As palaeoecologist, I work on data retrieved from natural archives, going back in time. It can be a lake sediment core or peat. The time scale of the data, actually, is built by interpolation of a few radiocarbon dates, measured at particular places along the core. This creates an age-depth model. I routinely use two R packages from Maarten Blaauw to do this:
clam, for classical age-modelling,
rbacon, for Bayesian accumulation (and it's an R package, obviously).
Each package defines its own function (
Bacon(), respectively) which only needs the name of the core, to find a CSV file on your computer with the same name. This is in this CSV file that one saves the results of the radiocarbon datings to use them with either clam or rbacon. However, each package expects the data to be presented in a slightly different fashion. But in a file having the same name. Computers don't allow that. And sometimes, I want to be able to use either clam or rbacon.
There are some workarounds, but they imply either to constantly manipulate the data file(s), or to maintain several files, eventually in different folders. The is neither an enjoyable situation, nor safe. The risk is high of forgetting what you've done last time (so, is it a clam file or an rbacon file???) or, worse, to lose data. :(
Script is much safer, because it is a written protocol of what you're doing! This is the structure of my directories:
~/Project Name/Chronologies/Core Name
Of course, an RProject file lies in the
Project Name directory to open RStudio and automatically set the working directory. Then, I maintain only one CSV file in the
Core Name directory. The rest is just a bit of R-magic.
library(tidyverse) library(clam) core_name <- "Core Name" ( radioc_data <- str_c("Chronologies/", core_name, "/", core_name, " Dating results.csv") %>% read_csv() )
# A tibble: 3 x 7 lab_ID age_uncal sd c_content delta13c method depth <chr> <dbl> <dbl> <dbl> <dbl> <chr> <dbl> 1 Ftsy-41930 890 20 992 -36.4 graphite 20 2 Ftsy-59862 970 50 992 -29.3 graphite 40 3 Ftsy-30725 1020 30 234 -28.5 graphite 75
Starting from this, I can easily produce a CSV file to be used by the function
( clam_file <- radioc_data %>% mutate(cal_BP = "", reservoir = "") %>% select(ID = lab_ID, C14_age = age_uncal, cal_BP, error = sd, reservoir, depth) %>% arrange(depth) ) clam_file %>% write_csv(path = str_c("Chronologies/", core_name, "/", core_name, ".csv")) clam(core_name, coredir = "Chronologies")
For using rbacon, the script is:
library(tidyverse) library(rbacon) core_name <- "Core Name" ( radioc_data <- str_c("Chronologies/", core_name, "/", core_name, " Dating results.csv") %>% read_csv() ) ( rbacon_file <- radioc_data %>% select(ID = lab_ID, age = age_uncal, error = sd, depth) %>% mutate(cc = 1) %>% arrange(depth) ) rbacon_file %>% write_csv(path = str_c("Chronologies/", core_name, "/", core_name, ".csv")) Bacon(core_name, coredir = "Chronologies")
See that the
rbacon_file data frame that this script produces is a bit different than
This is how I reconciled working with both clam and rbacon packages from the same radiocarbon result file!