Find nearest grid cell between two 2D matrixes of different resolution

Hello,

I have two matrices I would like to compare.

A is 192(lon) x 288(lat) B is 360(lat) x 720(lon)

How can I regrid B so that its resolution is the same as A?

I would like to know how to do this both by finding the nearest B lat&lon&values for A, and also by decimating B to the coarser resolution of A

Thank you,

Melissa

Answer 1

[Walter Roberson]

Assuming that they are regular grids with known latitude and longitude vectors, Along, Alat, Blong, Blat, then the approximation that neglects curvature of the Earth would be

[X, Y] = meshgrid(Blat, Blong);
[Xq, Yq] = meshgrid(Alat, Along);
Bapprox = interp2(X, Y, B, Xq, Yq);

The interp2() would need the longitude vectors phase-unwrapped if they cross 180 to -180:

degrad = pi./180;
unwrappedBlong = unwrap(Blong .* degrad) ./ degrad;

If you do need to take the curvature of the Earth into account, then possibly mapprofile might turn out to be usable for you -- I am too tired at the moment to figure out what a refmatrix is.

If mapprofile is not suitable then perhaps someone has already built a routine. If not then intrplon might help.

[Melissa]

thanks! What do I do if it says,

Error

The grid was created from grid vectors that were not strictly monotonic increasing.

[Walter Roberson]

That's when you have to do the unwrapping like I showed.