An Adaptive Strategy for Robust and Efficient Computation of Dispersion Curves in a Layered Half-Space

Since the inception of surface-wave theory, the root-missing problem in dispersion calculations for layered media has been a persistent numerical challenge, particularly pronounced in complex models exhibiting mode kissing. Although subsequent studies have proposed various sampling strategies to mitigate this issue, they often result in excessive sampling and, consequently, low computational efficiency. Here, we introduce an adaptive sampling strategy that dynamically refines the search based on the local variation of the secular function, automatically capturing roots while keeping computational costs under control. Numerical tests demonstrate that the new strategy exhibits excellent robustness and efficiency and naturally retains outstanding generality. Users can confidently adopt adaptive finer sampling for complex models without concern for excessive computational burden, offering a powerful forward modeling tool for surface-wave inversion of shallow complex velocity structures.