Many vintage reflection seismic surveys exist that have nonrepeated acquisition geometries or final-stacked/migrated sections are obtained using different or nonconsistent processing flows. This may lead to derived time-lapse attributes that are not internally consistent or even conflicting. For instance, we have focused on a subsurface gas blowout that occurred in 1989 in the Norwegian sector of the North Sea. The 2D site survey data were acquired in 1988 and 1990, and the 3D data were acquired in 1991 and 2005. The various sets of legacy data are plagued by poor repeatability among data acquisitions, application of different processing strategies, missing prestack data, and the presence of multiples. All of these factors severely complicate even qualitative interpretation of the gas migration associated with the underground blowout. Careful time-lapse processing may provide useful information even from such challenging legacy data by first computing numerous attributes including instantaneous amplitude differences, time shifts, time-lapse attenuation, and impedance inversions. Then, judicious quality control, invoked by comparing the various attributes, was used to check for internally consistent results.