Being new to geostatistics I am unsure about the various types and modes of spatial analysis. Is there a way to determine the probability of relatedness between points? Can the variogram be interpreted as a CDF? The geostatistical books, basic enough for me, do not really elaborate on probability. At this point I am unsure whether to search for classical or frequency based probability, to explain the relatedness of points.
Specific to my question, say: could I partition the variogram into three categories (1) range 0-10 very related, (2) 10 -20 related, and 20 – max distance not related. How would a person assign a specific probability to the very related category for points less than 10 units apart? Is it simple probability or does the joint distribution cause some issues?
I really want to determine if a treatment was applied to a given set of points, what would the probability be of near points responding in a similar manner. Surely with spatial continuity there is a way to extrapolate this value.
I am unsure how to proceed.
Thanks in advance.
OSU Environmental Sciences