Good point. I'd love someone to explain this.
I think what it pertains to is uncertainties in the measurement of the orbit of the asteroid. Asteroids are very, very hard to see in the sky, and sometimes only a handful of observations are available to extrapolate the path of the asteroid. And I think that there can be several different orbital paths - maybe 75? - that would fit those limited set of observations.
But the orbital path that fits all the existing observations of the asteroid’s position, AND which includes the surface of Mars, allows them to get a bead on when and where the asteroid would hit if it were to do so.
I presume that the "impact zone" is the half of the planet facing where the asteroid is coming from, at the time the asteroid approaches the planet.
So I then presume they mean the rover will be on the back (far) side of the planet when the asteroid will either a) hit the planet, or b) pass by it.
It's relatively easy to predict the date and time the asteroid will be at the planet, and the rotational position of the planet at that date and time. But the actual trajectory of the asteroid (i.e. whether it will hit the planet) is a LOT tougher to predict.
The rover could be on the other side of the planet at the time of impact.