Different calibrated data from OxCal, Calib and Rbacon

[Yiming Wang]

Dear all,

I have a rudimentary question and hope to get some support on it.

I have some 14C dates from a marine sediment core in the tropical Indian Ocean. I am using Marine20 for the calibration curve and accounting for the local reservoir age of 203 years with an uncertainty of 32 years.

I ran the data through OxCal, Calib, and Rbacon, but I got different calibrated median and sigma ranges (1 and 2). For example:

One of my 14C dates is 4160 years with an error of 35 years. When I ran the data through different calibration tools, I obtained varying results:

• OxCal: Bayesian modelled median age = 4040 years BP
• Calib: Modelled median age = 3772 years BP
• Rbacon Bayesian model: Modelled median age = 3868 years BP

Could someone give me a hint as to where I might have gone wrong?

Thank you,

Yiming

[Christopher Ramsey]

Dear Yiming

Thank you for your query.  Using marine curves in OxCal the code should either look like:

 Options()
 {
  Curve="marine20.14c";
 };
 Plot()
 {
  Delta_R(203,32);
  R_Date(4160,35);
 };


if you are using the quick calibration function or:

 Plot()
 {
  // Delta_R values checked for Marine20
  Curve("Marine20","marine20.14c");
  Delta_R(203,32);
  R_Date(4160,35);
 };


If you are using it within a model etc.  The median age should be ~3770 calBP as in the first attached plot.  Normally OxCal does run an MCMC process when there is a Delta_R (because this error is often shared between measurements) which means there is some noise.  An alternative approach is to subtract the Delta_R from the date and calibrated directly.  The errors should be combined in quadruture - so the corrected date is 3957 and the error sqrt(32^2 + 35^2) = 47.  The calibration is then just:

 Plot()
 {
  // Delta_R values checked for Marine20
  Curve("Marine20","marine20.14c");
  R_Date(3957,47);
 };

for which no MCMC is done.  This gives the second plot which is the same +/- a year or so from the MCMC noise - and possibly also the rounding of the error to the nearest year.

I cannot help with the comparisons - but the value you are getting for OxCal does not seem to be quite right.

Best wishes

Christopher

> --
> You received this message because you are subscribed to the Google Groups "OxCal" group.
> To unsubscribe from this group and stop receiving emails from it, send an email to oxcal+un...@googlegroups.com.
> To view this discussion on the web visit https://groups.google.com/d/msgid/oxcal/efad88e4-eb77-4a5e-b64f-0a7c4b38c220n%40googlegroups.com.

[Erik Marsh]

Hi Yiming,

I'm guessing you have checked that all three programs are using the same calibration curve and reservoir correction. My best guess is that your models are using different parameters – if you are running depth sequences in different programs, you can get quite different modeled medians. I would take a close look at each parameter, especially the assumed (or defined) sedimentation rate. And OxCal's k variable. Bacon tends to expect smoother translations between core sections (unless you define unconformities). It uses the deposition rates from both previous sections in your model as well as generalized trends from North American lake sediment cores in Goring et al. 2012.

Hope this helps,

Erik 

[Yiming Wang]

Hi Erik,

Thank you for the reply. It is very helpful. The reservoir corrections for each program were consistent. It is likely the parameters were difficult to find in each different calibration tool. For OxCal, I only tried the quick simple calibration without selecting any model yet. When I set different k variables in OxCal, it gives me similar calibration results.

What might cause the age difference when I calibrate the age one at a time and also enter them as a data list in OxCal? 

Thank you again,

Yiming

You received this message because you are subscribed to a topic in the Google Groups "OxCal" group.
To unsubscribe from this topic, visit https://groups.google.com/d/topic/oxcal/9uHPmWZ4Tbo/unsubscribe.
To unsubscribe from this group and all its topics, send an email to oxcal+un...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/oxcal/3b256777-dcbe-4e42-936d-30773e679eaen%40googlegroups.com.

[Yiming Wang]

Dear Christopher,

Thank you for the quick response to my query. You are right that the median age for this date is around ~3770 Cal BP in OxCal when I calibrate one date at a time. But when I enter the entire dataset using the import function, then the calibrated age for this date becomes ~4040cal BP. I assume other dates also become older too. Please see the the code and table output below. I wonder if I have missed something here, so the calibration did not take into account the Delta_R values?

Options()
 {
  Curve="marine20.14c";
  SD1=TRUE;
  SD2=TRUE;
 };
 Plot()
 {
  R_Date("Sample1",545,25);
  R_Date("Sample2",4160,35);
  R_Date("Sample3",6435,40);
  R_Date("Sample4",7575,40);
  R_Date("Sample5",8534,45);
  R_Date("Sample6",10470,50);
  R_Date("Sample7",12830,70);

  // Delta_R values checked for Marine20
  Curve("Marine20","marine20.14c");
  Delta_R(203,32);

 };

Thank you again for your support!

Yiming

You received this message because you are subscribed to a topic in the Google Groups "OxCal" group.
To unsubscribe from this topic, visit https://groups.google.com/d/topic/oxcal/9uHPmWZ4Tbo/unsubscribe.
To unsubscribe from this group and all its topics, send an email to oxcal+un...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/oxcal/458F30DF-8F48-4C29-99EE-999D7FFE2DF5%40arch.ox.ac.uk.

