I went to look and see if the changes we did to the Picket fence, like using the PNSN and NEIC networks has impacted our ability to see earthquakes before they arrive to the site and the results are promising with the small sample I grabbed. So, once again thanks to the folks at NEIC and PNSN for helping set up the lower latency connections.

I went to the summary pages and looked for times that the SEI_ENV state changed to 'Earthquake' to make this list, I presume those are the earthquakes we would've liked to act on anyways and later we can do a more in-depth analysis. There are 6 events between Oct 26th and 30th and they are all listed in the powerpoint attached.

Main (preliminary) takeaways:

- When the events come from far away (and we have a SEISMON alert), it seems like the earthquke mode is triggered at 400 nm/s of peakmon ground motion. There are 4/6 examples where the picket fence would provide a warning time longer than 25 seconds. The exact numbers are 25, 63, 102, and 137 seconds of extra time if we leave the trigger at 400 nm/s. This is in line with the main purpose of the picket fence: To provide accurate estimates for teleseismic events.

- The picket fence would have given a 25 second forewarning for the EQ from northern BC mentioned in LHO aLog 73786. This indicates that the PNSN seems to be working with the low latency we expect.

- There is one 'outlier' events, a very small EQ that did not trigger the picket fence, but no lockloss happened [see slide 5].

- There is also a local EQ mentioned in LHO aLog 73756.  The local event triggers EQ mode when peakmon nears 600 nm/s, and it shows on the picket fence with enough amplitude about 7 seconds later. We still need more data to see how effective the pickets can be at warning about local events. This would be a nice bonus if we could use it effectively.