Advancing obsidian provenance analysis through an integrated research infrastructure 

and support system

黒曜石産地推定のための共通基盤の構築と包括的支援体制の整備に向けて

Yoshimitsu Suda (Nagasaki University)

1. Introduction

Obsidian artefacts have been widely recovered from Palaeolithic and Jomon archaeological sites throughout the Japanese archipelago. Since the 1970s, the application of analytical techniques, such as X-ray fluorescence (XRF), has enabled the identification of geological sources. These studies have significantly advanced our understanding of prehistoric exchange systems, revealing the movement of raw materials from source areas to sites of consumption and providing insights into human mobility and social organisation.

Despite the recognised importance of provenance analysis, a vast number of obsidian artefacts—likely on the order of several million—remain unanalysed in museums and institutional collections across Japan. This backlog continues to grow, with thousands of new artefacts added each year. Currently, provenance studies are conducted by only a limited number of facilities, including some private laboratories, resulting in a national analytical capacity of only a few thousand samples per year. Consequently, the current system falls far short of meeting substantial demand. In addition, outsourcing analyses requires dedicated funding, which poses further constraints for many institutions.

To address these challenges, a distributed analytical framework, where cultural heritage institutions across Japan can independently conduct provenance analysis, offers a promising solution. Such an approach would greatly enhance analytical capacity and facilitate new archaeological insights into the past. However, its implementation requires the establishment of a central organisation capable of providing comprehensive support.

2. Fundamental principles of obsidian provenance analysis

Obsidian is formed through rhyolitic to andesitic magmatic processes, and its chemical composition varies systematically across geological sources. In particular, the concentrations of trace elements such as rubidium (Rb), strontium (Sr), and yttrium (Y) differ among sources, providing a geochemical basis for source discrimination.

By analysing the elemental composition of obsidian artefacts recovered from archaeological sites, typically using techniques such as X-ray fluorescence (XRF), it is possible to infer their geological origin. This approach has become the standard method for investigating the provenance of obsidian in archaeological research.

To support such analyses, we provide an online database that compiles both the geographic distribution (location data) and quantitative compositional data of obsidian sources, primarily from the Japanese Archipelago.

In our study, the term “obsidian provenance analysis” is used to refer specifically to the identification of the geological source of obsidian artefacts.

3. Research history of scientific approaches to obsidian provenance analysis in Japan

Provenance analysis of obsidian artefacts using physicochemical methods, such as X-ray fluorescence (XRF) and neutron activation analysis (NAA), was first developed in the 1960s by research groups at the University of Cambridge and the University of Michigan. In Japan, initial efforts began in the late 1960s when a research group at the University of Tokyo explored provenance determination based on chronometric techniques. However, this approach has not been practically implemented.

In the 1970s, provenance studies based on neutron activation analysis, utilising the reactor facilities at Rikkyo University, were conducted by research groups at Rikkyo University and Tokyo Gakugei University. From the 1980s onwards, XRF gradually replaced NAA as the primary analytical method, with active research conducted by groups at Kyoto University, Numazu College of Technology and Tokyo Gakugei University.

Entering the 2000s, these XRF-based approaches—particularly those developed at Numazu College of Technology—were further adopted and applied at institutions such as the Meiji University cultural heritage research facilities and the Nagasaki Prefectural Archaeological Center, leading to broader implementation in archaeological practices.

However, since the 1980s, XRF-based provenance studies in Japan have largely relied on relative X-ray intensities or theoretically derived concentrations based on the fundamental parameter (FP) method. Because these values vary depending on the analytical instrument used, they cannot be directly compared across different laboratories, a limitation commonly referred to as instrument dependence. Consequently, provenance methodologies have tended to develop independently within individual institutions, limiting data integration and broader comparability.

Against this background, since the 2010s, research groups at Nagasaki University and the Center for Obsidian and Lithic Studies, Meiji University, have been working to establish a new analytical framework based on absolute elemental concentrations, enabling instrument-independent and comparable XRF-based provenance analyses.

4. The need for concentration-based provenance analysis and the role of obsidian reference standards

Outside Japan, provenance analysis based on quantitative elemental concentrations obtained by X-ray fluorescence (XRF) had already been established by the 1980s. In contrast, within Japan, methods based on relative XRF intensities or theoretically derived concentrations using fundamental parameter (FP) approaches are still widely employed.

For provenance analysis to be implemented in a distributed manner across cultural heritage institutions nationwide, a shared analytical framework is essential, in which results generated at different facilities can be directly compared. Achieving comparability requires a transition to concentration-based approaches using absolute elemental values.

