That you can't find it is a good thing.
Lagrange interpolation is a nice thing for ONE purpose only: to teach students some basic ideas. What those teachers fail to followup with is that it is a bad thing to use when you really need to do interpolation. So then those students go into the world, and try to use it. Worse, then they want to do stuff like use it for 2-d interpolation. Don't do it. The idea has "bad" written all over it.
There are many schemes around for 2-d interpolation/modeling. Use one of them. Start with scatteredInterpolant, or you might use radial basis function interpolation (I recall it being on the FEX), or approximation tools like my gridfit (on the FEX), or neural nets. Or there is Kriging, or use splines.
