I recently ran some simulations with initial concentrations of air @ 878,900 ppm and contaminant gas @ 100 ppm.  Obviously these don't add up to 1,000,000 (I forgot that I deleted three other gases from the simulation, and did not update the initial volume fraction of air).  After solving, I noticed that flow simulation changed the initial concentration of air to 999,886.23 ppm and contaminant gas to 113.77 ppm.  Curious, I did a little math and figured out that flow simulation automatically adjusted the unallocated volume fraction for me:

878,900 ppm air / 100 ppm gas = 8,789 (the ratio of air:gas in the tank)

1,000,000 - 878,900 - 100 = 121,000 ppm (the unallocated volume of fluid in the tank)

121,000 / 8,789 = 13.77 ppm (the fraction of the unallocated volume that was assigned to the contaminant gas) --> 100 + 13.77 = 113.77 ppm contaminant gas

121,000 - 13.77 = 120,986.23 ppm (the fraction of the unallocated volume that was assigned to the air) --> 878,900 + 120,986.23 = 999,886.23 ppm air

To see what happens when the total volume fraction adds up to more than 1,000,000 ppm, I ran a similar simulation, but this time intentionally set the initial concentration of air = 1,000,000 ppm, then added three contaminant gases that totaled 600 ppm.  As expected, flow simulation adjusted the initial concentration of air = 1,000,000 - 600 = 999,400 ppm.

The takeaway message here is be careful when specifying your initial concentrations, but if you make a mistake it's better to err on the high side