> In 1971, I and 16 other people were taught in a short
> course at Fontainebleau given by Andre Journel and
> Charles Huijbregts, completely in terms of the
> semi-variogram with the covariance only being brought
> in as a special case when you could ensure
> stationarity of both mean and standard deviation.
> Where these conditions are met, the two are
> equivalent. 

We all know that when the covariance function exists, so does the
variogram, but the converse is not always true. Hence, for pedagogical
reasons, one can chose to teach geostatistics in terms of variogram, even
for simple kriging -- when it is not necessary.

> For simple kriging, the mean has to be
> stationary by definition 

this is correct

> but only the covariance of
> the increments (differences) has to be stationary to
> obtain a semi-variogram.

but this not correct. For simple kriging, it is necessary that
the covariance function exists: it is an 'if anf only if' condition (we
are in a second order stationarity context), hence the variogram exists.
The condition on the increments is not strong enough, because it does not
ensure that the covariance function does exist.

Now, to Isobel's first point. The trick that consists in transforming a
variogram gamma(h) in a so-called pseudo-covariance A-gamma(h) in OK and
UK is made for numerical reasons. It is well known that when inverting a
matrix it is much better (for numerical reasons) that the higher values
are on the diagonal and the lower values far off the diagonal. It is the
case for a covariance or a pseudo-covariance, but not for a variogram, for
which the terms on the diagonal are zeros.

As Pierre pointed out, the value A can be arbitrarly chosen because it is
canceled out in the equations thanks to the unbiasedness condition. In
practive, it is chosen to be larger than the maximum value of gamma_ij in
the matrix.

Regards,

Denis





.------------------------Denis ALLARD--------------------------------.
| Unite de Biometrie                       [EMAIL PROTECTED]    |
| INRA Domaine St Paul, Site Agroparc      tel:   (33) 4 32 72 21 77 |
| 84914 AVIGNON cedex 9, FRANCE            fax:                21 82 |
|                                                                    |
`--------- http://www.avignon.inra.fr/biometrie/welcome.html --------'



--
* To post a message to the list, send it to [EMAIL PROTECTED]
* As a general service to the users, please remember to post a summary of any useful 
responses to your questions.
* To unsubscribe, send an email to [EMAIL PROTECTED] with no subject and "unsubscribe 
ai-geostats" followed by "end" on the next line in the message body. DO NOT SEND 
Subscribe/Unsubscribe requests to the list
* Support to the list is provided at http://www.ai-geostats.org

Reply via email to