Integrating Archaeological Models and Data with Bayesian Data Assimilation

Summary

This is an abstract from the "The Expanding Bayesian Revolution in Archaeology" session, at the 88th annual meeting of the Society for American Archaeology.

Archaeological data are crucial for understanding how human societies shaped—and were shaped by—their biophysical environments. Yet these data are often sparse, noisy, and time averaged, making it difficult to uncover patterns of change across space and time. Process-based simulations are one way to fill the gaps in these imperfect proxy records, but they too require data for calibration and validation that are often lacking. Here I discuss how data assimilation, a Bayesian model-data integration approach used in engineering, meteorology, and other fields that work with imperfectly sampled complex systems, can be applied to the archaeological record. In data assimilation, an ensemble of simulation outputs serves as a Bayesian prior for the possible states of a past socioecological system and the uncertainty in its observation. I show how updating this simulated prior with real archaeological and paleoenvironmental data yields spatially explicit, multivariate reconstructions of past climate and demography. Any variable in the simulated prior can be reconstructed with this approach, as it leverages the internal consistency of the simulation to “spread out” information from the sparse proxy records. Crucially, this Bayesian approach to model-data integration enables a full accounting of uncertainty, helping target future research efforts that optimally reduce this uncertainty.

This Resource is Part of the Following Collections

• The Expanding Bayesian Revolution in Archaeology   •
• Society for American Archaeology 88th Annual Meeting, Portland, OR (2023)

Keywords

Record Identifiers

Abstract Id(s): 36554.0