To supplement Les's posting, here is the behaviour of the reflected Parry as a function of sun elevation. At sun elevations higher than about 11 degrees, the halo arc is below the sun. The halo is perhaps easiest understood by thinking of it as a Parry arc of the subsun. At the solar elevations given in these simulations, it stays all the time at 22 degrees above the subsun. So if you know solar elevation, you know also where it is located relative to the sun.
In simulations for 15 and 20 degrees sun elevation also the normal lower sunvex Parry is seen, being as much below the horizon as the reflected Parry is above the horizon. Simulations are made with Jukka Ruoskanen's software.