[Christopher Ramsey]

Dear Yiming

The Curve and Delta_R commands only affect statements after them in the code. This is so you can switch between curves and Delta_R values. The following should do what you wish:

Plot()
{
// Delta_R values checked for Marine20
Curve("Marine20","marine20.14c");
Delta_R(203,32);

R_Date("Sample1",545,25);
R_Date("Sample2",4160,35);
R_Date("Sample3",6435,40);
R_Date("Sample4",7575,40);
R_Date("Sample5",8534,45);
R_Date("Sample6",10470,50);
R_Date("Sample7",12830,70);

};

Best wishes

Christopher

> On 17 May 2024, at 09:46, Yiming Wang <ywalas...@gmail.com> wrote:
>
> Dear Christopher,
>
> Thank you for the quick response to my query. You are right that the median age for this date is around ~3770 Cal BP in OxCal when I calibrate one date at a time. But when I enter the entire dataset using the import function, then the calibrated age for this date becomes ~4040cal BP. I assume other dates also become older too. Please see the the code and table output below. I wonder if I have missed something here, so the calibration did not take into account the Delta_R values?
>
> Options()
> {
> Curve="marine20.14c";
> SD1=TRUE;
> SD2=TRUE;
> };
> Plot()
> {
> R_Date("Sample1",545,25);
> R_Date("Sample2",4160,35);
> R_Date("Sample3",6435,40);
> R_Date("Sample4",7575,40);
> R_Date("Sample5",8534,45);
> R_Date("Sample6",10470,50);
> R_Date("Sample7",12830,70);
> // Delta_R values checked for Marine20
> Curve("Marine20","marine20.14c");
> Delta_R(203,32);
> };
>

> To view this discussion on the web visit https://groups.google.com/d/msgid/oxcal/CAAS8STdx%2B3Kvad7bvkyTmyRdifjEDGTH251q6cKBwQNcoDFPdA%40mail.gmail.com.

[Yiming Wang]

Dear Christopher,

Yes, that's it! Thank you.

All the best,

Yiming

To view this discussion on the web visit https://groups.google.com/d/msgid/oxcal/B44C7CFB-38A7-42C8-9AD5-076D61F517E9%40arch.ox.ac.uk.

[Yiming Wang]

Dear Christopher,

Sorry to bother you again about this question. I had another close look at the results and had a different set of observations.

Plot()
 {
  // Delta_R values checked for Marine20
  Curve("Marine20","marine20.14c");
  Delta_R(203,32);
  R_Date("Sample1",545,25);
  R_Date("Sample2",4160,35);
  R_Date("Sample3",6435,40);
  R_Date("Sample4",7575,40);
  R_Date("Sample5",8534,45);
  R_Date("Sample6",10470,50);
  R_Date("Sample7",12830,70);
 };

And  then I added the Delta_R(203, 32) in front of the every 14C

Plot()

 {
  // Delta_R values checked for Marine20
  Curve("Marine20","marine20.14c");
  Delta_R(203,32);
  R_Date("Sample1",545,25);

  Delta_R(203,32);
  R_Date("Sample2",4160,35);
  Delta_R(203,32);
  R_Date("Sample3",6435,40);
  Delta_R(203,32);
  R_Date("Sample4",7575,40);
  Delta_R(203,32);
  R_Date("Sample5",8534,45);
  Delta_R(203,32);
  R_Date("Sample6",10470,50);
  Delta_R(203,32);
  R_Date("Sample7",12830,70);
 };

The results from both calibration methods differ by approximately 100 years. The calibrated age from the second method, which involves repeating the Delta_R correction for each sample, aligns exactly with the results from Calib. These results are slightly younger. Based on these findings, I'm curious about the necessity of repeating the Delta_R correction for each date. Could this approach influence the discrepancy? I'd appreciate your perspective on this.

Thank you again and all the best

Yiming

To view this discussion on the web visit https://groups.google.com/d/msgid/oxcal/B44C7CFB-38A7-42C8-9AD5-076D61F517E9%40arch.ox.ac.uk.

[Christopher Ramsey]

The two approaches are theoretically different. The first assumes that the Delta_R is the same for all samples but with a prior of 203+/-32; the second assumes that each Delta_R might be different.

Normally this does not make that much difference. However, here your first date, if it is on the Marine20 curve and thus older than 1950 must have a lower Delta_R value than 203 by quite some way. You will see there is a poor agreement for the Delta_R - and indeed the first date. This one point then pulls the Delta_R down for all the other dates - hence the effect.

If you miss off the first sample there is very little difference between the two approaches. You do expect some difference as, even for the Marine Curve, some radiocarbon values are more likely than others because of plateaux etc.

Best wishes

Christopher

> To view this discussion on the web visit https://groups.google.com/d/msgid/oxcal/CAAS8STcvi%3DvYkK4VeN68kbsRdfnNtm8o8N4D53f9TWUPkqqvWA%40mail.gmail.com.

[Yiming Wang]

Dear Christopher,

Thank you so much for the explanation. Yes, I see how the first date affects the calibration now. Perhaps I can leave the first date out for Delta_R correction ( 203+/-32  ) or find a more suitable Delta_R correction for the first date.

Best wishes,

Yiming

To view this discussion on the web visit https://groups.google.com/d/msgid/oxcal/BEF427E5-2F1C-4D43-971B-63D53DD3EEE5%40arch.ox.ac.uk.