Author: Jarrad Kowlessar
Kowlessar, Jarrad, 2023 A New Focus on Ancient Art: Digital Techniques to Record and Understand Rock Art, Flinders University, College of Humanities, Arts and Social Sciences
Terms of Use: This electronic version is (or will be) made publicly available by Flinders University in accordance with its open access policy for student theses. Copyright in this thesis remains with the author. You may use this material for uses permitted under the Copyright Act 1968. If you are the owner of any included third party copyright material and/or you believe that any material has been made available without permission of the copyright owner please contact copyright@flinders.edu.au with the details.
Rock art as a material cultural artefact, presents significant and unique challenges to all aspects of archaeological analysis and cultural heritage management, including site management issues, preservation methods, and disaster planning against the ever-increasing threats of climate change and industrial destruction.
Although much work has been done towards studying and managing rock art from various viewpoints, a holistic approach encompassing the many requirements, perspectives and outcomes is lacking. This thesis addresses that need by presenting new digital approaches that embrace emerging modern technological applications to offer fresh opportunities that overcome many of the previous barriers to engagement. Applying and demonstrating new methodologies from raw data collection, through cutting edge digital, geospatial and geophysical landscape analysis, this work provides an advanced, but cost-effective new means of gathering, analysing and communicating this information using various platforms.
Firstly, ground-based photography is used to create high-resolution, georectified three-dimensional models of rock art sites, while aerial remote sensing is used to generate models of the landscape around a target site. This reveals a complex Pleistocene landscape, which offers the potential to locate additional archaeological sites and so reveal more about the lifeways of the earliest Australians.
Additionally, this research conducts a new approach to stylistic motif analysis, the first ever machine learning analysis of Australian rock art, using data efficient transfer learning to identify distinguishing features within styles of rock art. By generating a stylistic chronology, it is shown that the model is sensitive to both temporal and spatial patterns in the distribution of rock art in the Arnhem Land Plateau region.
Further, building from thesis site recordings, motif analysis and palaeogeographic reconstructions, a Geographic Information Systems based spatial analysis is produced, applying detailed models to the spatial analysis of rock art site placement. The resultant elevation, land cover and visibility modelling reveal significant changes in the site placement strategies of the rock art in the region, highlighting changes in four key phases of the past environmental history, spanning from the late Pleistocene to the late Holocene.
Finally, this thesis develops a new approach to archaeological landscape visualisation through the use of game engine technology designed for detailed virtual reality immersion. This combines the separate data and archaeological models produced throughout the thesis and merges these into one holistic, virtual cultural landscape simulation that can be dynamically explored and interacted with. The resulting new workflow used to create this simulation makes a unique way for archaeological information to be explored and communicated and is especially useful in communicating complex landscape information in an intuitive way.
This research contributes innovative digital methods to each component of rock art analysis in a structured way that allows data to build constructively towards a unified model of an archaeological rock art landscape. Each section of the thesis research has contributed to detailed and unique knowledge about the rock art of Arnhem Land and together forms a cohesive approach that can be applied to the global discipline of rock art management and analysis.
Keywords: Rock Art, Archaeology, Machine Learning, Geophysics, GIS, Remote Sensing, Virtual Reality, Unreal Engine
Subject: Archaeology thesis
Thesis type: Doctor of Philosophy
Completed: 2023
School: College of Humanities, Arts and Social Sciences
Supervisor: Ian Moffat