[cmsmith]

>A single interrogator can cover some 40 kilometers of fiber, Biondi says, and monitor a virtual sensor every couple of meters.

Earthquake engineer here. This is the cool part. We've got plenty of data to detect and locate any given earthquake occurred. And we have maps that show approximately how hard the soil is under any given site. But the intensity that any building sees is still only known to +-50%, mostly due to variation in the subsurface conditions that we don't have enough resolution to model.

[evincarofautumn]

Would higher resolution detectors also help us develop better models for prediction?

[dmix]

We're not currently able to predict the actual earthquakes at all, only probabilities of likelihood on very large timescales. I could see it being useful to predict future damages to certain buildings/areas based on existing data, assuming an earthquake happened.