S Cooper, J Warner, S Dwyer, 

As part of the work we've been doing to try and understand how large earthquakes affect the interferometer, I've written MATLAB script that searches for the lock state of the interferometer as well as searching for watchdog trips in HEPI, the ISI and the suspension to try and narrow down the number of events that we need to do in depth analysis of. 

The attached images show the interferometer lock loss as a function of earthquake velocity across O1 and O2.There is 11 bins ranging from 400nm/s to 2000nm/s peak velocity, using the 0.03-0.1mHz ground BLRMS channel. To account for the different arrival times of different waves during an earthquake I added a start buffer of an hour to the data search to avoid more events going in the 'previous lock loss' bin. 

 From this we find that work done during the O1-O2 break has enabled H1 to remain locked in earthquakes up to 1.5 um/s peak velocity, whereas during O1 the interferometer locked in only two earthquakes at 400nm/s peak velocity, obviously the number of datapoints is quite small, so these numbers may change with more earthquakes.