1

I was wondering what sets of equations are used to fit an experimental variogram when I choose it to be a spherical model. For instance, I used the following parameters of a spherical model:

  model    psill    range
1   Nug 5384.346   0.0000
2   Sph 9042.061 944.9017

here, sill variance, c = 5384.34 + 9042.06 = 14426.41; and range, a = 944.9017. A spherical model is explained by the following equation (book name: Geostatistics for Environmental Scientists):

variogram = c*{(3h/2a)-0.5(h/a)^3} when, h<=a
variogram = c when, h>a

where, h = distance; therefore, if I plug-in the values 'a' and 'c' given a known distance 'h' using the above equation, the coordinate (h, variogram) should fall on the model. Unfortunately, it doesn't do that for different values of h! Is it possible to know what sets of equations are used in the case of different models (spherical, exponential, etc.) in the 'gstat' package?

ToNoY
  • 1,358
  • 2
  • 22
  • 43
  • 1
    Have you looked at the `gstat` documentation? Looks like thy have [a nice vignette](https://cran.rstudio.com/web/packages/gstat/vignettes/st.pdf), which includes a short biography. – Gregor Thomas Nov 19 '15 at 00:39
  • I should look into that again but I think they didn't explore the equations in great detail. – ToNoY Nov 19 '15 at 00:44
  • Oops, I meant bibliography. I would bet the peer-reviewed publications they cite have equations. – Gregor Thomas Nov 19 '15 at 01:14

0 Answers0