r.calib: return a random calendar age from a calibrated distribution

Description

Calculate the cumulative calibrated distribution, then sample n random uniform values between 0 and 1 and find the corresponding calendar ages through interpolation. Calendar ages with higher calibrated probabilities will be proportionally more likely to be sampled.

Usage

r.calib(
  n,
  y,
  er,
  cc = 1,
  postbomb = FALSE,
  deltaR = 0,
  deltaSTD = 0,
  as.F = FALSE,
  is.F = FALSE,
  thiscurve = NULL,
  yrsteps = FALSE,
  cc.resample = FALSE,
  dist.res = 200,
  threshold = 0,
  normal = TRUE,
  t.a = 3,
  t.b = 4,
  normalise = TRUE,
  BCAD = FALSE,
  rule = 2,
  cc.dir = NULL
)

Value

n randomly sampled calendar ages

Arguments

n: The number of calendar ages to sample
y: Uncalibrated radiocarbon age
er: Lab error of the radiocarbon age
cc: Calibration curve to use. Defaults to IntCal20 (cc=1).
postbomb: Whether or not to use a postbomb curve. Required for negative radiocarbon ages.
deltaR: Age offset (e.g. for marine samples).
deltaSTD: Uncertainty of the age offset (1 standard deviation).
as.F: Whether or not to calculate ages in the F14C realm. Defaults to as.F=FALSE, which uses the C14 realm.
is.F: Use this if the provided date is in the F14C realm.
thiscurve: As an alternative to providing cc and/or postbomb, the data of a specific curve can be provided (3 columns: cal BP, C14 age, error).
yrsteps: Steps to use for interpolation. Defaults to the cal BP steps in the calibration curve
cc.resample: The IntCal20 curves have different densities (every year between 0 and 5 kcal BP, then every 5 yr up to 15 kcal BP, then every 10 yr up to 25 kcal BP, and then every 20 yr up to 55 kcal BP). If calibrated ages span these density ranges, their drawn heights can differ, as can their total areas (which should ideally all sum to the same size). To account for this, resample to a constant time-span, using, e.g., cc.resample=5 for 5-yr timespans.
dist.res: As an alternative to yrsteps, provide the amount of 'bins' in the distribution
threshold: Report only values above a threshold. Defaults to threshold=0.
normal: Use the normal distribution to calibrate dates (default TRUE). The alternative is to use the t model (Christen and Perez 2016).
t.a: Value a of the t distribution (defaults to 3).
t.b: Value b of the t distribution (defaults to 4).
normalise: Sum the entire calibrated distribution to 1. Defaults to normalise=TRUE.
BCAD: Which calendar scale to use. Defaults to cal BP, BCAD=FALSE.
rule: Which extrapolation rule to use. Defaults to rule=1 which returns NAs.
cc.dir: Directory of the calibration curves. Defaults to where the package's files are stored (system.file), but can be set to, e.g., cc.dir="curves".

Author

Maarten Blaauw

Examples

Run this code

  r.calib(10,130,20) # 10 random cal BP ages
  plot(density(r.calib(1e6, 2450, 20)))

Run the code above in your browser using DataLab