A Case for the Use of 3D Attenuation Models in Ground-Motion and Seismic-Hazard Assessment