Microvibrations onboard greenhouse gases observing satellite (GOSAT) cause scan speed variations in the TANSO Fourier transform spectrometer. The associated periodic sampling errors generate ghost features in O2 A-band spectra, where surface pressure and aerosol properties are retrieved to determine the optical path through the atmosphere. A correction algorithm has been developed to re-compute the interferograms at equally spaced sampling intervals. The key is to determine iteratively the amplitude and phase of sinusoidal perturbations with predetermined frequencies to minimize the magnitude of the out-of-band ghosts artifacts after correction of the sampling grid. This correction algorithm drastically reduces errors in retrieved surface pressure and improves agreement with ground-based observations.