Hi all,
This is a follow up to
my previous post on summarizing a sequence with a KDE, and I can now see how to constrain the upper and lower limits of a KDEs.
Next, I'd like to summarize multiple sequences and phases from 10 sites. Some sites have long intervals without dates (as in the example below). So far, my attempts to combine these run into the same problem as before – at the edges of these long intervals, the KDE is smeared (seems to be the nature of the beast). This artificially shortens the long intervals without dates (for example, between the end of Sequence 1 the start of Sequence 2). There are also a few dates that are not part of sequences or phases.
What I'd like to do is summarize all dates across all sites, but still respect the sequences' discrete boundaries. Can I do this with a KDE? Or perhaps another approach? So far, my best results are 1) a Sum of all dates' posteriors (no smearing but tons of calibration noise, not pretty) and 2) a histogram of modeled medians (simplistic and discards ±, but visually clear).
Any suggestions are welcome! Thanks, –Erik
Plot()
{
Sequence("Stratigraphic sequence")
{
Boundary("Start",BC(2000));
Sequence("Sequence 1")
{
R_Date("date 1",3500,30);
R_Date("date 2",3400,30);
R_Date("date 3",3300,30);
R_Date("date 4",3200,30);
KDE_Plot("KDE Sequence 1");
};
Boundary("End Sequence 1");
Interval("Gap between sequences");
Boundary("Start Sequence 2");
Sequence("Sequence 2")
{
R_Date("date 5",1250,30);
R_Date("date 6",1200,30);
R_Date("date 7",1150,30);
R_Date("date 8",1100,30);
KDE_Plot("KDE Sequence 2");
};
Boundary("End",AD(1000));
};
};