in Danish waters. Bottom: The relation between model values and observed/expected ratio of known wrecks. An exponential curve is fitted.

25

Birger Thomsen 1945-2007

On March 16, maritime archaeologist Birger Thomsen died – all too soon. A virulent cancer put a sudden stop to an active and committed life in Danish maritime archaeology. With a background as historian and ethnologist Birger was employed by the Conservation Authority (later the Forest and Nature agency) in 1983. His job included protection at sea, which included to explore and record sites in Danish waters. In practice, he concentrated on historical wreck. Birger’s name will remain associated with that pioneering work. With great dedication he contributed to the creation of the socalled FREDSØ-register, a nation-wide database on cultural remains on the seabed. The register contains information that is important for professional maritime archaeological work, both in the office and at sea. Adequate 26

survey, reliable position fixing and adequate recording were the factors that Birger stressed in its creation: if we are to deal with heritage in relation to developments and planning at sea in a professional way, then the information should be correct and position coordinates should be traceable and convertible. The FREDSØ-register allows one to combine information gathered in the field with archival data such as the archives dealing with loss at sea. In many instances it has been possible to identify shipwrecks in that way. It was a specific and really good tool. When the Forest and Nature agency transferred its competences and registers to the National Museum is was decided not to continue with the FREDSØ-register, but to integrate information in DKC, the central register of cultural-historical values. The format was different and not particularly suitable for maritime information. Data was taken out of context and the conversion of geographical positions to UTM co-ordinates introduced so much uncertainty in the register that it severely undermined its professional usefulness. That was a great disappointment for Birger. Trips with the dory, the Terne, the Ryle, the Havterne and a wide variety of inflatable dinghies became one of Birger’s trademarks. Everywhere in Denmark positions were taken and checked, and sites were surveyed and recorded. Around Anholt for instance, one of the waypoints that caused so many to wreck, just like Skagen or Læsø. It was there that Thomsen was in his element, while developing new ways of survey. Once, for instance I was asked to join him on Anholt with a pair of Polaroid sunglasses. Using them, dark patches of sea-bottom stood out much clearer, which was a simple

Maritime Archaeology

but highly effective way of visual survey up to water depths of 6 m. This did not, by the way, mean that Birger rejected electronic survey. He experimented with seismics, magnetometer and advanced echo-sounders. Birger Thomsen readily made his extensive knowledge and experience available to sportsdivers and other players in the sea area. The trips with Lilla Dan and Fulton were the platform in which learning, seamanship, fellowship and maritime archaeology went hand in hand. Those trips were near his heart – and he led several. As recently as October 2006 Birger and I made a trip along the west coast of northern Jutland in order to assess the protection of 8 historical wreck sites. It was to be Birger’s last trip – but the commitment, the interest and the wondering were fully intact. He simply had to check the positions in the maritime register in the real world. During that trip we spoke about Birger Thomsen’s archive – and where, one day, it would have to end up, when he would be ready to sign off. Neither of us knew how soon that would be. It was Birger’s wish that Bangsbo Museum og Arkiv would take over the extensive data material in order that Danish maritime archaeology will benefit from it in the future. Let us cherish his memory Jan Hammer Larsen museumsinspektør og marinarkæolog Bangsbo Museum

Newsletter No. 22 • Summer 2007

from Denmark

ISSN 1902-0708 EDITORS: Thijs J. Maarleveld & Helle Kildebæk Raun Lay-out: Jens Lorentzen & Ewa Britt Nielsen DTP: Helle Kildebæk Raun PRINT: Athene Grafisk Aps., Albertslund © Centre for Maritime and Regional Studies and authors 2006

Maritime Archaeology Newsletter from Denmark is a continuation of Maritime Archaeology Newsletter from Roskilde, Denmark and is published twice a year by: The Maritime Archaeology Programme, University of Southern Denmark at the Centre for Maritime and Regional Studies Niels Bohrs Vej 9 • DK-6700 Esbjerg Tel. +45 6550 4177 • Fax +45 6550 1091 e-mail: [email protected] The Newsletter is supported by: University of Southern Denmark Centre for Maritime and Regional Studies Fiskeri- og Søfartsmuseet The National Museum of Denmark The Danish Institute in Athens Langelands Museum The Viking Ship Museum Ringkøbing Museum Strandingsmuseum St. George Bangsbo Museum Moesgård Museum Haderslev Museum

In June 2007, students from Esbjerg and Southampton jointly trained in Roskilde. Photo: CMRS.