[R-sig-Geo] spatial averaging of interpolation data

Sam Field sf2257 at columbia.edu
Mon Oct 2 18:39:39 CEST 2006

I was going to say bootstrapping, but I guess that wouldn't work
since you can't have duplicate locations with kriging (see previous
post), as you would if you sampled from your sample.


Quoting Wouter Buytaert <wouter at paramo.be>:

> Hi list,
> this is a rather theoretical question, so sorry if it is not very
> appropriate for this list, but with a bit of luck someone may
> feel
> challenged to reply.
> I am looking for the correct procedure to average spatially
> interpolated
> data over an area. Specifically, I used kriging to interpolate
> point
> rainfall over a grid covering a catchment and now I want to
> estimate the
> average rainfall over the catchment, including the uncertainty on
> the
> estimation.
> Obviously, taking just the mean and the variance of all the grid
> cell values doesn't work because it does not take into account
> spatial
> correlation and thus underestimates the uncertainty.
> So what is/are the best method(s)? Suggestions on how to
> implement them in
> R are very welcome, as well as suggestions for literature if the
> answer
> would turn out to be more complicate than I imagined.
> Many thanks!
> Wouter Buytaert
> Lancaster University
> _______________________________________________
> R-sig-Geo mailing list
> R-sig-Geo at stat.math.ethz.ch
> https://stat.ethz.ch/mailman/listinfo/r-sig-geo

Samuel H. Field, Ph.D.
Associate Research Scholar
Institute for Social and Economic Research and Policy (ISERP)
Columbia University
420 W. 118th Street
Mail code 3355
New York, NY  10027

More information about the R-sig-Geo mailing list