I am attempting to create surfaces of strata via Kriging in Arc GIS from borehole data. In addition to the predicted surface I also want surfaces showing 95% confidence above and below the predicted surface. The issue I'm having is with these 95% surfaces.
The surfaces are very "tight" to the known surface when within a couple of meters of a known value/ borehole but standard errors immediately become high once a few meters away from the borehole.
I have included a cross-section of our 3 surfaces. The predicted surface is represented by the middle line with the lower dashed and upper line the lower and upper 95% confidence limits. Please note how steeply we lose confidence at a very close distance to a known borehole.