topic Re: Ordinary Kriging and Universal Kriging in ArcGIS GeoStatistical Analyst Questions

Ordinary Kriging and Universal Kriging

KirstenFalzarano — Mon, 06 Aug 2012 17:46:48 GMT

Hi there,
Can anyone tell me what the difference is between ordinary kriging with a first order trend removal and Universal Kriging (also with first order trend)? Is there an advantage to using one over the other?

Re: Ordinary Kriging and Universal Kriging

EricKrause — Mon, 06 Aug 2012 20:21:35 GMT

If you do your trend removal well, you shouldn't see a big difference between Universal and Ordinary/Simple kriging. As a general rule, we suggest using Simple kriging because it supports Normal Score Transformations.

The difference between Ordinary/Simple and Universal is subtle and a bit confusing. The trend removal page is essentially just changing your input data by removing the trend that you fit. Then the semivariogram is fit to the detrended data. The difference between Ordinary/Simple and Universal is that Universal goes back and re-fits a global trend model to the data that has already been detrended (it has to do this for the Universal kriging equations to work). If you did your trend removal well, there should be very little global trend left to re-estimate, and Ordinary/Simple and Universal kriging should give very similar results.

Re: Ordinary Kriging and Universal Kriging

KirstenFalzarano — Tue, 07 Aug 2012 18:07:38 GMT

Thanks for the response!
If I am interested in comparing whether or not to use a trend removal (based on the cross-validation results), is it best to compare the results of an Ordinary/Simple Kriging interpolation (no trend removal) with the results of a Universal Kriging interpolation (with first order trend removal)? If the results of a first order trend removal with ordinary/simple kriging is very similar to Universal Kriging is this test redundant (in terms of a comparison), or should I compare the two variations of Ordinary/Simple Kriging instead of Universal Kriging?

Re: Ordinary Kriging and Universal Kriging

EricKrause — Tue, 07 Aug 2012 18:20:06 GMT

You can compare against Universal kriging if you want, but personally, I would just use Simple kriging. Build two Simple kriging models, one with trend removal and the other without. See if removing the trend gives better cross-validation.

Re: Ordinary Kriging and Universal Kriging

KirstenFalzarano — Tue, 07 Aug 2012 18:30:33 GMT

Okay, last question (I hope!!)

When I run simple kriging (with normal score transformation) without and with a trend removal and compare the results of ordinary kriging (with and without trend removal), ordinary kriging gives me better cross-validation results. Is this a justifiable reason to use ordinary kriging?

Re: Ordinary Kriging and Universal Kriging

EricKrause — Tue, 07 Aug 2012 19:12:06 GMT

If it gives better cross-validation, then that's a good justification for using the model. Ultimately, you need to justify that the model fits your data the best, and cross-validation is a good way to do that.

Re: Ordinary Kriging and Universal Kriging

KirstenFalzarano — Tue, 07 Aug 2012 20:35:06 GMT

Thank you!!!