Large 3D scatter interpolation: time x longitude x latitude

Dear All,

I have 2 Matrices:

Matrix 1: about 60 millions points (time x long x lat x windspeed) from unstructured grid models

Matrix 2: about 2 millions points (time x long x lat x windspeed) from Altimeter data

All points are randomly distributed. I want to interpolate model data into altimeter using scatteredInterpolant as follows:

F=scatteredInterpolant(modeldata(:,1), modeldata(:,2), wwmdata(:,3),wwmdata(:,4));
newwind=F(altimeter(:,1),altimeter(:,2),altimeter(:,3));

It was out of memory even if I ran the script on supercomputer.

Could you please suggest what wrong with this and how could interpolate this data effectively?

Many thanks

tanq