DendroCT

DendroCT is a research project, supported by the Nordic Culture Fond, which aims at developing methods for non-destructive dendrochronological dating of archaeological objects by means of CT-scanning.

Partners in the project are:

Prof., Ph.D. Jan Bill, Museum of Cultural History, University of Oslo (project development, project direction)

Ph.D. Aoife Daly, Department of Environmental, Social and Spatial Change, Roskilde University (dendrochronology)

Ing. Knut Dalen, Animal and Aquacultural Sciences, Norwegian University of Life Science, Ås (CT)

MA Carina Sjöholm, Göteborg City Museum (archaeological objects)

The first test scan of the project, an anchor stock from the Oseberg find. The picture is put together from nine 50x50 mm sections, each with a resolution of 512x512 cells.

Graph depicting the X-ray absorption along a radius from the outermost year ring towards the heart of the Oseberg anchor stock. The variation reflects the density of the wood and thus the year ring pattern.

Project aim

Dendrochronology is the most precise dating method available in archaeology, with the capacity of dating wooden artefacts with an accuracy of less than a year. Since it is based on ocular measuring of the year rings, it requires, however, in most cases destructive sampling of the objects. This strongly limits its applicability. The development of three-dimensional imaging techniques has in recent years allowed for non-destructive dendrochronology, but the equipment used for this – high-resolution 3D computer tomograms – are in general not accessible for archaeological purposes. The present project aims at developing a methodology for the use of low-resolution CT-scanners – the type which is found in large numbers in the health care sector – for imaging of year rings and dendrochronological dating. If the project succeeds, it will allow not only dendrochronological investigation of highly sensitive wooden objects like sculpture or art work, but also more extensive dendrochronological analyses of complex objects like ships, where the dendrochronological examination of large numbers of structural components will allow for the study of, e.g., the object’s biography and mode of construction.

Challenges

The main challenge in developing a reasonable successful method for the use of low-resolution CT-scanners for dendrochronology lies in the fact that these scanners today are constructed to calculate the X-ray absorption within a 512x512 matrix, the minimum size of which is 50 mm x 50 mm. The maximum resolution is thus ca. 0.1 mm, which is considerably less than that of a traditional microscope used for dendrochronology, and too little to identify very narrow annual growth rings. Thus CT-scan procedures have to be refined towards obtaining the best possible results, but it is also necessary to develop methods to develop with an inpredictable loss of year rings in the CT-scans.

Photo montage of the Äskekärr ship on display in the Göteborg City Museum. Ten planks selected for CT-scan for the project has been marked out with red. These planks are all made of radially split oak, and they are expected to be around 1000 years old.

Investigating the Äskekärr ship

The main action during the project is the CT-scanning and analyses of the year ring pattern in ten planks from the Äskekärr ship from Göta Älv in Western Sweden. The ship has been selected, because it is characteristic for many wooden museum objects. It was excavated in 1933, and underwent various conservation processes, including controlled drying immediately after the excavation, treatment with pentachlorophenol in the 1960’s, and a complete surface treatment with carnauba wax in the 1980’s. Today the remains are on display in Göteborg City Museum, where they have been laid out in an arrangement imitating the find situation. Since the indivdual parts have not been mounted in a reconstruction, it is easy to remove entire parts for scanning. Many of the parts are, however, very fragile and should be handled as little as possible.

Radiocarbon dating of inlaid caulking from the ship has produced dates in the 8th-11th centuries, and dendrochronological examinations of smaller timbers from the ship have, without being conclusive, supported a Viking Age dating. Oak planks from the Viking Age and earlier usually have very narrow year rings, since their parent trees have grown in dense forests, while planks from later periods generally display much wider annual growth rings. Since earlier attempts to use CT for dendrochronology have demonstrated that narrow year rings are difficult to identify with CT, it is important for the project that the Äskekärr ship possesses exactly this quality.

The investigation of the ten selected planks is planned to take place in September 2009. The planks will be removed from the exhibition and placed on transport stretchers, supported with packing material. During transport and scanning they will remain on these stretchers, in order to reduce handling to a minimum. After scanning they will be returned to the exhibition.

For the scanning, the optimal sections will be identified on the individual planks, by visual inspection and by trial scans. In order to improve the quality of the annual growth ring curves it is planned to scan two sections on each plank. By combining year ring measurements from two sections on the same plank, it is possible to increase the number of year rings recorded for that particular plank, as well as to reduce the impact of local irregularities in the growth ring pattern.

After scanning, the width of the year rings in the scanned sections will be recorded. Experiments will be made in order to identify the best method. The year ring width can simply be measured directly on the scans, or image analysis software can be employed to identify the borders between the year rings. An important step will be to develop methods to deal with the possibility of unidentified year rings in the scans through systematic test of shorter year ring sequences against each other within the find and against external curves. This is a problem well known also from conventional dendrochronology, but which is accentuated by the limited resolution of the CT-technology.