a chieftain's hall or a grave; ground penetrating radar in an
TRANSCRIPT
Abstract—Stiklestad’s rich archaeological and historical landscape contains ritual, mound and settlement sites some of which are recorded in Norse chronicles. The death of King
Olav in the Battle of Stiklestad in AD 1030 represents a turning point in Norwegian history marking the transition from paganism to Christianity. Interpretation of oblique aerial photography taken in 2007 identified a possible site for a long sought after Chieftain’s Hall on a farm at Haug near Stiklestad which belonged to one of the King’s servants at the time of the battle. The nature and form of the cropmark led to speculation as to whether it represented a Hall or a grave. A research programme of geophysical survey led to the use of ground
penetrating radar followed by a test excavation. These revealed a complex series of features some 50m in length and up to 15m in width that have been interpreted using comparisons with visible and excavated, longhouses, halls and graves. The overall results show that interpreted geophysical and excavated features have attributes which can be assigned to a building or a grave or in some cases both.
Index Terms—Archaeology, Cropmark, Excavation,
Geophysical Surveys, Ground Penetrating Radar, Norway,
Stiklestad.
I. INTRODUCTION
The battlefield at Stiklestad, Mid-Norway (Fig. 1) is a
central place in Norwegian history. The Stiklestad area had
been a stronghold for regional chiefs for a long period. The
Christian King Olav Haraldson was killed in a battle in AD
1030 in an attempt to conquer a pagan opposition in Central
Norway. The church canonised King Olav and he soon
became the most popular saint in Norway. This battle is very
famous in Norwegian history due to the sagas of Olav
Haraldson written by Snorre Sturlason. The battle marks the
transition from the Iron Age to Medieval Times in Norway.
Until recently, no one tried to explain why Stiklestad was
chosen for the final battle for the Norwegian throne [1]. A
study of the archaeological landscape surrounding Stiklestad
has however revealed a remarkable concentration of
monumental grave mounds. The power of the local chiefs thus
can be traced back to the Iron Age; this is demonstrated by
perhaps Norway’s largest concentration of monumental grave
mounds.
The largest of these mounds is between 40-50m in diameter
and 6-7m high. None of them has been excavated and due to
their size, it is almost impossible to carry out traditional
archaeological excavations. Geophysical methods, in
particular magnetic gradiometry with a depth of investigation
of less than 1m, have been tried on some of them and a feature
inside one of them is interpreted as a stone chamber [2].
II. DISCOVERY AND INTERPRETATION OF THE CROPMARK
Aerial photos have revealed cropmarks that have led to
discoveries of many graves that were thought to have been
destroyed in the past [3]. Most interesting was the cropmark
discovery in summer 2007 of a possible house site ca 50m
long and 15m wide with oval walls (Plate 1a). These walls
were represented by two gently curving areas of enhanced
crop growth that could be due to outer southern and northern
ditches (Plate 1b). This possible house is similar to a special
type of house called a “hall”. The hall had an administrative
function as well as being an arena for feasts and religious
services. The house site is located on a nearby farm, at Haug,
known in written sources as the farm of one of the King’s
servants at the time of the battle at Stiklestad.
Halls are known on several important sites in Scandinavia.
At Gudme in Denmark there is a 50m long hall dated to the
third century, several finds of the same size in Sweden are
described as halls and recently a hall measuring 35m x 11.7m
from the Viking Age at Huseby near Sciringsal in Vestfold has
been excavated [4].
The Stiklestad area seems to have an accumulation of
power symbols from prehistory. One problem is however that
there have been very few archaeological excavations at
A Chieftain’s Hall or a Grave; Ground Penetrating Radar in an Archaeological
Geophysics Survey to Target the Excavation of a Cropmark near Stiklestad, Nord-Trondelag,
Norway.
5th International Workshop on Advanced Ground Penetrating Radar IWAGPR2009,University of Granada,27th-29th May 2009
Kevin Barton Earthsound Associates,
Convent Road, Claremorris, Co Mayo, Ireland.
Lars Stenvik & Brynja Birgisdottir, Norwegian University of Science and Technology,
Section of Archaeology and Cultural History, Trondheim NO-7491, Norway.
Stiklestad. There have been several surveys of visible
monuments but we do not know their age or their contents. In
connection with plans for new roads, settlement traces from
the Pre Roman Iron Age, Roman Iron Age and the Migration
Period (400BC-600AD) were excavated in 2008 [5]. The
excavation revealed extensive prehistoric agricultural areas as
well as the remains of four longhouses. The largest of the
houses was only partially preserved but does show a large
building, possibly 40 metres long and 8 metres wide (Fig. 2).
