Calculating k for P_sequence

[Kim Newman]

I have been reading Bronk Ramsey 2008 'Deposition models for chronological records' as I am wanting to understand how to calculate an age depth model.  I do not have a background in math or statistics, so I am slower than most on the uptake of the formulas use for various calibrations and the like and benefit from seeing real life examples.

I am currently stuck on how to calculate the k parameter used for the P_sequence.  The example in the caption for Fig3 is based on two identified layers that have their thickness measured at different locations.

u= 45, 50, and 48

v= 55, 63, and 72

The given formula is k = μu μv/[σu σv (pi +qi)]

where:
μu is the mean of u = 47.7
μv is the mean of v = 63.3
σu is the SD of u = 2.5
σv is the SD of v = 8.5

But I have no idea what p and q are in this equation

The example in the for A.20 is k(est) =  μu μv/σu σv  1/(u+v)

In this version of the formula I do not know which u values and which v values to use.

 I would really appreciate it if someone could show me a copy of the formula with numbers in it so I can clearly see where each of the values come from.

 

I am also curious how this formula works with multiple layers and what you do if the sediments clearly change in type throughout a depositional sequence?  For example, intermittent rock falls or a midden deposit.  Do you need to do a different P_sequence each time the coarseness of the sediment changes significantly or is there a way to account for this in one model?

 

Would really appreciate any advice

Cheers

Kim

[Christopher Ramsey]

Kim

From the equation in A20 the (p_I + q_i) is the total distance between the two events in a core.  The attached workbook shows how to calculate k from the values in Figure 3.

Note that most often now we allow the model to find the value of k from the data using the variable k method.  See variable k values in the manual under:

https://c14.arch.ox.ac.uk/oxcalhelp/hlp_analysis_oper.html#deposit

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/96a2a968-37da-41f3-833a-b4d61a1d6225n%40googlegroups.com.

[Kim Newman]

Thanks for that Christopher, the variable k is exactly what I was looking for.  This has been a great help

Cheers

Kim