Is the fault core-damage zone model representative of seismogenic faults? Pre-existing anisotropies and fault zone complexity