Kriging

Developer
Jan 28, 2012 at 11:28 AM

I've seen that there is a kriging plugin, in DotSpatial and in MapWindow. I start using DS without previous using MW, so I don't know very well how plugin are structured.
Is the kringing plugin source code available ? i have not see them in this list

http://www.mapwindow.org/apps/wiki/doku.php?id=download

Many thanks

Tummi

Jan 30, 2012 at 3:53 PM
Edited Jan 30, 2012 at 3:54 PM

The code source for Kriging plugin  in MapWindow is available in  http://svn.mapwindow.org/svnroot/MapWindow4Plugins/KrigingPlugin.  For DotSpatial is 

http://geostatisticaltool.codeplex.com/.

Carlos

Developer
Jan 30, 2012 at 5:37 PM

I'm trying to write an "automatic" kriging like Surfer that creates a linear variogram, to automate some process of interpolation so I was making some esperiment and I was writing a c# version of this

http://www.nbb.cornell.edu/neurobio/land/oldstudentprojects/cs490-94to95/clang/kriging.html

and also on GeoEAS

I was also reading this

http://www.goldensoftware.com/variogramTutorial.pdf

And the help of surfer that explicates how they creates a linear standard interpolation of variogram (you can find it also in demo version)

But is not very clear how the line is created and the theory is on this interpolation, then is different in  different versions of Surfer.

Did you never think on this ? suggestions ?

Many Thanks

Tommi

Jan 30, 2012 at 6:54 PM

Hi..

In the Mapwindow version 4.8.6, the linear model is implemented..  I don't include this linear variogram in the geostatistical tool.

The automatic kriging is not "really" possible, there are infinite solutions for a variogram taking into account parameter like number of lags, size of lag,

the anisotropy, type of theoretical model (i.e Spherical, linear, circular). When one of these parameters changes, there is a new solution.. 

By default all Geostatistical packages generate a default initial solution, but it doesn't mean that it is the best solution.  

For obtaining the best theoretical model is needed to fix the empirical semivariances or Covariance  to selected model.  There are some 

algorithms used to determine the best model, WLS or OLS,

(Cressie, N., 1985. Fitting variogram models by weighted least squares. Math. Geol. 17, 563–586.) In this article you can find the most common method to fit a variagram..

(Least-squares ftting from the variogram cloud, muller,1998)

In the Geostatistical Tool is used an OLS to get the best model.. 

Carlos

 

Developer
Jan 31, 2012 at 6:43 AM

I'll to read your source code and the documentation you suggest.

Many Thanks.

 

Jul 12, 2013 at 9:38 AM
Is there any chance that also a co-Kriging algorithm will be available through geostatistical tools?

Thanks
Oscar