MOTIVATION AND PROGRAMS

This model study is motivated by the observation that refraction inversion results often do not make sense geologically. Typically the inverted refractor model thickens over regions of elevated terrain. This is counter intuitive. The low velocity layer is expected to thin out or disappear in elevated regions and thicken in basins. This would be consistent with the occurrence of sediments in valleys and outcropping formations in elevated regions. Refraction inversions often converge to solutions which exhibit the exact opposite behavior.

I wrote a ray tracing program to calculate realistic first break pick times for models with complex topography. The program calculates the arrival time of refracted head waves by integrating the ray path along the interface. Once the refraction paths are calculated, the program checks for the possibility of valid transmitted paths. In areas of large topographic relief, energy impinging upon the interface between two media at angles slightly less than the critical angle is propagated under the refractor as transmitted waves. If these paths exist, their travel times are calculated and designated as first arrivals. These arrivals can be substantially faster than the head wave arrivals.

The program RINSE (Ray Inversion for Near-Surface Estimation) is used to demonstrate the effect of transmitted arrivals on refraction inversion. RINSE utilizes the slope of the first break travel time curves to determine the direction of rays arriving at a group and rays leaving a shot. Ray travel times from shot to group are then compared to the actual travel times to determine the depth to the refractor and refractor velocity. These results are then used to calculate near surface static delays. A detailed description of the algorithm is given by Jones and Jovanovich (1985).