Conventional 3D and machine learning-based 2D surface wave tomography are used to obtain a data-driven velocity model which helps illuminate the extent of the fault zone damage. Geological mapping studies (e.g., Hough et al., 2020) find extensive distributed faulting in a zone several kilometers wide around the 2019 Ridgecrest ruptures. While, the depth extent is unknown, we assume that this documented off-fault deformation in Ridgecrest is the surface expression of underlying fault damage characterized by a low-velocity zone (LVZ).Accurate assessment of such fault zone damage is important for many applications, including earthquake location and seismic hazard analysis. For example, impedance effects and trapped waves in fault zones can cause ground motion amplification (e.g., Parker et al., 2020), and rupture dynamics simulations suggest fault zones can affect the rupture pulse (e.g., Harris & Day, 1997) and cause super-shear rupture (e.g., Gabriel et al., 2012). Furthermore, nonlinear rheology in the fault zone may affect the rise time and shape of the rupture