Does this mean that the scaling factor of the velocity should correspond to the physical time-stepping between the flow shots? And second, when calculating the particle age, should the velocity scaling factor be added back?
The particles are just moved using the vector (interpolated on the particle location) from one timestep to the next, so that really depends on your units, which ParaView has no knowledge of.
when calculating the particle age, should the velocity scaling factor be added back?
I dont think so, age should depend on time only, not on velocity.