The excavated postholes were nearly a metre in diameter and
depth, indicating a large, tall building (Plate 2). In a central-
Norwegian perspective this is a large building and can indicate
a settlement of high social status. Findings of ornamented
pottery sherds as well as a Roman glass sherd indicate that the
settlement was of a high economic status. The excavation
merely revealed a small section of the prehistoric settlement at
Stiklestad.
The discovery of the cropmark at Haug (haug means
mound) led to speculation amongst archaeologists as to the
type of archaeological feature that lay beneath the ground. The
speculation, based on excavation results (Fig. 2) and the
visible archaeology in the locality, focused on the cropmark
being due to a house or a grave. In Ireland, in the absence of
excavation, the use of multi-method geophysical survey [6]
can be the answer to the challenges we meet in identifying
important elements in the landscape. The discovery of the
cropmark, and the archaeological speculation arising from it,
provided us with an opportunity to carry out and evaluate a
multi-method geophysical survey in Norway. The objectives
of the survey that was carried out in summer 2007 were to
define the spatial footprint of the cropmark, to investigate its
sub-surface structure and target a test excavation to identify
the source of the cropmark.
III. INITIAL GEOPHYSICAL SURVEYS
The cropmark was found in a field sown with barley. The
field has been subject to intensive cultivation for many years
and there is sandy topsoil forming an agricultural layer some
25cm in depth. Initial magnetic gradiometry surveys did not
reveal a coherent pattern that directly correlated with the
cropmark. In order to assess the potential magnetic response
of the topsoil a detailed topsoil magnetic susceptibility survey
was carried out. This survey showed the area of the cropmark
had a susceptibility enhancement when compared to its
immediate surroundings. The magnetic susceptibility response
encouraged us to carry out a second magnetic gradiometry
survey at a higher spatial resolution than the initial survey.
This survey revealed a series of possible postholes or pits that
formed two gently inward curving arcs that ran ‘parallel’ to
the long axis of the cropmark. The northern arc lay in the
northern ditch whilst the southern arc lay in the zone between
the southern and northern ditches. The interpretation of
postholes raised the possibility that the cropmark represented
a building.
In order to further investigate the possible postholes and the
ditches interpreted from the magnetic surveys an earth
resistance survey with a depth of investigation of less than
0.5m was carried out (Fig. 3a). This survey confirmed the
presence of ditches as low resistance anomalies caused by
increased water retention and mapped other features in the
vicinity that have been interpreted as low resistance cooking
pits and possible small diameter graves. The latter are
indicated by the high resistance anomalies found to the
northwest and southwest of the cropmark (Fig. 3a). The high
resistance could be due to largely stone construction. The
survey did not confirm the presence of postholes.
IV. GROUND PENETRATING RADAR SURVEYS
The results of the magnetic and earth resistance surveys
together with the sandy nature of the soil indicated that ground
penetrating radar survey (GPR) might be able to provide a
three-dimensional view of the source of the cropmark and
investigate the relationship between the sub-surface features
interpreted from the other geophysical surveys. A test or
reconnaissance survey was carried out using a Sensors &
Software Noggin Plus with a 250 MHz antenna system
mounted on a cart. The transmitter was triggered at 2cm
intervals by an odometer or survey wheel. The earth resistance
survey data was used to position 30m long GPR
reconnaissance transects which were spaced 5m apart (Fig.
3b). The data were processed using Sensors & Software
EKKO software and four selected time sections 3, 5, 7 & 9 are
shown in Fig. 4.
The resistive, sandy soil provided good propagation
conditions for the GPR survey which resulted in time sections
containing a number of features that could be interpreted as
being due to sub-surface archaeology. Transect 3 (Fig. 4a) is
at the western end of the cropmark where it is truncated by a
relic field boundary (Plate 1b and Figs. 3a & 3b). The
southern ditch is seen between 15 and 20m along the section.
The northern ditch is not seen in this section. Transect 5 (Figs.
3b & 4b) shows both the southern and northern ditches
separated by a platform in which there is a small, central
depression which may be caused by a pit. In the section, at the
base of each ditch, are a number of narrow hyperbolas which
may indicate a point source such as a stone or boulder.
Transect 7 (Figs. 3b & 4c) is similar in appearance to Transect
5 except there is no distinctive central depression. There are
narrow hyperbolas again at the base of each ditch perhaps
indicating that the ditch has a stone lining. Transect 9 (Figs.
3b & 4d), near the eastern extent of the cropmark, is broadly
similar to Transects 5 & 7 including the narrow hyperbolas at
the base of the ditches. In this section the southern ditch, seen
at about 5m along the section, is considerably reduced in
width and depth. This correlates well with the weaker
response seen in the earth resistance data (Figs. 3a & 3b) in
this locality.
