Reconstructing Deposits and Recontextualising Artefacts at Creag a' Chapuill Cave, Skye

Published: 9 October 2023| Version 1 | DOI: 10.17632/b44mhb82nr.1
Callum Murray


This dataset contains a 3D model of Creag a' Chapuill cave, Skye, Scotland. The 1932 excavation at Creag a' Chapuill cave by W. Lindsay Scott revealed significant archaeological evidence with the potential to inform archaeologists about settlement and trading contacts along the western seaboard of Scotland. 3D Computer Aided Design (CAD) software, Graphical Information Systems (GIS) and legacy data was used to virtually reconstruct excavated deposits and recontextualise artefacts. Volumetric analysis revealed a high density of worked-stone (225 artefacts per m3) in Region 4C and a high slag density in Regions 4B (17.2 kg/m3), 4C (5.6 kg/m3) and 2B (1.4 kg per m3). Exploratory Spatial Data Analysis (ESDA) was used to detect potential patterns of artefact distribution in order to investigate the degree and extent of disturbance at the time of excavation, chronological phases, and activity areas. Join count statistic indicated global spatial autocorrelation within the worked-stone (z-score = -2.144) and Bronze Age pottery (z-score = -2.109) data and Local Moran’s I statistic indicated local spatial autocorrelation within the worked-stone and slag data. The data can be used to confirm patterns recognised by the original excavator, as well as revealing new patterns which highlight the complex depositional and post-depositional formation of the cave deposits. This allows easier comparisons to be made with more recently excavated sites therefore increasing our understanding of the local prehistoric landscape and contributing to our knowledge of prehistoric cave use within Scotland.


Steps to reproduce

A survey of Creag a' Chapuill Cave was undertaken between 9th - 10th April 2021 resulting in the creation of 36 profile drawings and a plan drawing. A traditional method of surveying outlined was used. Horizontal and vertical baselines were set up within the cave and measurements were taken every 20 cm along the base of the cave and every 10 cm along the cave walls. The XYZ co-ordinate for each surveyed point was stored in a text file. This was imported into FreeCAD and the points were automatically connected with lines to create profiles. The profiles were lofted to form a solid shape and the corners of the cave were created by extruding the plan drawing. Boolean functions were used to combine them into a single 3D model. Next, the 1934 drawings were scaled, aligned, and digitised. Scaling was initially achieved by using the scale bars on the published drawings. However, alignment accuracy was improved by matching known points on the new drawings to points on the 1934 excavation drawings. Excavated deposits were reconstructed by extruding each area to the depth of each spit and using the 3D model as a template to shape the deposits. This process produced a 3D model which is accurate at the surveyed profiles, however the space between profiles approximates the true shape of the cave. The volume of each region was calculated within FreeCAD. The 3D model was exported using the VRML file format, displayed in Esri ArcScene, and joined with the volume and artefact data (digitised from the original excavation report).


University of Ulster


Archeology, 3D Analysis, Spatial Analysis, 3D Geographical Information System, Prehistoric Archeology, Cave and Rock Shelter