With high primary and secondary useism the SEI_ENV node has been transitioning to earthquake mode based on higher peakmon (H1:ISI-GND_STS_CS_Z_EQ_PEAK_OUTMON) signals, but not due to an actual earthquake. I've raised this threshold from 600 to 850 for now. It looks like the waves should calm down in about a day so I'll plan to revert this tomorrow.
TJ reverted this threshold earlier today. Around the same time (11:22pst) I made a change to the filter used for the peakmon calibration, that should reduce the confusion between microseism and earthquakes.
First picture show 3 ground spectra, red is a more normal frequency higher microseism time on the 16th, blue is a spectra take when both the primary and secondary microseism were elevated from a large mass of wave activity right off the coast and pink is a largish eq from a couple days ago. On the blue spectra the .1-.3hz peak has spread down to 80 or 90mhz, and was bleeding into the band peakmon uses.
Second image compares the bandpass peakmon was using to the one I installed today. Blue is the new filter, a cheby1 lowpass with a zpk([0],[.015],1) high pass, green is the old total filter. The problem this week with the old filter was the higher primary microseism peak at 50-60 mhz and the broader peak around .1-.3hz extending down to 80mhz. The blue filter should greatly reduce the chance of that secondary peak confusing the SEI guardians.Rise time for this filter is important, and the blue filter has some lower frequency poles than the green, but the foton step response for the filters isn't dramatically different. The new filter takes a couple seconds longer to respond, but that should be easy to tweak.
We've already had one successful earthquake transition about an hour after I changed the filter. Wave forecasts for next week suggest more high microseism starting this next Tues-Wed.