Digby
That is the 'traditional' cross semi-variogram as
discussed in Matheron's original work. Now also known
as a co-located cross semi-variogram.
There is a non-co-located cross semi-variogram which
goes something like:
gamma(h)=1/2N(h) SUMi,j(vi-uj)^2
which is always positive. However, you
To those who are interested in "interpolation" for
any dimensions (1D, 2D, 3D,..). We have good news to you.
To date, there are no existing publications
regarding interpolation methods that can be deduced analytically and extended
easily to arbitrary dimensions. In fact, there is no
