I searched and tried to find the answer, but couldn't. I know that orbits fall back to Keplerian orbits at high time steps so that our space stations don't de-orbit during Mars transits, but is the same true for orbit decay caused by exosphere drag?
From the PDF on dynamics (doc\Technotes\dynamics.pdf)
With a typical frame rate of 50 fps and at a selected time compression
of 1000x , the step interval is 20 s...During normal simulation runs, the main
problem are however isolated long steps, caused for example by disc I/O. During such
steps, the orbit stabilization mode keeps orbits from deteriorating.
This was speaking on how Orbiter interpolates positions and orbits. If the time step is too large (either due to time acceleration or laggy hard drives), Orbiter prevents orbits from decaying.
Therefore, Orbiter must ignore any exosphere drag that occurs at large timesteps.