Estimation of Lateral Correlation Length: A Quantitative Approach for Deep Reflection Seismic Interpretation

Abstract

Summary Deep reflection seismic interpretation is challenging due to the heterogeneity of the crystalline crust. Statistical approaches are suggested as a potentially powerful tool for deep seismic interpretation as the reflection wavefield can represent at least the spatial velocity heterogeneity of the geology. The lateral correlation length is an essential parameter of the von Karman stochastic model which is considered as a proper description of the heterogeneity of crustal velocity. To estimate the lateral correlation length, it is crucial to obtain accurate autocorrelation function from limited data sets. We introduce an innovative algorithm to estimate lateral correlation length from deep seismic reflection data. First, the power spectrum is obtained through autoregressive spectral estimation. Then inverse Fourier transform is applied to the power spectrum to obtain the autocorrelation function. Moreover, average multitrace one-dimension autocorrelation is used to replace the two-dimension autocorrelation function. A synthetic seismogram is used to verify the resolution of our method. The geological meaning of different seismic reflection patterns was discussed after applying our method to deep seismic reflection field data. The result shows that lateral correlation can identify Moho discontinuity, crustal reflections, and some structure in sedimentary strata