Dendrochronological Provenance Patterns. Network Analysis of Tree-Ring Material Reveals Spatial and Economic Relations of Roman Timber in the Continental North-Western Provinces

Over the last decades, Roman wood in various shapes and sizes has been excavated in the region of the continental north-western provinces of the Roman empire. However, it is often unclear whether wood has been transported. Most dendrochronological provenance studies rely on simple comparisons between tree-ring series based on a single similarity measure. In addition, most fail to consider the complex system of relations that is the result of the variables that influence tree-ring patterns. Network analysis is a solution to this problem, because it allows to both visualize and analyze the complex (provenance) relations of tree-ring series as a whole. A network makes it impossible to ignore existing (statistical) relations between tree-ring series. Although networks can be build using any (combination of) similarity measure(s), in this study a combination of the Synchronous Growth Changes (SGC), its related probability of exceedance (p), correlation (r) and overlap define the edges. This paper focusses on networks with site chronologies as nodes, although networks can also be constructed using individual tree-ring series or a combination with site chronologies. A combination of these can also help to refine the (archaeological) interpretation. The location of any tree-ring series in a network reflects its provenance. Material that is placed closer together in the network has similar growth patterns and is generally from the same region. Therefore, network communities reflect wood that has a similar provenance. If tree-ring material was found on different spatial locations, but in close proximity in the network, this indicates that wood has been moved. To determine which wood has been transported, a combination of archaeological and spatial arguments is used. The method is sound, simple and gives insight in the complexity of all tree-ring relations in a simple diagram. The resulting patterns show that most wood was obtained in the region where it was used, but that transport of wood in the Roman period did take place. Three scales are defined to describe the Roman wood economy: local, provincial and imperial. While transport of wood over long distances is attested for both military and civilian sites (provincial and imperial scale), it seems that wood that was transported beyond the provincial borders was only used in civilian sites (imperial scale). The combination of network science, dendrochronology and archaeology is a powerful method to understand patterns in the Roman timber economy.