Linearized Rock Physics Inversion Based on Geostatistics

Summary Prediction of reservoir’s physical parameters using seismic data is a significant research topic in seismic interpretation, as it can characterize quantitatively the link between reservoir properties and seismic response. Geostatistics plays an important impact on improving the effect of reservoir prediction since it can provide useful prior knowledge for seismic inversion. In this paper, a novel linear rock-physical inversion strategy is proposed based on the Bayesian theory and geostatistics. Firstly, the prior information is characterized according to the geostatistical theory. Then, Aki-Richards formulation and the first-order approximate Kuster-Toksöz inclusion model are used to determine the relationship between porosity and seismic data. Finally, the inversion results are acquired under the Bayesian framework. The model test illustrates inversion results have a high resolution and agree well with the true model.