Imaging distributed acoustic sensing-to-geophone conversion data: A field application to CO2 sequestration data

Abstract

Compared with conventional geophone data, distributed fiber-optic sensing, including distributed acoustic sensing (DAS), can provide better spatial coverage for imaging the subsurface with finer spatial sampling. Because DAS measures subsurface seismic responses differently than the geophone, imaging technologies (e.g., reverse time migration and full-waveform inversion) that are developed for conventional geophone data cannot be readily applied to original DAS data without causing uncertainties in phase or depth, especially when one compares the DAS imaging results against the usual geophone imaging results. Based on vertical seismic profile field data from a CO2 sequestration site, we have compared the imaging results of the CO2 storage reservoir associated with the DAS and the geophone data, respectively, and we illustrate the differences between the imaging results of the DAS and geophone data. The difference between the DAS and geophone imaging results could be critical in obtaining time-lapse signals for monitoring reservoir changes, e.g., in subsurface CO2 sequestration. We propose to convert DAS to geophone data so that we can reduce the discrepancies between DAS and geophone imaging results and we therefore can reuse existing seismic imaging technologies. Two conversion methods, one physics-based and one deep-learning (DL)-based, are used for the DAS-to-geophone transformation. Field data results demonstrate that the DL-based approach can better successfully improve the alignment between the DAS and geophone images, whereas the physics-based solution is constrained by its assumption.