And the comms to the unmanned missions are mostly for forensics (outside of active science or other active payload functions) -- if a problem arises that the automation on board cannot solve it is highly unlikely that a difference between real-time and an hour later warning would matter.
It's one thing to have to wait 16 minutes to grab telemetry and issue commands to something on Mars, but having only a few seconds lag seems like something mission designers/operators would take full advantage of. The idea of actively controlling a rover rather than having it perform pre-programmed routines seems highly desirable.
Rover, probably (which I considered payload functions). For delivery vehicle during passive parts of the trajectory the value is doubtful.
Plus, the standard way of getting a ground station outside a country's land footprint is to put one on a ship. That's what Soviet Union did and I suspect this is what China is doing now.
Never heard of a ship based ground station, and that's saying something as I'm in the industry. How does that work? I'd imagine the antenna would need to be on a massive gimbal, yeah?
In principle this isn't that different from ships designed to track ICBMs, and I've seen designs both with spinning parabolic radar antennas in radomes and regular antennas on gimbals. In fact many of those tracking ships were used in space programmes, but I'm not certain, whether just for one way telemetry or two-way comms
My guess is that they would need a third location for unbroken 360 degree coverage. I would gues somewhere in Africa would do it. They have lots of development projects ther.