I don't quite agree with Roger. It looks like the Chile 8.8 had some effect. None of the others look like they did anything to those sensors. I may get this wrong, but M9 quakes have some slight long term effects in things like length of day, so it does not seem impossible that extremely precise instruments in mines might measure something. It would be interesting to see these curves through the Japan M9 March 11, 2011, and after then.
Someone please correct me if I am wrong on some tilt being measurable in a mine if the instrument is precise enough.

But this is a lot different from predicting earthquake. If there is a relation between the Chile 8.8 and these tilts, it has nothing to do with what would have to be many orders of magnitude smaller precursor, which would be much better measured in the epicentral area. (and in many cases there is no precursor measured with local precise instruments, although some quakes may speculatively be triggered by deep creep/slow quakes.

Chris