Hi all
As part of a larger study I?m going to test the difference in CI for
estimated extreme flood quantiles (Q with a return level of 100 years)
for different samples of the same population to illustrate the effect of
sample size. I wonder if anyone have some advice on how I can do this
by boostrapping in S-plus?
The AMS-series is 112 years. I have estimated Q100 based on sample
sizes 10 years, 30 years and 100 years. My question is especially related to
how I should boostrap. Is it best to select one 10 years sample (or more)
and bootstrap from this (parametric?)and compare them with the same
procedure for the 30 and 100 year sample? Any suggestions on how to do
this in S-plus?
I use S-plus 6.0 for Win. I?m applying the Gumel distribution and the
L-moment method for parameter estimation, both of which I have
programmed in S-plus.
Thank you in advance
Line J. Barkved
|