Is cokriging available in the current implementation?

[flydream0428]

Is cokriging available in the current implementation? Is there a plan to add this if not?

[mmaelicke]

Hey there,

No, unfortunately not. In fact, scikit-gstat only implements ordinary kriging. But there is an interface to gstools and pykrige and both have way more kriging algorithms available. I.e.:

import skgstat as skg

vario = skg.Variogram(coords, values, ...)
krige = vario.to_gs_krige(**args)

# i.e. use a structured grid
x = y = range(100)
field, sigma = krige.structured([x,y])

The **args part is passed down to gstools, refer to their docs for available parameters (model, cond_pos, cond_val are already set by Variogram)

Gstools does not have cokriging as far as I know, but external drift kriging which might be an option as well?

I'll close this issue, if there are further questions, feel free to open again