Kevin
There is something badly wrong with your software.
With ordinary kriging, it is not possible to get
negative variances.
With lognormal kriging, it is not possible to get
negative varianse or negative estimates unless you are
using a large additive constant.
Isobel Clark
Hello all,
Yesterday morning I asked for pointers to software and
algorithms, preferably MATLAB, that could simulate spatially-
autocorrelated and/or fractal 2D patterns. Here is a summary
of replies.
1) Brian R Gray [[EMAIL PROTECTED]] suggested simply using PROC
SIM2D in SAS.
2) Yetta Jager
