RE: [ai-geostats] kriging interpolation time vs depth
It is possible to calculate a semivariogram for one-dimensional data with the same method used for two dimensional (spatial) data. The problem arises in kriging because of the string effect, in which the data points at the extreme end of the search window are weighted too high. This can be corrected by modifying the kriging algorith as described by Deutsch, 1993. Essentially, you wrap the search window around to a “circle". I don’t know Idl, so I can’t help you with those specifics.
Deutsch, C.V., 1993. Kriging in a finite domain. Mathematical Geology, 25: 41-52.
I used the process to modify the GSLib program to interpolate timeseries of groundwater levels, and it is described in:
Barabás, N. and P. Goovaerts. 2004. Comparison of geostatistical algorithms for completing groundwater monitoring well timeseries using data of a nearby river. In X. Sanchez-Vila, J. Carrera, and J. Gomez-Hernandez, editors, geoENV IV - Geostatistics for Environmental Applications. Kluwer Academic Publishers, Dordrecht , pp. 199-210.
From: vv [mailto:vincenzo.vellucci@...]
Sent: Wednesday, November 23, 2005 1:19 PM
Subject: [ai-geostats] kriging interpolation time vs depth
I'm new in this list, and I'm coping with a 2D interpolation.
My data are irregular and I have time on the abscissa and depth on the ordinate.
I would like to perform a kriging interpolation by using KRIG2D function in Idl.
My question is: how may I, if possible, calculate the variogram for such data?
and, if not, can anybody suggest me the correct routine to use to perform a correct
interpolation for this kind of data?
thanks in advance