# Is there a recursive version of Kriging or Inverse Distance spatial interpolation?

Classic use-case of Kriging: you have a 2d space, you have $n$ observations, each of them representing an exploratory dig. It has a $x$ and $y$ coordinate, and a $V$ representing the value discovered by that dig.
You can use Kriging (or IW) to interpolate these observations and get an estimate for the $V$ at any location $(x,y)$.

Now my issue is as follows: every day, I dig a new hole and get a new observation $(x_{n+1},y_{n+1},V_{n+1} )$ .
Is there a way to compute the new spatial interpolation without recomputing the whole thing?
Basically is there a spatial interpolation equivalent of what the Kalman Filter does for linear regressions?

I am very open to alternative interpolation algorithms, but best if they compute some confidence interval as well.

• I believe that if you search google scholar for "sequential kriging", you might find some hits that could point you in the right direction. – Edzer Pebesma Oct 18 '15 at 16:56
• Thank you for the suggestion. Looks like just what I was looking for. – CarrKnight Oct 22 '15 at 11:28