What constitutes a valid intercomparison of satellite and in situ stratospheric H2O measurements?