User:Nicolas Perrault III/Radiocarbon calibration

For a variety of reasons, archaeological dates obtained with the method of radiocarbon must be calibrated to be correctly interpreted. This is important, for there can be differences of more than 4000 years between an uncalibrated date and its calibrated counterpart. For instance, one of the earliest putatively Iberomaurusian dates is 21,240 ± 130 14C BP (Tamar Hat, OxA-27506, Hogue and Barton 2016). This date is uncalibrated and calibrates to 25,845-25,270 cal BP (95.4% confidence) with the IntCal 13 calibration curve.

Unfortunately, dates reported in the literature are not systematically calibrated, nor systematically left uncalibrated. It is often difficult to figure out whether a given date in a book or scientific article is calibrated or not. When writing a Wikipedia article, it is nevertheless crucial to figure this out. Figuring this out allows calibrating dates for Wikipedia if this has not already been done by the cited author, and then specify that the date given in the Wikipedia article is calibrated, using "cal BP" instead of the ambiguous "BP".

How to verify that a radiocarbon date is calibrated
Verifying that a radiocarbon date is calibrated involves four steps.
 * 1) Make sure that the date is indeed a radiocarbon date. Dates cited as 50,000 BP or more can be assumed to be non-radiocarbon dates, because that is the limit of the radiocarbon dating range. To the contrary, the vast majority of dates younger than 50,000 BP will be found to be radiocarbon dates.
 * 2) Determine whether the radiocarbon date is presented as calibrated or uncalibrated (see next section).
 * 3) If it is found to be uncalibrated, calibrate it using, for instance, the OxCal Firefox extension. This will give a range of possible dates (e.g. 25,845-25,270 cal BP). You may decide to round it (e.g. "c. 25,600 cal BP").
 * 4) Specify inline that the date is now calibrated. Do this by writing "cal BP" instead of "BP".

Determining whether a cited date is calibrated or not
The following can help determine whether dates given in the literature are calibrated or not. Even with the key given below, it is sometimes not possible to determine whether this is the case. An editor that cannot resolve whether a cited date is calibrated or not may very well consider removing the date from the article.


 * 14C BP or uncal BP: the date is a radiocarbon date and is uncalibrated.
 * cal BP: if the date was obtained on radiocarbon, it has been calibrated. The "cal" in "cal BP" means "calendar years", which is more general than "calibrated years". Dates not obtained on radiocarbon and hence not needing calibration are calendar years and therefore "cal BP".
 * BP: Used alone, this is often ambiguous, but there are at least five ways of resolving the ambiguity:
 * 1) The source giving the date was published before 1998, when the IntCal98 calibration curve was published. In these cases, the given dates can be assumed uncalibrated.
 * 2) The text specifies that all dates therein are either calibrated or uncalibrated. In practice, unless one is wiling to read much of the text, this is hard to verify.
 * 3) The date is given with a margin of error. Almost always, uncalibrated dates use the plus-minus symbol (e.g. 21,240 ± 130 BP), whereas calibrated dates use the dash (e.g. 25,845-25,270 BP).
 * 4) The laboratory number is given (e.g. c. 21,000 BP (OxA-27506)). In these cases, it is often possible to google the laboratory number and determine whether the cited date is calibrated or not.
 * 5) If the date is referenced to have appeared in earlier publications, verify that reference and apply criteria 1 to 4 to that earlier publication.

Example
Consider the following: "In Italy, the Epigravettian industry is traditionally said to begin c. 20,000 BP."

Step 1: Determine whether the date actually is a radiocarbon date
Is the date younger than 50,000 BP? If it is, it is almost certainly a radiocarbon date. (Radiocarbon dating is both considerably cheaper and more precise than other methods.) If it is older than 50,000 BP, it is not a radiocarbon date, because the range of radiocarbon dating is 50,000 years. It is hence not ambiguous to write "75,000 BP".

Very occasionally, one will encounter a non-radiocarbon date that is younger than 50,000 BP. For example Vermeersch (2000) gives a thermoluminescence date of 25,000 ± 2500 BP for the beginning of the Shuwikhatian culture of Egypt. In this case, one could format the date as "25,000 ± 2500 TL BP" so that a Wikipedian does not assume it is a radiocarbon date and wonder whether it is calibrated or not. In the example above, the date is younger than 50,000 years, so it can reasonably be assumed to be a radiocarbon date.

Step 2: Determine whether the date is calibrated or not
In the example, the cited text (Palma di Cesnola 2001) is a French translation of a 1993 Italian book. It is therefore likely that dates therein are uncalibrated, since the original book was published before 1998. Nevertheless, it is possible that the dates were calibrated for the French translation, so the dates cannot be assumed uncalibrated. The text does not seem to discuss radiocarbon calibration, which is a hint, but not proof, that the date is not calibrated. The text gives neither a margin of error nor a laboratory number.

The text does refer for that date to two previous publications (Bartolomei et al. 1979, Bietti 1990). Because these two publications are earlier than 1998, it is probable that the date is uncalibrated. But it is best to verify, since the 2001 book might have secretly calibrated it. Verifying either of these two earlier publications confirms that the cited date of 20,000 BP is uncalibrated.

Step 3: Calibrate the date
Follow this step only if you have determined that the date is not calibrated. If it is calibrated, go to step 4.

This step assumes you use OxCal, which is available as a Firefox extension. You may also choose to calibrate dates in a different way. In OxCal, type "20,000" in the "date" field and use an artificial error (±) of 1 year. Choose a calibration curve, preferably the most recent IntCal curve, which should be the default. Click calibrate. You will see a bell curve and a note above it that reads:

''95.4% probability 24231 (95.4%) 23887calBP''

The calibrated date is 24,231-23,887 cal BP. Change in the text the uncalibrated date for the calibrated date. The calibrated date notation is cumbersome and you may choose to round the preceding calibrated date to "c. 24,100" or "c. 24,000" cal BP.

Step 4: Specify inline that the date has been calibrated
So that other Wikipedians don't have to repeat steps 1 and 2, specify that the date given in the text is now calibrated. To make your calibration more easily verifiable, you may add a hidden comment in the source code:

The sentence now looks like this: "In Italy, the Epigravettian industry is traditionally said to begin c. 24,000 cal BP."