Joint Analysis of Refractions and Reflections (JARR) Method for Quantitatively Deriving Velocity Models

Abstract

Joint analysis of refractions and reflections (JARR) is introduced and is developed as a quantitative method for improving the accuracy of near-surface velocity model functions. These accurate near-surface velocity functions are essential in various reflection processing flows. With reflection processing, accuracy is vital to produce accurate subsurface stacked sections. To demonstrate the method, it was evaluated by calculating long-wavelength statics and post-stack depth migration on high resolution data. Seismic data were statically corrected and a migration was applied using velocities derived from the JARR method facilitated by First Arrival Tomography (FAT). One sample dataset is from the Wellington petroleum field, Wellington, Kansas and was intended to image to the basement structure (1500m). A second dataset was acquired along Highway 61, Inman, Kansas and intended to image solution features in the near-surface (upper 500m). The JARR method utilizes a specialized processing flow designed to produce more accurate near-surface velocity functions than traditional velocity analysis methods. The method starts from raw interval normal move-out (INMO) velocity functions determined from reflections utilizing standard velocity estimation techniques (velocity panels, semblance). After a traditionally defined velocity function is selected, first arrivals are picked and saved. First arrival tomography is performed using the a-priori damped reference INMO velocity function. Synthetic first arrival rays are passed through the reference model. Time differences from calculated and observed first arrival rays are then inverted to produce a new velocity function. The process is repeated iteratively. The aforementioned method has many potential applications; however, it has the greatest impact in determining long-wavelength static corrections and high-resolution migration velocities. Results, in challenging areas, show the JARR method to be a unique and novel way for calculating accurate near-surface velocity models when other approaches have shown marginal effectiveness. Results also support the utility of the JARR method for calculating velocities used to determine long-wavelength statics and migration of data with complex near-surface lithology and structures. These achievements were due to the JARR methods capability of producing velocity models with increased spatial and temporal accuracy.