In the previous Part of this study, an attempt to estimate the c.o.v. of the pedestrian density approaching a footbridge has been made. In analogy to wind engineering, where statistics on the incoming wind are complemented with the description of the flow around the obstacle to provide the aerodynamic forces on the structure, the obtained result is not conclusive. In pedestrian dynamics, the density along the footbridge cannot be confused with the incoming one if the pedestrian dynamics is affected by the walkway features (side barriers, walkway geometry, obstacles). On the other hand, differently from fluid dynamics, pedestrian motion is a multiscale phenomenon, which can be described both at the macroscopic scale (continuous medium) and at the microscopic scale (granular medium). The pedestrian density described in Part I can be ascribed to the former. Conversely, the uncertain location (in space and time) of each pedestrian at the entrance of the footbridge must be sampled and treated at the microscopic scale. Such uncertainty on positions further propagates spanwise along the walkway, being transported by pedestrians themselves. Evidences of this are available in literature based either on in-situ observations or on lab experiments. In this part of the study, the span-wise propagation of the uncertainty generated by the incoming pedestrian density is studied in the framework of the Monte Carlo method. In particular, a statistical analysis of repeated microscopic simulations of a first order crowd model accounting for body size is performed. Inlet conditions are prescribed on the basis of the statistics of the crowd density (described in part I), by assigning probability laws to the pedestrian chord-wise inlet positions and to the inter-arrival times. The final goal of the model is to estimate the propagation of the incoming variability along the footbridge. The application of the model to an ideal footbridge is provided.