Stochastic Seismic Inversion

MPSI UltraFast stochastic inversion scheme is a joint venture project with Earthworks reservoir which is considered the fastest stochastic seismic inversion scheme currently available.

Product overview

• Full bandwidth model based on seismic inversion technology
• Low cost, no fuss unlike most stochastic inversion algorithms which are slow, expensive or both
• Available as an OpendTect plugin
• Rapid, robust, inexpensive

Product description

These ARK CLS / Earthworks OpendTect plugins are collectively called MPSI and include deterministic and UltraFast stochastic seismic inversion. Whether deterministic or stochastic, the inversion procedure is model-based. A 3D broadband impedance model of the sub-surface is constructed using well and picked seismic horizon data. Geostatistical gridding (kriging) is constrained by 3D anisotropic variograms. In addition, the user specifies a 2D error grid which is used to provide spatial variation in the constraints used in the subsequent seismic inversion. The error grid is a geostatistical standard deviation map based on the wells. The user can specify the relative importance of the seismic data and the model in the inversion, allowing the model to take priority close to the wells (where the impedance is known) and to relax the model constraint and use more seismic information away from the wells.

Using a 3D impedance model and error grid, subsequent inversion can be either deterministic or stochastic. In a deterministic inversion the objective is to estimate the unique mean or average impedance, based on least squares linear inversion. In a stochastic inversion the goal is to estimate a set of realisations of impedance representing the uncertainty in the seismic inversion. The impedance realisations from stochastic inversion can then be used to estimate the uncertainty in reservoir properties such as lithology or porosity in the form of probability maps and volumetric uncertainty. The stochastic realisations can also be used to estimate reservoir connectivity and associated connected or swept volume uncertainty. The stochastic inversion menu provides a comprehensive suite of utilities to allow the user to analyse the realisations in this way.