The clarity and apparent depth of investigation of the
reconnaissance GPR time sections encouraged us to further
investigate the complete cropmark using GPR. A very detailed
3-D survey was carried out over the entire 50m x 30m area in
order that horizontal time slices could be created. A series of
200 transects each 30m long and spaced 0.25m apart was
collected using the Noggin Plus GPR with 250 MHz antenna.
The resulting dataset was processed using GPR-SLICE
Software [7] with 65 successively overlapping slices each
averaged over 3ns intervals being created. Horizontal slices 2,
16 & 33 illustrating the major interpreted archaeological
features are shown in Fig. 5.
Horizontal time slice 2 (Fig. 5a) shows the response of the
25cm deep agricultural layer which contains a strong WNW –
ESE ploughing overprint. At the NW of the slice the plough
traces are truncated against the field boundary (Plate 1b) and
change direction to the NW of the boundary. Horizontal time
slice 16 (Fig. 5b) clearly shows the top of the south and north
ditches. At about 1028E there is a south to north thin vertical
line which cuts both the truncated cropmark and the field
boundary. This is a buried pipe. Outside the ditches there is a
strong ploughing overprint. Between the ditches is a complex
pattern of small, coherent reflectors which could indicate
stone cobbles or boulders which might be indicative of a floor
or stone base. Horizontal time slice 33 (Fig. 5c) shows the
response at or just under the ditches. There is a thin pattern of
coherent reflectors that describe the shape of the cropmark.
These relate to the hyperbolas seen in the time sections (Figs.
4b, c & d). This pattern may confirm that the base of the
ditches are stone-lined or have a hard base. On the northern
side of the central part of the feature is a cluster of coherent
reflectors. This cluster could represent a ploughed-out or
partially destroyed base of a chamber.
In order to both test the geophysical interpretation and
establish whether the mapped feature is a house or a grave or
both, a test trench was opened in September 2008. The
location and orientation of the trench was based on the earth
resistance image and GPR horizontal time slice data. The
trench was positioned to intersect the ditches perpendicularly
in order to obtain their true width and also to avoid any nearby
features which might not be directly related to the footprint of
the cropmark. The trench location is shown on the horizontal
time slices (Fig. 5).
V. EXCAVATION
The 2m x 25m test trench [8] orientated NNE-SSW was
opened across the cropmark (Fig. 6). The trench was dug by
hand in order to observe any anomalies or features in the
agricultural layer. The layer was homogenous brown sandy
soil and could easily be distinguished from the more coarse
sand and gravel in the subsoil. In the agricultural layer, there
was some modern waste from the modern cultivation and
nearby farms, as well as scattered burnt stones, charcoal and
burnt clay which could represent remains of a prehistoric
settlement. Due to the humidity and colour of the subsoil the
ditches were not visible in the trench immediately after it was
dug. When the subsoil had dried for a few hours the ditches
appeared as wide slightly darker areas across the trench. The
ditches were excavated stratigraphically in 1m wide sections.
The ditches were 3-3.5m wide at the top and narrowed down
to about 30cm in the bottom at 100-130cm depth. The ditches
were filled with sand and turf in horizontal layers (Fig. 6). The
northern ditch seemed to be lined with wood in a horizontal
layer in a middle level of the filling. The origin of the wood is
not yet known. Scattered in the sandy layers and in larger
patches underneath the wood was an iron pan/iron oxide.
Between the ditches there was a platform of lighter fine
grained sand (Plate 3). On top of the platform and cutting into
it were several features. The features were not postholes as
expected. They can be described as clusters of burnt clay,
charcoal and burnt stones. These features can be interpreted as
disturbed remains of hearths and cooking pits. Some of the
burnt clay was apparently the remains of daub, with markings
of small posts and branches and therefore represents
prehistoric dwelling on the site. There are other cropmarks
and known remains of longhouses in a nearby field. No
features that would clearly belong to a large Chieftain’s hall
were found in the trench. The ditches are too wide to be wall-
trenches as they appear in the excavation. They could yet
prove to be the remains of drainage trenches surrounding a
large building. However, the field in question is very dry and
there are no modern agricultural drainage trenches.
In size and shape the ditches resemble well known foot-
trenches which surround and mark some large grave mounds.
An example of a 50m long boat-shaped mound with foot
trenches is found at Sømna in Northern Norway (Plate 4).
Currently there is no visible or firm excavation evidence of a
burial mound on the site, although the name “Haug” might
suggest otherwise. In the top soil, a small ceramic bead of an
Early Iron Age type was found. The bead could be from a
disturbed grave on the site or the ploughed cultural layer from
the prehistoric settlement in the area.
VI. CONCLUSION
The interpreted geophysical data over the cropmark
confirmed the presence of ditches and indicated some pits,
areas of burning, a possible stone floor and base of a chamber.
Some possible small graves may be found outside the
cropmark. The interpreted basal “stones” were not
encountered in the excavation of the ditches. It is possible that
the narrow hyperbolas are a response to iron pan, an example
of this was discovered near the base of the northern ditch.
