Question about Age function

[Chris Harpel]

Hi Everyone,

I am working on a model to date an eruption that incorporates an age from a different radiometric dating method and have run into a problem that I am trying to understand.

The model incorporates 14C ages from multiple sites that are cross-referenced, but for simplicity, I have trimmed that all out to illustrate the problem.

For reference, TBJ is a dated tephrostratigraphic marker bed and Eruption is the age from another radiometric dating method.

Using the following model:

 Plot()
 {
  Curve("IntCal20","intcal20.14c");
  Sequence("Site 6")
  {
   Boundary("Start Site");
   Age("Eruption",N(2300,300))
   {
    Year=CE(2025);
   };
   Date("Paleosol");
   R_Date("Age 1",1620,20);
   Date("TBJ",U(CE(429),CE(433)));
   Boundary("End Site");
  };
 };

The model yields an age for Eruption that can be as young as 460 CE, which is younger than the ages of the overlying units, including the TBJ.

I have found, however, that when I put a 14C age beneath Eruption in the following model, it changes:

Plot()
 {
  Curve("IntCal20","intcal20.14c");
  Sequence("Site 6")
  {
   Boundary("Start Site");
   R_Date("Age 2",2000,25);
   Age("Eruption",N(2300,300))
   {
    Year=CE(2025);
   };
   Date("Paleosol");
   R_Date("Age 1",1620,20);
   Date("TBJ",U(CE(429),CE(433)));
   Boundary("End Site");
  };
 };

The model then yields a minimum age of 310 CE for Eruption, which is consistent with the overlying ages but is a bit of an artificial fix that I would prefer to avoid.

In the bigger model, there are plenty of 14C ages beneath Eruption, but they are cross-referenced and Eruption still comes out with a minimum age that is younger than overlying units.

So, I am a bit stumped on why this is happening and would like to understand a bit better.

What am I missing here?  Does anyone have any ideas?

Cheers,
Chris

[Erik Marsh]

Hi Chris –

So I found a solution but I don't understand why it works. This problem is really only apparent when I zoom to a decadal scale (see attached image). The TBJ range is shorter than the default 10-year resolution, so resolution seems to be the issue. When I set the resolution to 1 year, everything looks good (see below). I changed the command from Age to C_Date. This makes the eruption respect the following events with no anomalous tail that runs past later events in the sequence.

What I don't understand is that two other ways of expressing the same date did not work:

Age("Eruption",N(2300,300))   {  Year=CE(2025);  };

Date(“Eruption Date”,N(2025-2300,300));

For both of these, the curve was angular (see image), suggesting that they were stuck at the 10-year resolution and not respecting the resolution option, which I thought was global. Could this be a bug? Fixed by setting the time scale? Maybe someone else can answer that.

Erik

Options()  {  Resolution=1;  };
Plot()
 {

  Sequence("Site 6")
  {
   Boundary("Start Site");

   Date("Paleosol");

   C_Date(“Eruption Date”,2025-2300,300);

   R_Date("Age 1",1620,20);

   Date("TBJ",U(CE(429),CE(433)));
   Boundary("End Site");
  };
 };

[Christopher Ramsey]

Dear Chris & Erik

This is a bin size issue. Because the N(2300,300) has such a large uncertainty the bin size for that distribution (prior and marginal posterior) is set to be large. On the other hand the implication of Date("TBJ",U(CE(429),CE(433))) is that you know the range very precisely and possibly even within the year. The Boundaries etc will be set with the default 5 year bin size unless specified in the options.

This can be resolved with:

Options()
{
Resolution=0.5;
};

Plot()
{
Curve("IntCal20","intcal20.14c");
Sequence("Site 6")
{
Boundary("Start Site");

Age("Eruption",N(2300,300,0.5))
{

Year=CE(2025);
};
Date("Paleosol");
R_Date("Age 1",1620,20);
Date("TBJ",U(CE(429),CE(433)));
Boundary("End Site");
};
};

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 visit https://groups.google.com/d/msgid/oxcal/fbe4becc-66cd-4adf-a19c-b01c870f0e7fn%40googlegroups.com.

[Chris Harpel]

Hi Christopher and Erik,

Many thanks for both of your replies.

Christopher, your solution solved the problem in both the small model snippet that I sent and the larger model that incorporates it.  I very much appreciate your responsiveness!

Cheers,

Chris