It's a response spectrum, your looking at a system's response to the event. A rigid mode will track ground acceleration (ZPA in fact) and a very, very flexible mode will reveal the ground displacement. (I the first sentence I reference system response and in the second I emphasize ground input.)
If not entered in multi-spectra input, CAESAR II sets this anchor displacement on the lowest entered frequency. Some users ignore the low frequency end of their spectrum thinking there is no system mode that low to be excited. If you don't go low enough in frequency, you may end up with large values for this derived displacement and then big stress. Look for the "P" in the label for the largest "modal component" of the total stress.
In many installed piping systems the strain component does more damage than the inertia component.