[ai-geostats] modeling cross-semivariogram
I intended to seek for scale-dependent correlations among soil properties by performing a multivariate geostatistical analysis (factorial kriging analysis, FKA). GS+ software (version 5.1.1) was used to compute and model the experimental simple and cross-semivariograms. However, I found a wrong cross-semivariogram result (i.e. the wrong result of computed experimental cross-semivariogram, and so a wrong fitted semivariogram model) was produced during covariance analysis when selecting “Nonuniform intervals”. Does anyone know how to fix this problem?
In order to obtain results of cross-semivariogram analysis, I tried to fit an authorized semivariogram function such as a spherical model by using the “nlinfit” (nonlinear least-squares data fitting) in Matlab. The model fit was not always good enough by checking the displayed figure with the fitted model as well as the regression coefficient. To be honest, I am a beginner and know little about Matlab. Can anyone tell me whether it is appropriate to use “nlinfit” in Matlab for cross-semivariogram model fit or not?
Is there any other free software (easy to use for beginner like me) for multivariate geostatistical analysis?
I should really appreciate anyone’s reply!
Thanks a lot ahead!
Do You Yahoo!?