The order of dates and duration of a phase

[סיגל לביא - אלבז‎‎]

Hi everyone,

I'm trying to find out whether in the simple model of sequence > boundary > phase

the order of the dates matter?

I tried to take only two phases from a site, once in stratigraphic order and once in chronological order and the results are almost identical.

And also - what is the best query for estimating the duration of the phase?

Thank you very much.

 

This is my code:

 

Plot()

 {

  Sequence("Site")

  {

   Boundary();

   Phase("2")

   {

    R_Date("RTD 10951",12335,40);

    R_Date("RTD 11029",12369,37);

    R_Date("RTD 10942",12231,40);

    R_Date("RTD 11148",12241,37);

    R_Date("RTD 10943",12332,40);

    Span("phase 2");

    Date("Phase 2");

    Sum(phase 2);

    KDE_Plot("Kernel phase 2")

    {

    };

   };

   Boundary();

   Phase("1")

   {

    R_Date("RTD 11145",12205,36);

    R_Date("RTT 6115",11570,75);

    R_Date("RTD 11147",12313,37);

    R_Date("RTD 10946",12727,41);

    R_Date("RTD 10945",12530,41);

    R_Date("RTT 6117-2",12300,70);

    R_Date("RTD 8021",12138,30);

    Span("phase 1");

    Date("Phase 1");

    Sum(phase 1);

    KDE_Plot("Kernel phase 1")

    {

    };

   };

   Boundary();

  };

 };

[Helene Agerskov Rose]

Hi סיגל לביא - אלבז,‎

To answer some of your questions: within a sequence events follow the order you put them in, but the order of dates within a phase does not matter. You can estimate the duration using the Span() command.

Your model agreement is very low and I am unsure what prior information you are basing the model on? The samples are very old and from what type of context? The group can help you better if you provide some more information.

Best,

Helene

[סיגל לביא - אלבז‎‎]

Hi Helene, thanks a lot.

Thanks for the answer regarding the order of the dates within the phases.
The dates are from an Epipaleolithic site with 8 layers, I only included two of them here, but my model includes them all.
If I tag some of them as outliers, it increases the agreement. Is this what is recommended to do?

thanks, Sigal

‫בתאריך יום ג׳, 17 בספט׳ 2024 ב-9:34 מאת ‪Helene Agerskov Rose‬‏ <‪heleneager...@gmail.com‬‏>:‬

--
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/36e848a8-3244-4542-b3c4-5a089565d6fcn%40googlegroups.com.

[Helene Agerskov Rose]

Dear Sigal,

The model agreement can always be increased by omitting problematic dates, but ideally you will have some other information to base this on rather than "it doesn't fit". Such as suspicious quality measurements (%C, d13C etc.), an uncertain find context (might the material have a long intrisic age or de redeposited?), or might samples have been swopped? Or stabilized at the museum? All that to say, there is no easy fix to your problem, but hopefully one emerges when you look at the details.

Best,

Helene

[Erik Marsh]

Hi Sigal,

Helene is right about outliers. Tagging all the dates as outliers does improve the agreement, but you'll have to look into each date.

You're right, Span queries the duration of each phase (Date is something different). You could even group phases 1 and 2 as a single phase (or make them overlapping phases, not contiguous phases) – this will model better but this has to be based on the priors from stratigraphy, etc. I also wonder if your phases are inverted? The deepest/oldest phase should be first – in this code, phase 1 is first, and models well, but there are still three major outliers in phase 2. I've grouped the phases under the command KDE_Plot, which treats them as a Phase and makes a KDE (the KDE is much better than Sum here). The command Interval queries the time between phases.

Maybe this code helps?

Erik

 Plot()
 {
  Outlier_Model("General",T(5),U(0,4),"t");
  Sequence("Site")
  {
    Boundary("Start");
    KDE_Plot("Phase 1") 
   {
    R_Date("RTD 10951",12335,40) Outlier(0.05);
    R_Date("RTD 11029",12369,37) Outlier(0.05);
    R_Date("RTD 10942",12231,40) Outlier(0.05);
    R_Date("RTD 11148",12241,37) Outlier(0.05);
    R_Date("RTD 10943",12332,40) Outlier(0.05);
    Span("Span phase 1");
   };
   Interval("Gap between phases");
   KDE_Plot("Phase 2") 
   {
    R_Date("RTD 11145",12205,36) Outlier(0.05);
    R_Date("RTT 6115",11570,75) Outlier(0.05);
    R_Date("RTD 11147",12313,37) Outlier(0.05);
    R_Date("RTD 10946",12727,41) Outlier(0.05);
    R_Date("RTD 10945",12530,41) Outlier(0.05);
    R_Date("RTT 6117-2",12300,70) Outlier(0.05);
    R_Date("RTD 8021",12138,30) Outlier(0.05);
    Span("Span phase 2");
   };
    Boundary("End");
  };
 };

[סיגל לביא - אלבז‎‎]

Thank you very much, Helene and Erik.
Yes, it does seem that Phase 1 has earlier dates than Phase 2, but the site's stratigraphy is quite good, and Phase 2 is actually earlier. It might be that some of the samples moved a little across the sequence on the site.
Many thanks for the help and the code.

‫בתאריך יום ד׳, 18 בספט׳ 2024 ב-17:42 מאת ‪Erik Marsh‬‏ <‪erik....@gmail.com‬‏>:‬

To view this discussion on the web visit https://groups.google.com/d/msgid/oxcal/d8c9d5a8-a9e5-43f7-bd7b-98d59556bbf4n%40googlegroups.com.