Efficient seismic imaging with the double plane wave data
Seismic imaging is critical in providing the image of the Earth’s subsurface, and it plays an important role in hydrocarbon explorations. Obtaining high resolution images with accurate reflectivities and accurate positions of subsurface structures is the goal for exploration geophysicists. Reverse time migration (RTM), which solves the two-way wave equation, can resolve all wavefield propagation phenomena. In geologically complex regions, RTM has been proven to outperform other imaging methods in correctly revealing the subsurface structures. However, implementing the traditional pre-stack shot profile RTM is computationally expensive. Time consuming wavefield propagation processes need to be performed for each shot gather to obtain high resolution images. The traditional RTM can become extremely expensive with increasing shot numbers. In this dissertation, I focus on improving the migration efficiency of the RTM using the double plane wave (DPW) data, which are the fully decomposed plane wave data. Three RTM methods are developed to migrate the DPW data, all of which can improve the migration efficiency comparing to the traditional shot profile RTM. Two of the methods utilize the adjoint state method, and they are known as the time domain DPW-based RTM and the frequency domain DPW-based RTM. A third migration method using the DPW data is derived under the Born approximation. This method employs the frequency domain plane wave Green’s functions for imaging, and it is named as frequency domain DPW RTM. Among the three proposed RTM methods, the frequency domain DPW RTM is the most efficient. Comparing to the traditional shot profile pre-stack RTM, the frequency domain DPW RTM can increase migration efficiency of RTM by an order of magnitude, making the frequency domain DPW RTM a preferable option for migrating large seismic datasets. All of the three proposed migration methods can image subsurface structures with given dips, which makes them target-oriented imaging methods. The proposed methods are beneficial to migration velocity analysis. To improve the resolution of migration results, a least squares RTM method using the DPW data is proposed. A Born modeling operator that predict the DPW data at the surface and its adjoint operator, which is a migration operator, are derived to implement the least squares RTM. Both of the operators require only a limited number of plane wave Green’s functions for the modeling and the migration processes. The proposed least squares RTM substantially increases the efficiency of the least squares migration. In the DPW domain, the applicability of the reciprocity principle is also investigated. The reciprocity principle can be applied to the seismic data that are processed with proper seismic processing flow. Utilizing the reciprocity principle, a DPW dataset transformed from one-sided shot gathers can approximate a DPW dataset transformed from split-spread shot gathers. Therefore, I suggest that one-sided acquisition geometries should be extended to the largest possible offsets, and the reciprocity principle should be invoked to improve subsurface illumination. Migration efficiency can be further improved with the help of the reciprocity principle.