The residual dispersion is calculated from the dispersion present in the fiber and the dispersion caused by introducing the grating. The figure is simply demonstrating that using a FBG to account for the total fiber dispersion gives maximum performance.
However! You could conceive a simulation that would attempt to find the amount of dispersion experienced by a pulse. You would not be able to separate the dispersion from nonlinearity, unless you disabled that part of the fiber model, but you could use the optical time domain visualizer to view the chirp/phase for a pulse and from that value calculate the residual dispersion, assuming an input pulse without chirp.
Regards
.
