In richardjtelford/bio250: What the Package Does (One Line, Title Case)

knitr::opts_chunk$set(echo = TRUE, eval = FALSE)

Setup

Make a new project (File > New Project). This helps keep track of files and working directories.

Load the tidyverse package with library()

Calibrating individual dates

Fitting an age-depth model for the Trettetjørn data.

1) Download the dates_trettetjorn.csv from mitt uib.

2) Put the file dates_trettetjorn.csv into the directory “chronology_practical/dates_trettetjorn” inside the directory where your project file is. This is because Bacon and clam are very inflexible about where the data should be.

3) , then import the dates into R with read_csv2() and assign to the object trettetjorn with the arrow <-

4) Calibrate dates individually with IntCal::calibrate(). To calibrate the first date, you will need code like

library(IntCal)
calibrate(age = trettetjorn$"14C_age"[1], error = trettetjorn$error[1])

Fitting a model with clam

skip this section - the data need reformatting a lot for it to work (unfortunately Bacon and clam need the data formatting differently).

If you still want to do it, edit the data according to instructions in the help file ?clam::clam.

5) Fit a model with clam

#skip this section
clam("dates_trettetjorn", coredir = "chronology_practical/")

Try different types of model

Bchron

Bchron models are easier to fit than Bacon models because you can give it the data you have imported, whereas Bacon needs to find the data in a particular directory.

Load the Bchron package with library()

Fit the Bchron model with the function Bchronology()

Important arguments you need to set are ages, ageSds, and positions (i.e depth). See the examples in the help file for how to do this with ?Bchronology(). The calCurves argument will default to "intcal20", which is what we want, so it does not need changing.

After fitting the Bchron model, and assigning it to an object, you can plot it with plot() or make predictions for the depths you have proxy data at with predict()

Running Bacon test examples.

An alternative package for generating age-depth models is rbacon.

1) Load the rbacon package with library(rbacon). 2) Run the command Bacon() to make an age-depth model for the ^14^C dates on the default core MSB2K. Answer "y" to the questions. 3) This will take a a few minutes (depending on how fast your computer is). When it is finished, run agedepth() to create a plot.

The first plot indicates how well the Monte Carlo procedure has converged – you want this to resemble a hairy caterpillar (white noise).

The second and third plots show the prior (green) and posterior (grey) for the sedimentation rate and memory respectively.

The main panel shows the calibrated ^14^C dates (blue) and the age depth model, where darker greys indicate more likely ages. A red line picks out the most likely age for each depth.

4) Run proxy.ghost() to see the proxies with chronological uncertainty.

Using Bacon

dates_trettetjorn.csv is formatted correctly for rbacon

The file must be saved in the expected directory or Bacon will not find it (see above for instructions).

Fit a model with Bacon. Answer “y” at the prompt.

You need to tell Bacon the name of the file (without the .csv), which must also be the name of the directory containing the file.

The sep argument is needed because the columns are separated by semicolons rather than the expected commas.

Bacon("dates_trettetjorn", coredir = "chronology_practical/", sep = ";")

Examine the age-depth plot and check for convergence.

Bacon will put some output files into the data directory.

richardjtelford/bio250 documentation built on Oct. 12, 2022, 2:16 a.m.