Accueil > INTERNET > Emplois et stages > Offres d’emplois et de stages de formation > Post doc : Bayesian model for predicting the spatio-temporal trends in ferrous metal production and distribution

Post doc : Bayesian model for predicting the spatio-temporal trends in ferrous metal production and distribution

- Subject and scientific context


The Khmer Empire, based at Angkor in Cambodia, rapidly extended their political influence across mainland Southeast Asia between the 11th and 13th c. AD. Traditionally, Angkor’s power base is attributed to an elaborate road system and the lack of key resources around the capital suggests this network provided materials necessary to enable agricultural output, supply of armies, constructions, and trade with foreign nations. Iron is viewed as an ideal – and currently undocumented – medium to investigate these empire-building processes.


The post-doctoral project is proposed in the frame of the ANR project IRANGKOR that aims to intensively evaluate the raw material supply, distribution and trade of iron within the periods of expansion and ultimate collapse (15th c. AD) of the Khmer Empire. Metal trading modalities are a complex interaction between the archaeological, historical, technical, social and environmental contexts. Data sources that may be considered as key indicators of trends of the production and distribution organization are numerous and relate to characteristics of the iron product itself (e.g. composition, sourcing), the context of discovery and the temporal context. Reconstructing the exchange networks therefore requires a multidisciplinary approach integrating and combining this heterogeneous set of data (quantitative, qualitative, presence, absence data). In this context, the development of a predictive statistical framework that integrates the most diverse range of available evidence for modelling diachronic and synchronic production-exchange systems is of major importance.


The postdoc will be responsible for developing a model capable of reconstructing the exchange networks by combining the quantitative and qualitative data, within their archaeological and historical milieu, generated early in the frame of the project IRANGKOR. He/she will have to set up a statistical and comprehensive framework to address two core issues : (1) predict the probability for an iron object (used on a consumption site) to come from a production zone with changes in scale of production sites through time ; (2) determine the spatio-temporal patterns of production and distribution. To address this spatio-temporal problem, Bayesian statistical methods will be implemented. Based on Bayes rule, this approach will permit to establish a relationship between data and a set of parameters. Parameters that should be considered in the model are those representing geographical characteristics such as mineral sources, production loci, regional centers and discovery/consumption site. The objective will be to estimate, a posteriori, how these parameters are related through time, knowing the data. Each of these parameters will be described by production materials (i.e. geochemical, typological and technological data). A main part of the work will therefore be to correctly define prior knowledge about the different geographical parameters. Parameters of the model correspond to spatial coordinates/areas (geographic location) and time (14C in-slag dating, 14C dating of metal itself, temporal data coming from historical or stylistic sources). This project will aim to determine the posterior correlation between these parameters that will permit to better understand the diachronic and synchronic organization of the iron exchange system within medieval Southeast Asia. The development of such an innovative statistical model represents a key methodological advance for archaeometallurgical studies. This tool could be implemented in the frame of other international projects for which a reconstruction of iron exchange networks in the past is crucial.

- Duration :


18 months starting in September 2016.

- Salary :


The monthly salary is 2020 Euros (net).

- Work supervision


LANOS Philippe (DR CNRS), Université Bordeaux 3, Ecole Doctorale « Montaigne-Humanités ». UMR CNRS 5060 IRAMAT (Institut de Recherche sur les Archéomatériaux), CRPAA (Centre de Recherche en Physique Appliquée à l’Archéologie) and UMR CNRS 6118 Géosciences-Rennes, Université Rennes 1.

LEROY Stéphanie (CR CNRS), Laboratoire Archéomatériaux et Prévision de l’Altération
UMR CNRS 5060 IRAMAT (Institut de Recherche sur les Archéomatériaux), LMC (Laboratoire Métallurgies et Cultures) and UMR CNRS 3299 SIS2M/NIMBE, CEA Saclay.

- Location


This project will be carried out jointly by two laboratories of the IRAMAT (Institut de Recherche sur les Archéomatériaux) : the CRPAA (Centre de Recherche en Physique Appliquée à l’Archéologie) at Rennes and the LAPA (Laboratoire Archéomatériaux et Prévision de l’Altération) at Saclay. The post-doc will be based at the Université Rennes 1 (Laboratoire Géosciences-Rennes, groupe Modélisation chronologique et Archéomagnétisme). During the duration of the project, many work meetings between LAPA and CRPAA will be organized (e.g. 2 days every month either in Saclay or in Rennes).

- Candidate profile


We look for a candidate specialized in applied mathematics, more specifically in statistics with skills in programming and stochastic numerical calculus. Moreover, the candidate must be very interested by applications in anthropology and questions involving spatio-temporal problems.

- QUALITIES/SKILLS NEEDED

  • PhD Thesis in statistics and probability
  • Marked interest in Human sciences, in particular in archaeology and dating techniques in archaeometry
  • Autonomy, open mind, capacity to work in a team
  • Knowledge of programming (C++)

- Applications


Please send us a CV and a research statement to contacts :


LANOS Philippe : philippe.lanos@univ-rennes1.fr


LEROY Stéphanie : stephanie.leroy@cea.fr

Recommendation letters should be sent to us directly in electronic form.

- Application deadline : 31 May 2016