Skip to content

can we use georob package to make a 3D nested variogram?

6 messages · Bingwei Tian, Edzer Pebesma

#
Dear list,
It seems only gstat can make a 3D nested variogram but can not auto fit it
until now, .
Recently I find georob is also a useful package to do 3D estimation with
trend data, 
But I did not find the way to create a  3D nested variogram, Does anyone
kown about this,
or any other package for fit  a  3D nested variogram?
Thank you in advance!



-----
Bingwei

Ph.D. Student 
Kyoto University
C-1-2-225, Katsura Campus, Kyoto University, 
Nishikyo-ku, ?615-8530, Kyoto, Japan
--
View this message in context: http://r-sig-geo.2731867.n2.nabble.com/can-we-use-georob-package-to-make-a-3D-nested-variogram-tp7587522.html
Sent from the R-sig-geo mailing list archive at Nabble.com.
#
On 12/05/2014 05:08 AM, Bingwei Tian wrote:
?fit.variogram gives me

Description:

     Fit ranges and/or sills from a simple or nested variogram model to
     a sample variogram

Or do you mean, by "auto fit", that it should also chooses the number of
nested structures, and the corresponding models automatically?

  
    
#
Thanks edzer, As you said , I hope it can choose the number of
nested structures, at least keep the eyefitted number and the corresponding
models automatically.
But, for my case , the fit seems do noting except change the second Nug to 
1.

R:> uk.eye1  <- vgm(psill = 0.155,  model = "Gau",  range=700,  nugget=0.01)
R:> uk.eye   <- vgm(psill = 0.125,  model = "Sph",  range=35000,  nugget=0, 
add.to=uk.eye1)
R:> uk.eye
  model psill range
1   Nug 0.010     0
2   Gau 0.155   700
3   Nug 0.000     0
4   Sph 0.125 35000

R:> uk.fit  <- fit.variogram(uk.vgm, uk.eye, fit.sills = TRUE, fit.ranges =
TRUE,
+ fit.method = 7, debug.level = 1, warn.if.neg = FALSE)

Warning: singular model in variogram fit

R:> uk.fit
  model psill range
1   Nug 0.010     0
2   Gau 0.155   700
3   Nug 1.000     0
4   Sph 0.125 35000



-----
Bingwei

Ph.D. Student 
Kyoto University
C-1-2-225, Katsura Campus, Kyoto University, 
Nishikyo-ku, ?615-8530, Kyoto, Japan
--
View this message in context: http://r-sig-geo.2731867.n2.nabble.com/can-we-use-georob-package-to-make-a-3D-nested-variogram-tp7587522p7587524.html
Sent from the R-sig-geo mailing list archive at Nabble.com.
#
Well, yes, you can't fit two nuggets independently, can you? Either fix
one of them, or remove one of them. Note that these two are different
models, when it comes to fitting parameters:
model psill range
1   Sph     1   900
model psill range
1   Nug     0     0
2   Sph     1   900
On 12/05/2014 08:56 AM, Bingwei Tian wrote:

  
    
3 days later
#
Yes, you are right, only one nugget effect can be fitted.
I fitted the variogram with one nugget, but still did not better than eye
fitted.



-----
Bingwei

Ph.D. Student 
Kyoto University
C-1-2-225, Katsura Campus, Kyoto University, 
Nishikyo-ku, ?615-8530, Kyoto, Japan
--
View this message in context: http://r-sig-geo.2731867.n2.nabble.com/can-we-use-georob-package-to-make-a-3D-nested-variogram-tp7587522p7587534.html
Sent from the R-sig-geo mailing list archive at Nabble.com.
#
On 12/09/2014 02:13 PM, Bingwei Tian wrote:
If your criterion for "better" is visual, then you must be right. If
your criterion involves minimising a weighted sum of squares similar to
what gstat::fit.variogram uses, I would be happy to receive from you a
reproducible example.