Kriging - Variable search and variogram rotation

Rupert (Moderator (AUS)) 5 years ago in Resource Estimation updated by Metin Yurdakul 4 years ago 5


I noticed Micromine added the ability to vary the rotation of the search ellipse in the 2016 release. I was speaking to a Micromine employee recently who told me that this function also rotates the variogram axes but I can find no documentation of this in the help pages.

Could you let me know if it does or doesn't rotate the variogram model please?

Thanks and all the best


Hi Rupert,

Variable search ellipsoid rotation has been a feature of Micromine's IDW interpolation for quite some time now -- if memory serves, it was added around 2012 or thereabouts.

In Micromine 2016.1 we also added variable search orientations to the various kriging interpolations. The orientations in the block model affect the orientations of both the search ellipsoid and the variogram model. Another way to think of it is that the variogram model axes are always aligned with the search ellipsoid axes.

Clearly, this should have been documented in the help. It is an oversight and I will ensure it is corrected.



The Help page now mentions that the variogram models are rotated but it is unclear exactly how.

The variogram axes are defined as Main Direction, 2nd Direction and 3rd Direction.

Could you confirm that:

Main Direction is the Y axis of the search ellipse

2nd Direction is the X axis and

3rd Direction is the Z axis?

In other words, if the Azimuth, plunge and rotation are all 0:

Main Direction: Az=000, plunge=0

2nd Direction: Az= 090, plunge= 0

3rd Direction: Az=180, plunge=90



Hi Rupert,

Yes, that's correct, the variogram axes mirror the search ellipsoid axes.

Another way to look at is is to consider a series of left-handed rotations in Z - X - Y order, where Z, X, and Y are the axes being rotated around.


A left-handed rotation around the Z-axis is the same as setting the azimuth, followed by

A left-handed rotation around the X-axis, which is the same as setting the plunge, followed by

A left-handed rotation around the Y-axis, which doesn't have a direct geological equivalent (it does help to determine the pitch of Axis 1 in the plane defined by Axes 1 and 2).

Hi Frank

Are you saying the rotation of variogram axes and search axes are hardwire linked or are you saying the rotation conventions for the two are the same?  Not clear in your post.




Hi Gary,

Yes, when Variable Search is enabled, the variogram orientations and search ellipsoid orientations both adopt the azimuth/plunge/rotation numbers assigned to each block in the block model.

And, yes, the rotation conventions are the same for both.