Stochastic Modelling of the Velocity Structure: Beyond Joint Inversion Methods

The traditional approach to modelling seismic velocity structure of the Earth with multiple data sets (such as receiver functions and surface wave dispersion curves) is to perform a joint inversion. This method produced a single solution and little information on uncertainty in the models. We explore an alternative approach enhancing the multiple objective forward modelling method with a stochastic search and optimization. A genetic algorithm is well-suited to the velocity modelling setting as it supports parallel optimization, maintains diversity in the search space, and is more likely than random search procedures to converge to the global minima.