Damage zones are important to the rupture dynamics, evolution and fluid coupling of earthquakes. However, information about the damage zone at depth is limited. It is unclear if damage zones increase or decrease in intensity with depth. Here we use marine 3-D seismic surveys and modern fault detection methods to address the depth-dependent structure of damage zones. We use two overlapping legacy industry seismic volumes collected offshore of Los Angeles span approximately 20 km of the Palos Verdes strike-slip fault. The data here allows visibility of the damage zone in the sedimentary formations to 2,200 meters depth, which is comparable to the constraints provided by SAFOD and other studies. Using both interpreted mapped primary fault strands and seismic attributes to identify subsidiary faults, we map and quantify spatial variations in damage zone size and intensity. The damage zone consists of subsidiary faults, or linked discontinuities in the seismics selected within assigned ranges of geometries to the primary strands. Damage was identified using a variation of the seismic attribute semblance, or multi-trace similarity. This method allows interrogation of damage zone in response to changes sedimentary lithology and fault geometry. Subsidiary faults delineate the damage zone to approximately 1 km in width and fracture density decays with distance from the primary fault strands for all sedimentary lithologies in the study area. The damage zone narrows with depth, but fracture density increases because the intensity of fracturing more than compensates for the decreased width. In the thickest formation we find that fracture density increases as Z1.8, where Z is depth in meters. These results are then compared to resolution changes with depth. The damage intensity increase and localization potentially provides a strong constraint for efforts to determine an appropriate rheology for producing damage zones and studying their effects.