It definitely would work. The distance to the event is irrelevant, it's the light travel time between the detectors compared to the accuracy with which you can pin the event down in time that matters. The light travel time across the Earth is of order a hundredth of a second, which is a significant fraction of an event that takes ~ a tenth of a second.

However, the error ellipse will probably be quite larg, and given that they come from cosmological distances it's unlikely that they would be anything but isotropically distributed (like gamma ray bursts are).