Improved shortest‐path ray tracing with locally linear velocity variations

Abstract

Shortest‐path algorithm has been used widely to calculate first arrival traveltimes and ray paths. It assumes that the velocity in each model cell is constant. When the velocity field varies continuously, models using too few constant‐velocity cells will result in too much error, and models using too many constant‐velocity cells will be inefficient. We made an improvement to this problem by allowing velocity increases linearly with depth at each model location. This improvement leads to a more accurate and faster ray tracing, and incorporates fewer model parameters for tomographic inversion at places of continuous velocity variation. The improved method is shown using 2D and 3D ray tracing examples.