In the case of geological obsidian samples, high-precision quantitative analysis can be achieved by powdering the material and preparing fused glass beads for wavelength-dispersive X-ray fluorescence (WDXRF) analysis, typically using commercially available rock standards. However, archaeological obsidian artefacts must be analysed non-destructively. This creates a critical requirement for solid reference materials with well-characterised compositions to enable reliable quantitative analysis without damaging the specimens. In this study, “reference standards” refers to materials for which elemental concentrations have been accurately determined through high-precision analysis.

To address this need, we established, for the first time in Japan, a suite of 39 obsidian reference standards (Fig. 1), enabling the non-destructive quantitative analysis of obsidian artefacts using energy-dispersive X-ray fluorescence (EDXRF) with a calibration-curve approach. These standards primarily consist of geological obsidian collected from Japan. Their certified values were determined using WDXRF at the Centre for Obsidian and Lithic Studies, Meiji University, based on the low-dilution-glass bead method.

In parallel, we are collaborating with a domestic XRF instrument manufacturer, Rigaku Corporation, to develop and implement an application for quantitative obsidian analysis on EDXRF instruments (NEX DE) using these standards. Detailed information on the reference standards, including their certified compositional values, is publicly available in our online database (Nagasaki University, n.d.).

5. Provenance determination using reference obsidian specimens

In obsidian provenance studies, results are often expected to be reported in terms of named geological sources. However, the outcomes of quantitative compositional analysis are, in practice, expressed as “chemical composition groups.” This distinction arises because a single geological source does not necessarily correspond to a single homogeneous compositional group, an important point that is not always fully recognised by provenance data users.

In Japan, even primary geological sources—where obsidian is originally formed—can include materials belonging to multiple compositional groups. Furthermore, secondary deposits, such as those found in riverbeds or coastal environments, frequently contain mixtures of obsidian derived from different primary sources. As a result, obsidian with similar or identical chemical compositions may be found across multiple locations. Therefore, “source names” and “chemical composition group names” cannot be treated as equivalent.

Based on this understanding, we organise obsidian not by geographic source labels but by chemical composition groups. These groups correspond to the fundamental units of provenance discrimination in this study. For each group, we defined a set of 12 reference specimens, and their geographic information and compositional data were made publicly available through our online database.

By utilising either physical reference specimens (through a loan system) or their associated compositional datasets, institutions equipped with XRF instruments can perform provenance analysis without the need to assemble their own geological reference collections. Currently, the preparation of these reference specimens has been prioritised for western Japan, with ongoing efforts to extend coverage to Hokkaido and eastern Japan.

6. Development of analytical algorithms and user interfaces for provenance determination

In many cultural heritage institutions across Japan, staff members do not necessarily have specialised training in statistical analysis or analytical chemistry. Moreover, developing data-processing algorithms for provenance determination is a technically demanding task. However such methods are available, complex or unintuitive interfaces can hinder their practical adoption. To overcome these barriers, we aim to develop robust analytical algorithms for provenance determination and user-friendly interfaces that enable seamless implementation. The goal was to provide a system that can be readily used by non-specialists in routine workflows.

In practice, users will be able to upload CSV files generated by XRF instruments into the interface and simply issue instructions such as “display the provenance results.” The system then generates a summary of the results for each sample. In addition, commands such as “plot the results on discrimination diagrams” will automatically produce visualisations showing the classification outcomes.

Importantly, the interface was designed to function in an interactive and conversational manner. Even when user instructions are imprecise, the system employs artificial intelligence (AI) to interpret the intent and execute the appropriate analytical procedures. We plan to implement this environment as a web-based platform accessible through standard browsers, such as Google Chrome.

7. Shared use of reference standards and geological reference specimens

Our program aims to contribute to archaeology and cultural heritage research through the preservation and effective use of lithic raw materials, with a particular focus on obsidian. As part of this effort, a suite of 39 obsidian reference standards, together with geological reference specimens covering the Japanese archipelago, is made available for shared use.

In accordance with the internal regulations of the Faculty of Education, Nagasaki University, these materials can be loaned on a fee basis for academic research. This system enables institutions to access well-characterised reference materials without the need to establish their own collections.

In addition, analytical services using the instrumentation installed in the Earth Science Laboratory, Faculty of Education, Nagasaki University are available on a fee basis. Further details on both the loan system and analytical services are available in the online database.