Scattered array, 2D integration?

[Nicholas Jankowski]

I have model data that is a non uniform mesh over a rectangular area (x,y), and a single nodal value at each point. (The data is a n x 3 point-value array, [x,y,S]. I need to do a 2D numerical integration over that domain.  If it was a full rectangular grid, even if nonuniform, I could just call trapz twice.  With current functions implemented in Octave, is my only option to create such a grid a la meshgrid and griddata? Level of refinement on one corner of the mesh makes interpolating to a uniform mesh as per the scattered data example in the manual [1] fairly unwieldy (mapped mesh without loss of refinement can square the amount of data in each vector).

Matlab 2013 introduced a scatteredInterpolant function that they recommend using as an intermediary between the data and other integration functions, but no compatible option implemented in octave yet.  

Was hoping someone would have a less resource intensive method, or a good manual implementation of the interpolant scheme.

[1] https://octave.org/doc/v5.1.0/Interpolation-on-Scattered-Data.html

Nick J.