Hello!
I am analyzing sampling points and trying to determine the distance at which they are no longer correlated to ensure that the data obtained is meaningful. The issue is that the sampling grid I used as a baseline has approximately 20 points, and the values recorded at each of these points are very similar.
I considered using the semivariogram to analyze spatial autocorrelation, but the results I obtained seem inconsistent, as they indicate distances at which sampling would not be feasible. Additionally, the cross-validation results reflect the same inconsistencies as the semivariogram. Out of all the datasets analyzed by date, only a few showed an acceptable pattern.
This leads me to my question: Is it possible for kriging or even the semivariogram to produce inconsistent results? I understand that this could be due to the limited number of points in the sampling grid and the fact that the data is nearly homogeneous, but could this be considered a valid justification?
Also, is there any other method I could use to determine the distance at which the sampling points are no longer correlated?
I would greatly appreciate your guidance.
