[R-sig-Geo] quick question re: fit.variogram in gstat

[Edzer Pebesma]

What you see is the range parameter; as the Gaussian model reaches the
sill asymptotically, the value sqrt(3) * a is often used as the
"effective range", or distance value where 95% of the (partial) sill is
reached. Just plot the curve of 1-exp(-(h/a)^2) to see why this is so.

On 08/09/2010 05:06 PM, Kerry Ritter wrote:
> I have a follow up question regarding the range parameter specified for
> the gaussian variogram model in gstat..
> First, thank you to Jaime Grion who riplied...
> 
> Dear Kerry, if you write in the option model "Gau" you just have to
> specify 200 [for the range].
> 
> However, when I fit the variogram and the fitted value returns "200" as
> the range parameter, the variogram looks like the true range exceeds
> 200.  Is the returned range parameter the "practical range" = the value
> at which this model reaches 95% of its asymptotic value or the value
> where spatial correlation ceases to exist?.  In Bivand, et. al's book
> "Applied Spatial Data Analysis with R", the authors state that the
> practical range for the gaussian variogram model is sqrt(3)*a where a is
> the range parameter for most other models.  So my question remains.  Is
> the fit giving me the range or the practical range?
> 
> Thanks,
> Kerry
> 
> Kerry Ritter wrote:
>> Hi all. I have what I hope is a quick question concerning the use of
>> "fit.variogram" in"gstat".  If you assume a a Spherical model for the
>> variogram and the estimated range is say 200m and then you switch
>> models to a Gaussian but want to keep the range fixed at 200m, would
>> you specify 200 for the range parameter or 200/sqrt(3)?
>>
>> Thanks,
>> Kerry
>>
>> **********************
>> Kerry Ritter, Ph.D.
>> statistician
>> Southern California Coastal Water Research Project
>> 3535 Harbor Blvd., Suite 110
>>
>> work: 714-755-3210
>> cell: 714-420-3346
>> fax:  714-755-3299
>>
>> email: kerryr at sccwrp.org
>>
>> _______________________________________________
>> R-sig-Geo mailing list
>> R-sig-Geo at stat.math.ethz.ch
>> https://stat.ethz.ch/mailman/listinfo/r-sig-geo
>>
> 
> 

-- 
Edzer Pebesma
Institute for Geoinformatics (ifgi), University of Münster
Weseler Straße 253, 48151 Münster, Germany. Phone: +49 251
8333081, Fax: +49 251 8339763  http://ifgi.uni-muenster.de
http://www.52north.org/geostatistics      e.pebesma at wwu.de