Sampling density and spatial analysis: a methodological pXRF study of the geochemistry of a Viking-Age house in Ribe, Denmark

Abstract

This study explores the significance of spatial sampling resolution on portable X-ray fluorescence (pXRF) analysis of an archaeological settlement site with favorable preservation conditions and clearly defined stratigraphic contexts as a benchmark study to interpret geochemical mapping of anthropogenic elemental markers. We present geochemical elemental mapping of a Viking-Age house floor in Denmark based on an unprecedented sampling density of a 0.25-m grid. In order to establish a fast, cost-efficient, comparable approach of how different sizes of data resolution affect the spatial elemental patterns, the data is analysed using three different grid sizes: 0.25 m × 0.25 m, 0.5 m × 0.5 m, and 1.0 m × 1.0 m. We analysed each grid size with selected anthropogenic markers (CaO, Cu, P2O5, and Sr) using ordinary kriging. The CaO, P2O5, and Sr patterns display a strong inter-correlation between data points up to a distance of 1–1.5 m from one another. At the highest resolution (0.25-m grid), all of the elements display a high degree of detail in the variation of the elements across the indoor surface with low standard deviations. Hence, the precise position of hot and coldspots, and spread of bounded concentration zones, is easily recognized in the maps. With the low resolution (1.0-m grid), the borders between high and low concentrations become more blurred and the indications of smaller hotspots (possible activity areas) are completely lost. Especially, Cu displays a high degree of clustering, which the high-resolution sampling could best reveal. This benchmark study shows that it is realistic to perform large-scale geochemical surveys of archaeological settlements using pXRF spectrometry in a standard archaeological excavation context, but also that sampling distances of 0.5 m × 0.5 m or finer are best suited to in indoor contexts.