Borehole-consistent sparse-spike seismic deconvolution and frequency enhancement

Abstract

One issue that can be found when working with seismic data is the lack of temporal resolution, which complicates the geological interpretation. One way to alleviate this issue is by means of the so-called frequency enhancement techniques, whose main goal is to widen the data original bandwidth. Starting from a sparse reflectivity series obtained from well-log information, we first carry out a sparse-spike deconvolution of the observed seismic data trace-by-trace moving away from the well location. To this purpose, we use a warping function to modify iteratively each reflectivity until the resulting synthetic trace fits the observed data, just like in some welltying techniques. We then convolve the reflectivities with a suitable broadband wavelet to obtain seismic data with an increased time resolution. Lateral continuity is guaranteed by assuming that two contiguous reflectivities contain the same number of spikes with similar positions. We test the method on synthetic and field data. The results show that we can obtain high-resolution images with good lateral continuity that honor the observed data. One key aspect of the proposed technique is given by the fact that the resulting sparse-spike reflectivities and high-resolution images are consistent with the borehole information. Because of the aforementioned assumptions, the best results are expected for models with low-to-moderate structural complexity.