Alternatively the hyperbolas may be an artefact of the width
and geometry of the ditches as they narrow to 30cm at the
base. The GPR surveys did not confirm the presence of
postholes. The excavation revealed the ditches, pits, burnt
areas and hearths. The excavated test trench was too narrow to
provide a satisfactory image of the main feature, and the
hearths could not be securely linked to it such that a building
or a grave could be defined.
The geophysical surveys and test excavation did not
decisively prove the cropmark is due to a Chieftain’s Hall or
due to a grave. The combined results indicate the features
have attributes which can be assigned to a building or a grave
or in some cases both. It will be necessary to excavate a
larger area in order to get a better context image of the ditches
and other interpreted features on the site. A larger scale
excavation is planned to take place in autumn 2009.
Samples from the ditches as well as the hearths are
currently being dated and will provide important information
in order to interpret the archaeological features excavated.
VII. ACKNOWLEDGMENT
Eskil Folstad, Stiklestad Nasjonale Kultursenter AS, for aerial
photographs of the cropmark and for participation in the excavation. Chris Randolph, Earthsound Associates, participated in the geophysical survey and excavation. Heather Gimson and James Bonsall, Earthsound Associates, participated in the geophysical survey.
REFERENCES
[1] L. Stenvik, ”Jernalderen,” Trøndelags Historie. Landskapet blir
landsdel. Fram til 1350. Trondheim, 2005.
[2] R. Binns, ”Gradiometer avdekker stor gravhaug på Stiklestad,” Spor
nr 1/2004. Trondheim, 2004. [3] L. Forseth, “Flying into the past in Nord-Trondelag,” AARGnews,
35, September 2007.
[4] D. Skre,” Kaupang in Skiringssa,” Kaupang Excavation Project. Publication Series, Volume 1. Norske Oldfunn XXII, Oslo, 2007. [5] B. Birgisdottir, ”Stiklestad 2008, Rapport fra arkoleogisk utgraving
på Stiklestad Østre 29/1, Verdal k. N-Trl,” Unpublished report. NTNU. Trondheim, 2009.
[6] K. Barton & J. Fenwick, “Geophysical Investigations at the Ancient Royal Site of Rathcroghan, County Roscommon, Ireland,”
Archaeological Prospection, 12, 3-18. Wiley : Chichester. 2005. [7] D. Goodman, “GPR-SLICE v5.0 Ground Penetrating Radar
Software,” Unpublished User’s Manual, Geophysical Archaeometry Laboratory, 20014 Gypsy Ln, Woodland Hills, CA 91364, USA.
2008. [8] B. Birgisdottir, ”Rapport fra Arkeologisk forskningsgraving. Haug
24/6, Verdal k. N-Trl,” Unpublished report. NTNU. Trondheim, 2009.
Fig. 1. Location of Stiklestad and the survey area at Haug. (Maps : Brynja Birgisdottir)
(a)
(b)
Plate 1. Oblique aerial photographs of the cropmark at Haug taken June 2007. (a) view from the west (b) view from the south (Photos : Eskil Folstad)
Fig. 2. Plan of longhouse with postholes excavated at Stiklestad 2008. (Birgisdottir [5])
Plate 2. Excavated posthole from a roof bearing post in the longhouse at Stiklestad. Ruler is 60cm
(Birgisdottir [5])
(a)
(b) Fig. 3, (a) Earth resistance image of the cropmark at Haug;
data range 302 Ohms (black) to 660 Ohms (white). (b) Location of GPR reconnaissance transects 3, 5, 7 & 9.
(a) (b)
(c) (d)
Fig. 4. Selected time sections from GPR reconnaissance transects (a) 3, (b) 5, (c) 7 & (d) 9. South at the left of each section. Location of transects shown in Fig. 3b.
(a)
(b)
(c)
Fig. 5. Selected GPR horizontal time slices averaged over : (a) 2-5 ns; showing cultivation traces, (b) 31-34 ns; showing the north and south ditches and possible internal stone cobbles & (c) 65-68 ns; Showing the basal “stones” in the ditches and possible base of chamber. The location of the 2m x 25m test
excavation is indicated on each slice.
Fig. 6. Drawing of the east face of the excavated trench
showing the ditches and features cut in the platform. (Birgisdottir [8])
Plate 3. The test excavation trench photographed from the south showing the eastern face of the trench. The ditches are visible on both sides of the lighter coloured central platform. The northern ditch is at the top of the picture. A hearth is visible in the central platform between the north and south ditches. (Photo : Brynja Birgisdottir)
Plate 4. A boat shaped mound at Sømna in Northern-Norway. The mound is clearly marked by a wide foot trench. (Photo : NTNU-Vitenskapsmuseet)