Select to view content in your preferred language

Help with kriging

3234
4
09-24-2012 02:21 PM
SamLevy
Deactivated User
I'm not sure that this is an appropriate use of this forum, but I have some data that I am trying to interpolate using the Geostatistical Wizard.  I have looked at a number of tutorials and forums that explain how to use it, but the data that they generally use seems to be a lot cleaner than what I am working with.  So when I try to do similar things with my data, I am not sure if the drastic differences in the final product are because their data was just cleaner or because there is something wrong with my data as far as interpolation is concerned.  I was hoping that if I supplied one of the areas that I was attempting to interpolate that someone could show me what some of the choices that they would make are and why they made those choices.  Things like whether or not to remove a trend or to transform the data, which type of kriging to use, whether to turn anisotropy on, what the lag size and number of lags should be.  I understand that I will probably be directed to a number of forums or tutorials, but I assure you I have read many of them already.  The reason I am asking for so much information is not because I haven't read these things but I guess because I fundamentally don't understand them well enough to be able to apply the concepts from the tutorial data to my data.  If somebody would be kind enough to help me through this first group of data, I think I should be able to figure out the rest of the areas I am working with for myself.

Almost all of my 122 data points are spaced at 50x50 meter intervals in a grid.  The data is of nitrogen levels in small area in Arizona in parts per million.

Any help is of course appreciated.
0 Kudos
4 Replies
EricKrause
Esri Regular Contributor
I took a look at your data, but I wasn't able to fit a good model.  There are several issues I saw that make it difficult to fit a model.  First, it looks like the variable to interpolate (NReclass) is a categorical classification.  There are many repeat values: more than half the data has NReclass = 10, and another quarter of the data has NReclass = 15.  With so many repeated values, the distribution is extremely right-skewed, and none of our transformations can handle such extreme skewness and repeated values.

Also, Moran's I with a Fixed Distance Band gives a p-value of just over .05, which means that the spatial autocorrelation is fairly weak.

If you have the original data (before the reclassification), it may give you better results.
0 Kudos
SamLevy
Deactivated User
Ok thanks a lot.  At least now I know that it isn't just me.  Unfortunately that data is essentially all that I have.  I had ranges instead of exact values.  Knowing that this is essentially all that I have, is there anything that I could do with the data along the lines of an interpolated surface or am I just kind of out of luck?  Again, thank you for taking a look at it.
0 Kudos
EricKrause
Esri Regular Contributor
If you really need a surface, you might try IDW or Natural Neighbors.  I definitely would not try to do kriging.  However, know that no matter which interpolator you use, the results will probably have large errors.
0 Kudos
SamLevy
Deactivated User
Thanks for the help.
0 Kudos