Introduction and Definitions
Age models and chronologies can be built in Tilia. In addition, Tilia supports the building of age models with several popular age modeling programs and can read the results from these programs. The following definitions are necessary to understand age models and chronologies in Tilia and the Neotoma Paleoecology Database. These definitions follow the complete workshop report from the PAGES Age Models, Chronologies, and Databases workshop held in 2014.
Age control – an estimate of absolute age, often with a specified uncertainty, for a sample within a core or stratigraphic profile that is used to constrain an age model for that core or profile.
Age model – an algorithm used to estimate the age-depth relationship for a series of stratified paleoenvironmental data points (e.g. depths within a core or stratigraphic profile), whose relative chronological relationships are known but for which only a limited amount of absolute chronological information is available from a set of age controls. Age models are used to make estimates of ages for depths not directly associated with an age control or to resolve discrepancies among age controls. Two kinds of age models are distinguished, classical and Bayesian. Some archaeological and paleontological sequences may have age models that do not explicitly define depths, only that sections are chronologically related to other sections. An age model incorporating sample depths may thus also be called an age-depth model.
Classical age model – an age model in which a curve or line is fitted to a series of age-depth points with no prior assumptions about sediment accumulation rate or monotonicity of ages (Blaauw 2010). If the age controls are derived from radiocarbon determinations, they may be calibrated or uncalibrated. Any calibration is undertaken before the curve is fitted, and outliers are rejected a priori or after producing the model. Common classical algorithms include linear interpolation, linear or polynomial regression, and various splines.
Simple age model – a classical age model that does not provide any estimate of the errors for interpolated ages.
Modern age model – an age model that provides uncertainty estimates for interpolated sample ages, regardless of whether the modeling is classical or Bayesian.
Bayesian age model – an age model that makes prior assumptions about sediment accumulation rates, stratigraphic superposition, and thus monotonicity of ages, and provides fully probabilistic estimates of the uncertainties in sample ages via application of Bayes’ theorem. Programs that implement Bayesian age models include Bacon (Blaauw and Christen 2011), OxCal (Bronk Ramsey 2008), and Bchron (Haslett and Parnell 2008). Age controls may be uncalibrated radiocarbon ages or calendar ages with uncertainties. These age models produce calibrated or calendar ages, and they can automatically deal with most cases of outlying dates.
Chronology – a series of estimated ages and associated uncertainty estimates for samples in a stratigraphic sequence. Such estimates usually derive from an age model and its associated age controls.
Age Models in Tilia