How can the origin of lead from multiple sources be determined? A new statistical approach.
Determining the origin of terrestrial materials is a major scientific issue that concerns a wide range of disciplines. It is an issue for geochemists seeking to understand the origin of rocks throughout the Earth's history; it is central to the research of archaeologists trying to determine the movements of populations; and it is one of the preferred approaches of researchers attempting to reconstruct past climatic conditions or the impact of human societies on their environment.
Publication date: 30/04/2018
Press, Research
Related teams :
External Envelopes Geochemistry
Related themes : Origins
Many chemical or isotopic tracers are used for these purposes, but the study of variations in the concentration and isotopic composition of lead is one of the most effective tools for a wide range of applications. Pb has four isotopes, only one of which (204Pb) is stable, while the other three (206Pb, 207Pb and 208Pb) result from the natural radioactive decay of 232Th, 235U and 238U with very different half-lives. The signature based on the combination of the three isotopic ratios (206Pb/204Pb, 207Pb/204Pb and 208Pb/204Pb) of a source is therefore unique.
If the system under study is relatively simple and only two or three potential sources are suspected, the proportions and respective roles of these sources can easily be constrained by combining the three Pb isotope ratios.
Things become more complicated when the number of potential sources increases, as the mixing system then becomes under-determined and poorly constrained. To deal with such cases, sophisticated statistical tools for processing large amounts of data exist and are used by the community of biologists working on the multiple food sources of predators, for example. These Bayesian-type mixture models have never really been used in the fields of archaeology, sedimentology or palaeoclimatology.
In the article cited below, a group of British, Romanian and French researchers explore the possibilities offered by a Bayesian treatment of a mixture model by taking three examples published in the literature: the sources of dust present in pre-anthropic environmental samples, ore mining during the Roman period, and finally the source of lead in a Roman mining artefact. The new approach will enable us to better understand and quantify variations in Pb sources in relation to climatic conditions during pre-anthropic periods, when the natural Pb cycle was dominant. It also enables us to better characterise the origin of anthropogenic pollution in the Roman period.
Finally, its application to the archaeometric study highlights the contribution of Pb from Romanian mines, a hypothesis previously mentioned but never proven.
This promising approach can be applied to a wide range of scientific problems and provides a new method for robustly determining the sources of Pb observed in rocks or materials of various origins and ages.
