spatial averaging of interpolation data
Wouter, you could resort to simulation, i.e. generate conditional simulations, compute the catchment aggregated value and obtain their distribution experimentally. If you're after the catchment mean, you better use block kriging for irregular blocks (the catchment). R package gstat provides this. If you're after another quantity, e.g. the catchment fraction above a given threshold you need to simulate, and e.g. use overlay() to select points in the catchment, select them, calculate statistics, etc. Hth, -- Edzer
Wouter Buytaert wrote:
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