I don't know if I'm going mad or if I'm just that stupid, but when I look at...
I don't know if I'm going mad or if I'm just that stupid, but when I look at the vertical distribution of the POLCOMS data, it's now *not* upside down (and matches the depths). So, I've taken out the flipdim/flipud to leave the POLCOMS data as is. Also add the interpolated FVCOM profile which falls closest to the sample location in the debugging figure to check that the output is as expected (i.e. the depth is the right way around)