Improve horizon mapping of legacy seismic data using high-resolution impedence data from stochastic inversion
Amplitude changes on seismic records are driven by the impedance contract in the subsurface layers.
The process in which seismic interpreters identify the lithological interface between the prominent subsurface intervals using seismic reflection data is referred to as seismic horizon picking.
Depending on whether the top reservoir and the overburden layer form a strong or a weak impedance contrast, the horizons may be picked at the peak, through or even zero crossings. In many cases, you may find that a sequential stratigraphic boundary doesn't necessarily associate with an impedance contrast, increasing the guessing element in horizon picking. As a general point, it is important to understand the seismic horizon does not equal to geological horizons.
In recent months, we have received an increasing number of inquiries from clients on our stochastic inversion services who had problems tracking horizons on poor seismic data.
If the seismic data is carefully processed, the problem of untrackable horizons is not necessarily caused by the poor signal-to-noise ratio. It is often caused by a high degree of lateral variation of the amplitude and weak seismic reflection associated with thickness variation, tuning effect, lateral lithological changes and even fluid content.
When the continuous amplitude of a horizon is absent, horizon picking can be a challenging task. Many may choose to simply fill the gaps between the observable sections (seed horizons) while assuming the reservoir interval being equal thickness.
For a thin reservoir, this practice will produce huge errors in estimated reserves. For example, if the average thickness of the reservoir is 20 metres, a 2-meter misinterpretation can result in a 10% difference in reserves. When you find that it is difficult to match the reservoir model predictions with the historical pressure data or flow rate, you may want to check the horizons used as input for geomodelling.
For a field redevelopment project, special attention needs to be paid to horizons as the remaining potentials can't be identified without a good understanding of the reservoir structures.
One good method to improve the structural interpretation of a reservoir with lateral thickness variation and relatively small thickness is using impedance data from inversion data.
Unlike a full-blown pre-stack stochastic inversion workflow designed for characterization of lithology variations, in order to improve seismic horizon interpretation, you often only need the p-impedance from stochastic inversion applied to the post-stack processed data.
When compared to colour inversion, sparse-spike inversion or 3D volumetric seismic attribute, high-resolution impedance from post-stack stochastic inversion has a clear advantage in providing a clear image of the thin reservoir layers.
Left: Sparse-spike inversion. Right: Stochastic inversion.