Surface-consistent amplitude corrections for single or multicomponent sources and receivers using reciprocity and waveform inversion

In land seismics, near-surface conditions often vary within surveys, resulting in differences in source strength and signature. Furthermore, discrepancies between closely spaced recordings are also commonly observed. Processing and interpretation of recorded data require that data are corrected for these source and receiver perturbations in the early stages of processing. However, existing surface-consistent deconvolution techniques are applicable to primary reflection data only, and therefore require that ground roll and multiples are suppressed prior to the application. This is usually performed with multichannel filter operations. The performance of these filter operations, however, rapidly deteriorates in presence of acquisition-related amplitude and phase perturbations. We propose an alternative approach to compensate for acquisition-related amplitude perturbations, which has the advantage of being purely a pre-processing step. It has the following characteristics: (i) it can be applied to complete recordings, hence does not require the isolation of primary reflections in the data, (ii) no assumptions are imposed on the subsurface and (iii) it is applicable to multicomponent data. The procedure is based on reciprocity of the medium response, so that differences between normal and reciprocal traces can be attributed to source and receiver perturbations. The application of reciprocity requires symmetric data acquisition, that is, identical source and receiver patterns, identical locations, and the source orientations have to be identical to the receiver components. Besides reciprocity, additional constraints are required to determine the lateral source and receiver amplitude variations fully. We use criteria based on minimizing total energy differences between adjacent common source and common receiver gathers, and in common offset panels of the medium response. Synthetic tests demonstrate that acquisition-related amplitude differences can be significantly reduced using this method.