I have estimated local cdfs for a continuous variable using indicator
kriging, and would like to go beyond the local variance of the random
variable to estimate the standard error of the E-type estimate at each
location. Is there a way to calculate this using GSLib programs?
Simulation/postsim, perhaps? It doesn't jump out at me from looking at the
parameter files. Is there a way to simulate the local mean the way you can
calculate the local mean using kt3d?
Alternatively, is there an analytical expression for this knowing the local
ccdf? I think the local variance, and the number of datapoints in the local
neighborhood as well as their weights should be part of this calculation.
Thanks very much!