The purpose of a reservoir offset is to enable the application of calibration data ($mu(theta)$, emph{e.g.} shortciteNP{stuiver:98}) developed for one reservoir (primary reservoir) to CRA's from another (secondary reservoir). The usual approach has been to define the activity of the secondary reservoir as some form of constant offset (with error) from the primary reservoir (emph{e.g.} citeNP{stuiver93:_model_bc}). In this case CRA's from a secondary reservoir are not independent. However, the standard procedure for incorporating offset error into calibrated distributions assumes that the CRA's from secondary reservoirs are independent ({it e.g.} citeNP{stuiver93b}), accordingly the calibrated distributions are incorrect. In many cases this calculation error will be insignificant, however the calculation error will be significant in some situations and approaches such as sample based Bayesian inference need to be adopted if a non independent reservoir offset is applied. Corresponding author, Martin Jones, "Martin Jones" <martin@analytic.co.nz> |

__Keywords__

radiocarbon, Bayesian inference, reservoir correction

__Math Review Classification__

Primary 62F15 (primary),
; Secondary 62-01 (secondary)

__Last Updated__

__Length__

7

__Availability__

This article is available in: