Hello

On the following link, You can find an example of code for the geostatistical analysis. But it is written in Spanish. I hope it's helpful

http://downloads.tuxfamily.org/tuxgis/geoblogs/geoestadistica_mapa_de_prediccion/EjemploMapaGeoestadistica.rar

Thank you Miguel.

Hi, Glory,

maybe you can find this useful...

http://cran.r-project.org/web/packages/gstat/vignettes/gstat.pdf

Are you wanting to interpolate separately in the three layers or are you wanting to treat it as a 3-D problem?  I suggest that you do some exploratory statistics before deciding

1.  Are the histograms of the data sets for the three layers quite similar or quite different?

2. Are the sample variograms for each of the three layers quite similar or quite different? 

3. Depending on the attribute(s) you are interested in, the expected value (of the random function) may change with depth which would mean that you need to use different residuals for each layer in fitting the variogram model(s). Try fitting a regression based on depth separately for each layer. The possible effect of depth may depend on how great the differences in depth are for the three layers.

gstat will allow you to use a variogram in 3-D and it will also allow you to use a polynomial depth regression model in the kriging.

Avoid simplistic answers and try to understand the phenomena you are trying to model and in particular the effect of the different depths.

Finally for gstat you want a plain text file, not an EXCEL file and you don't want replications, that conflicts with the basic assumptions implicit in the use of gstat

Thank you Prof Myers. Your answers are always very insightful and thought provoking. I sincerely appreciate your response