When an arcing fault occurs downstream from two 100A fuses in parallel, can the fuses modelled by a single 100A fuse by shifting the current axis to the right by a factor of two to represent the time-current curve for the two parallel 100A fuses ?
The above assumes that the current divides equally between the two 100A parallel fuses.

Yes you can do this, as long as your assumption holds (no difference in cable lengths, same source, etc.). I don't know which software you use, but properly modeled it may already account for the division. The arcing fault is calculated at a bus so it's not immediately obvious that the time is determined with half of it.

When plotting the TCC a 2x multiplier for a single curve will help visualize the device's actual response to an arcing fault. This also makes it easy to visually check if your calculation is correct.