External drift kriging basic question

classic Classic list List threaded Threaded
2 messages Options
Reply | Threaded
Open this post in threaded view
|

External drift kriging basic question

ccvert
Hello,

I am new to geostatistics. I was hoping someone could answer the following question:

When trying to do external drift kriging, I understand that you create the variogram from the residuals. Then, my question is when you actually do the kriging, do you interpolate the residuals and add them to an ordinary kriged field?

Or, can one just take the variogram derived from the residuals and apply it to the raw data when kriging.

Thank you for your time.

Chris
Reply | Threaded
Open this post in threaded view
|

Re: [52N Geostatistics] External drift kriging basic question

Paul Hiemstra
On 03/07/2011 03:36 PM, ccvert wrote:

> Hello,
>
> I am new to geostatistics. I was hoping someone could answer the following
> question:
>
> When trying to do external drift kriging, I understand that you create the
> variogram from the residuals. Then, my question is when you actually do the
> kriging, do you interpolate the residuals and add them to an ordinary kriged
> field?
>    

Hi,

When you perform external drift kriging, the variogram is indeed of the
residuals. When predicting at an unobserved location the prediction
comprises of a trend part and a residual part. The trend part is based
on the predictors and its estimated trend coefficients, only information
at the unobserved location is used for prediction at this stage.
Ofcourse, the other observations are used to estimate the trend
coefficients, so indirectly all the observations are used. The residuals
are interpolated using simple kriging with use of the residual
variogram. Adding those two components, predicted trend and residual, up
results in a prediction at an unobserved location.

For me, looking at the math behind kriging helped me to understand what
was going on exactly. Any textbook on interpolation can be a source for
this math, e.g. Chiles and Delfiner or Journel and Huijbregts.
Alternatively, in one of my papers[1] I also show the math for kriging
with external drift, although I call it universal kriging in that paper.

hope this helps,
Paul

[1] http://intamap.geo.uu.nl/~paul/files/PHiemstra_CandG_paper.pdf

>
> Or, can one just take the variogram derived from the residuals and apply it
> to the raw data when kriging.
>
> Thank you for your time.
>
> Chris
>
> --
> View this message in context: http://geostatistics.forum.52north.org/External-drift-kriging-basic-question-tp2646342p2646342.html
> Sent from the 52° North - Geostatistics Community Forum mailing list archive at Nabble.com.
> _______________________________________________
> Geostatistics mailing list
> [hidden email]
> http://list.52north.org/mailman/listinfo/geostatistics
> http://geostatistics.forum.52north.org


--
Paul Hiemstra, MSc
Department of Physical Geography
Faculty of Geosciences
University of Utrecht
Heidelberglaan 2
P.O. Box 80.115
3508 TC Utrecht
Phone:  +3130 253 5773
http://intamap.geo.uu.nl/~paul
http://nl.linkedin.com/pub/paul-hiemstra/20/30b/770

currently @ KNMI
paul.hiemstra_AT_knmi.nl

_______________________________________________
Geostatistics mailing list
[hidden email]
http://list.52north.org/mailman/listinfo/geostatistics
http://geostatistics.forum.52north.org