This is an old revision of the document!

Homestake Meeting Minutes, 11/17/17

Attending: Pat, Victor, Daniel, Andrew, Vuk, Ross

Agenda:

• Next Meeting: Dec 15
• Data analysis topics
  • Data corrections, frame production (Gary/Daniel/Pat)
  • Event catalog (Gary/Ross)
  • Radiometer updates (Pat, Andrew)
    • slides
  • Parameter estimation for Rayleigh and Love (Pat)
    • slides
    • Pat presented his slides describing the calculation he does to estimate the rayleigh eigenfunctions.
    • Victor: should incorporate a prior on a's forcing them to be positive (i.e. different from zero). Could use simulations from a range of models to get an idea of what values a's can take.
    • Victor: a4 seems to be close to zero, seems unphysical. So should think about the priors…
    • Pat: 2000 seems odd, the phase varies a lot.
    • Vuk: check if different stations at 2000 are consistent with each other
    • Pat: would like to add EQ events, more blast data. Expanding the catalog of events would be very useful.
    • Ross: Gary completed computing particle motions for events in the catalog, but no additional events - i.e. catalog only includes 2015 events.
    • Vuk: Should follow up with Gary.
    • Ross: it should take a week or two to get 2016 events.
    • Daniel: Could use IRIS for teleseismic.
    • Vuk: Is this enough for a paper?
    • Victor: this would be a nice confirmation of what was believed for how these eigenfunctions look like. Maybe noone would be surprised by this result, but should be able to stand on its own.
    • Daniel: could use this to estimate the velocity and compare to velocity estimates by other means. But this could be a rabbit hole, so maybe leave the result focused on eigenfunctons is better.
    • Victor: Bulletin of Seism. Soc. of America would be the appropriate place for this paper. BSSA allows any length, so it would be up to us how we do it.
    • Vuk: Should we start writing?
    • Victor: Yes. Could add Love waves, take care of priors. Probably don't need more events.
    • Vuk: Who would take the lead on this?
    • Victor: could write some parts, not the whole thing.
    • Ross: probably won't have much time.
    • Daniel: will be happy to help.
    • Victor: Pat should be first author, Tanner should be high too
    • Vuk:
  • Teleseismic events and Wiener filtering (Daniel, Michael)
  • Very-long-periods (Ross)
  • Velocity from noise correlations (Daniel)
    • Model selection?
  • Pull together all velocity measurements to start making comparisons (all)

• Papers
  • Homestake Array Summary paper:
    • Leads: Vuk, Gary, Victor
    • Submitted to Seism. Research Lett.

• Active excitation overview paper
  • Leads: Gary, Ross
  • Others: Victor, Daniel,…
  • Draft of the paper exists, as a report for IRIS. Gary will be adopting it to a publication. Don't have a particular date in mind.

• Surface Active Excitation
  • Leads: Victor, Boise
  • Others: Gary, Ross
  • Plan to invite the Boise group to give us a presentation in Sep or Oct 2017, when they are ready with their paper.

• Underground Active Excitation:
  • Lead: Gary, Ross

• Parameter estimation for the Rayleigh and Love eigenfunctions (amplitude vs depth dependence).
  • Leads: Pat, Daniel
  • Also contributing: Victor, Vuk, Levi
  • Novel, but maybe insufficient for a publication. Could use the eigenfunction measurement for Rayleigh and assume something about Rayleigh field to make preliminary estimates of the Newtonian noise.

• Estimating directional and modal composition of the seismic noise. (Pat, Levi, Vuk)
  • Leads: Pat, Vuk, Levi

• Wiener filter application
  • Leads: Jan, Michael
  • Vuk:

• Teleseismic EQs and near-surface scattering
  • Leads: Daniel, Victor
  • Daniel: had some results on this, Micheal helped with wiener filtering. Didn't do much in 6 months, but would like to get back to this and push it forward. Were able to predict the waveform that matched the Homestake data well. Could compare to other data, from other arrays. There is some literature on single transfer functions, from one station to another. This approach offers using multiple stations to make predictions, which would be entirely new in the literature.
  • Victor: related to a more general effort of using large arrays of stations, and model a wavefield as it travels across the array. Part of this is the idea to predict the waveform at one station given other stations.
  • Pat: machine learning could be interesting here, going beyond linear Wiener filtering.
  • Pat to send an email to Michael, Rich and others, to get this started. If it takes off, we can discuss more at the next meeting.

• Time-variable velocity changes
  • Lead: Victor, not started yet

• Very low frequency harmonics (Ross)