Geophysics Locate Royal Halls at Borre

I had a meeting with my geophysicist buddy Immo Trinks of the National heritage Board the other day, and he showed me an amazing Ground-Penetrating Radar (GPR) survey from Borre in Vestfold, Norway.

Borre is Norway's equivalent of Old Uppsala, with a large cemetery with huge barrows. One was obliterated by road workers in 1852, yielding a fairly well-preserved Viking Period ship burial of the Oseberg / Gokstad / Tune type, which sadly does not survive. Some copper-alloy metalwork from the grave gave the Borre Style its name, defined by knotwork with nicked ridges and Mickey Mouse heads.

The royal manor site that must accompany the graves at Borre has not been identified despite many decades' attempts. Not until Immo came along, that is. Terje Gansum is head of the visitor centre at Borre, and he had a hunch that some barely visible earthworks on the edge of the mound cemetery might be a ploughed-out building terrace. Bjørn Myhre had cut a long narrow search trench across the spot in 1991, identifying a few post holes and radiocarbon-dating one to about AD 700. But his trench hadn't been wide enough for him to be able to identify what kind of structure the post holes belonged to. The GPR solved that problem: the ground-plans of two large large hall buildings appeared! The animated plan loops a number of horizontal slices through the ground at successively greater depth: even the stones in the bottoms of the post holes are visible. The star-shaped black things in the eastern area are the roots of trees that spread wider in deeper slices.

Immo is also directing a magnetometry survey at my site in Kaga parish, Östergötland. GPR would probably give me more useful (that is, more easily interpreted) data, but unfortunately that method is much slower and thus more expensive than motorised magnetometry. Basically, to identify a house plan like those at Borre with magnometry, we would have to find a house whose postholes were full of burnt material such as pottery or daub. But magnetometry is good for hearths and rubbish pits, and sometimes waste scatters that show up in magnetometry can preserve the outline of a building. Thus, such a survey can help me place my excavation trench (and use my fieldwork time & money) more fruitfully.

Update 16 January: Added Immo on 10 January:

Actually, the sample frequency of a georadar system is much higher than that of a magnetometer.

Georadar is more expensive than magnetometry for the following reasons:

• Georadar data is recorded by measuring many parallel, closely spaced vertical radar sections, resulting in a 3D data volume and several hundreds of samples in depth per measured surface point.
• Magnetometer data is one sample per measured surface point.

Therefore processing, analysis and the archaeological interpretation of georadar data are considerably more time consuming than magnetic data analysis.

While it is common to run several magnetometers in parallel at the same time (e.g. four probes with 50 cm horizontal spacing), only few multichannel georadar systems exist today for similar system configurations. In recent years three interesting, approximately 2 m wide antenna arrays have been developed by some of the main georadar system manufacturers. These systems are pulled or pushed by motorized vehicles while data positioning is implemented using highly accurate Real Time Kinematic GPS (+/-2 cm positioning accuracy) or robotic total stations. Such multi-channel systems will certainly in the not too far future become the tool of choice for large-scale archaeological prospection, despite their rather hefty price tag. We plan to test some of these systems on archaeological sites with known structures in the coming year.

[More blog entries about archaeology, norway, vikings, vikingperiod, geophysics, gpr; arkeologi, vikingatiden, vikingar, norge, geofysik, georadar.]