Ordinary Kriging

2D-interpolation using geostatistics

Description

This function completes what I have started with the functions variogram and variogramfit. It is not intended to be a highly optimized code for performing kriging but may have educational value. Note that for using kriging, you should download the latest version of variogramfit.

kriging uses ordinary kriging to interpolate a variable z measured at locations with the coordinates x and y at unsampled locations xi, yi. The function requires the variable vstruct that contains all necessary information on the variogram. vstruct is the forth output argument of the function variogramfit.

This is a rudimentary, but easy to use function to perform an ordinary kriging interpolation. I call it rudimentary since it always includes ALL observations to estimate values at unsampled locations. This may not be necessary when sample locations are not within the autocorrelation range but would require something like a k nearest neighbor search algorithm or something similar. Thus, the algorithms works best for relatively small numbers of observations (100-500). For larger numbers of observations I recommend the use of GSTAT.

Acknowledgements

The author wishes to acknowledge the following in the creation of this submission:
Experimental (Semi-) Variogram, variogramfit

• MATLAB release: MATLAB 7.10 (2010a)

[Mohammad Javad Tourian]

 

[Yadong]