Joint seismic deconvolution and geostatistical extrapolation in a Bayesian framework using a B-spline discontinuous Galerkin method

We propose a method for simultaneously performing seismic deconvolution and geostatistical well log extrapolation in order to optimally estimate subsurface reflectivity. Traditional zero-offset seismic deconvolution and Kriging (geostatistical extrapolation of point samples) can be seen as special cases of this more general problem. We model the reflectivity field as a continuous function and incorporate prior geological knowledge in the form of local correlation information and discontinuities (faults) obtained from the processed seismic data.