Today there have been a good number of G stations active on 6m WSPR allowing a really good opportunity to explore GDX paths out to 134km. It seems that if stations are active out to close to 200km (and probably further) then even a few watts of WSPR (polarisation probably unimportant?) will be decoded at the far end of the path eventually, depending on aircraft reflections and favorable Doppler.
This probably means JT65 or JT9 ranges on 6m probably extend to around 200km at least if the stations know the most favourable alignment of aircraft. The same probably applies on 10m, 4m and 2m ,as well as 70cms and microwave bands.
What I'm not (yet) clear about is the normal tropo range without aircraft reflections. For true tropo, polarisation is more likely to be important: one does not want to throw away valuable dBs with cross-polarisation losses. I don't think this matters where signals are randomly scattered off aircraft bodies.