Application of SWI and FWI for Building Near-surface Velocity Model and Improving Seismic Image Quality

Abstract

Summary Building near-surface velocity model for static computation has traditionally been a crucial step in land data processing. As PSDM technologies expanded, building an accurate shallow velocity model has become an important step in the construction of full PSDM velocity model. The most popular velocity modelling techniques are based on the analysis of refraction and reflection kinematics. Each of these techniques has its inherent limitations. Refraction-based methods rely on the first break picking, which cannot always be done at the required quality level because of absorption, velocity inversions and interference with other types of energy. The use of reflected events is often impeded by limited offset range and low signal to noise ratio in the shallow intervals of data. Majority of the reflection-based methods make use of the kinematics of a single strong event, which often occurs at a substantial depth. Velocity distribution above this depth is approximated by a simplified model. Besides, the dynamics of data stay totally beyond the scope of these techniques. The paper presents new near-surface modelling methods - Surface Wave Inversion and Full Waveform Inversion. Based on different physical principles, these methods overcome limitations of the conventional approaches and allow creating detailed and accurate velocity models.