## Re: AI-GEOSTATS: Choosing Lag Distance and Angular Tolerance

Expand Messages
• Andrew You can apply standard geostatistics if the measurements are the average (or some similar feature) over an area. It makes interpeting the
Message 1 of 5 , May 15, 2001
• 0 Attachment
Andrew

You can apply 'standard' geostatistics if the
measurements are the 'average' (or some similar
feature) over an area.

It makes interpeting the semi-variogram extremely
tricky if you combine many different sizes of sample,
but common sense is the main thing here. The trick is
to derive a point semi-variogram model from which any
size can be derived (see Chapter 3, Practical
http://uk.geocities.com/drisobelclark/practica.html)

Kriging is modified to reflect that the samples are
averages, mainly by changing the diagonal elements in
the equations so that they are non-zero. I don't know
any software package (off hand ) that does this,
though.

Isobel Clark

--- Andrew Mullens <andrew_mullens@...>
wrote: > I have a question relating to this question,
> certainly not to question the
> previous writer, it just seems like a good time to
> bring it up. Will
> variograming and other such techniques work for the
> data the previous writer
> described, e.g samples aren't at points, but areas
> (and areas that might
> have very little to do with the question). If they
> did use points in the
> calculations where would the points be placed, at
> the center of the county,
> at the major population center, at some arbitrary
> point (e.g most northerly
> point).
>
> I may be miss reading the description, perhaps the
> sample are point samples,
> but were taken with one sample in each county.
>
> Obviously the point samples are never really point
> sample, they must be
> taken over some area, approximating a point, but
> does this design seem to
> push the boundaries on that assumption.
>
> Andrew
> ----- Original Message -----
> To: <ai-geostats@...>
> Cc: <waghei_y.MED.TMU@...>;
> <mohsen_m@...>; <waghei@...>;
> <mohsen_m@...>; <kazem_an@...>;
> Sent: Tuesday, May 15, 2001 7:33 AM
> Subject: AI-GEOSTATS: Choosing Lag Distance and
> Angular Tolerance
>
>
> > Hello dears
> > I have a spatial data set contaning n=262
> observarion (The variable of
> interest is Rate of Tuberculosis in 262 counties of
> Iran). I want to fit
> some models to Directional semi-variograms,and then
> build anisotropic
> semi-variogram.
> > Then questions are
> > - Is there any rule for choosing Lag Distance and
> Angular Tolerance?
> > -Also,how we can balance between Lag Distance and
> Angular Tolerance?
> > -Do you agree that both must be very small,as
> possible?(Such that number
> of pairs in each lag>20, for example)
> >
> > Thank you
> > Dep.of Biostatistics
> > Tarbiat Modarres Univ.(Tehran)Po.Box: 14115-111
> > Tel:8011001-3872 Fax:8007989
> >
>
___________________________________________________________________________
> > Visit http://www.visto.com/info, your free
> web-based communications
> center.
> > Visto.com. Life on the Dot.
> >
> >
> > --
> > * To post a message to the list, send it to
> ai-geostats@...
> > * 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
> majordomo@... 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
> >
>
>
> --
> * To post a message to the list, send it to
> ai-geostats@...
> * 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 majordomo@...
> 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

____________________________________________________________
Do You Yahoo!?

--
* To post a message to the list, send it to ai-geostats@...
* 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 majordomo@... 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
• ... wrote: ... This is something I learned about in quantitative ecology in the 1960 s ;-) You take each sample in turn and find the sample closest to it.
Message 2 of 5 , May 15, 2001
• 0 Attachment
wrote: >
>
> Could you explain what you mean by "nearest
> neighbour distances between samples locations"?
This is something I learned about in quantitative
ecology in the 1960's ;-)

You take each sample in turn and find the sample
closest to it. Note the distance. A histogram of the
distances will give you a sort of natural inter-sample
distance for use in the semi-variogram calculation.

The histogram can also be used to check whether
sampling is randon, uniform, clustered etc.

I don't have a full reference but it should be easy to
find: Pielou, Cressie or any other book with Spatial
Statistics in the title.

Isobel Clark

____________________________________________________________
Do You Yahoo!?