Age-depth modelling workshop by dffhrtcv3

VIEWS: 4 PAGES: 9

									Basic age-modelling

   Find ages for dated and undated depths
   E.g., linear interpolation, regression, spline (gaps)
   Choose which one looks nicest...
   How treat point estimates? (mid/max, multimodal)‫‏‬
Bayesian age-modelling

    Bayesian = combine data with other info
      14C   dates and depth info
        stratigraphical ordering / position
               •   e.g., wiggle-match dating
        Constraints on e.g. likely accumulation rates
        Other dates, e.g. pollen events, 210Pb‫)…(‏‬
        Outlier analysis
    Usually done by millions of simulations
Wiggle-match dating
Outlier analysis


•   Reasons: site, error, lab?
•   Give prior outlier probabilities to dates
•   Iteration i: is date within 2 lengths sd?
•   If not, label date and shift to fit
•   [Labelled / total]  posterior outlier prob.
•   No need to remove outliers!
•   Fit F: 1 – mean(posterior outlier prob.)‫‏‬
OxCal


•   Extract OxCal directory to C:\Program Files
•   Open .../OxCal/Index.html in Firefox
•   R_Date(‫“‏‬test”,‫;)05‏,0542‏‬
•   Save file, run
•   Run examples from manual
Bpeat


•   Extract Bpeat.zip somewhere
•   Open R there (or change dir)
•   source(“Bpeat.R”)
•   SetCore(“MSB2K”,2)
•   TestRun()
•   FinalRun( 0.1 ) # just a short run...
•   DepthChron()
The future of Bpeat: Bacon
Bacon

   Muscles and fat – robust, yet flexible
   Floppy/crusty – flexibility can be adapted
   Can be cut to your liking – hiatuses
   Cured – Bpeat bugs repaired
   MacBacon – multi-platform
   Pigs are smart – combine prior info + new data
   Pigs can fly – workshop in Mexico
Age-modelling‫‏…‏‬your‫‏‬own‫‏‬data?

    Try the different software pieces
    What are best settings for your site?
    Do you agree with the age estimates?
    Differences between approaches

								
To top