Homestake Meeting Minutes, 07/28/16

Attending: Victor, Vuk, Daniel, Tanner

Agenda:

• Planning the next visit:
  • Constraints from Jaret:
    • Maintenance days: Sep 23,26 and Oct 7,10. The Lab also observes the Native American Day holiday on Oct 10.
    • Tom's availability: Sep 24-26 (weekend), traveling say Fri Sep 23 (there is another group earlier that week and Tom needs to manage his hours). Also, week of Oct 3 looks good.
  • It looks like Sep 23-29 or so should suffice.
  • Travelers: Gary, Pat, Daniel and Vuk (probably not the whole trip).

• Paper plans:
  • Gary/Ross:
    • In the early phases of a data report paper on the practical issues we uncovered in the active source experiments. It will double as a data report required by IRIS. This probably should also include the weight drop experiment. Publication options are a new AGU journal Earth and Space that encourages this kind of paper. Alternatively, Seismological Research Letters recently announced it will consider similar papers. This particular story might be a bit long for the later.
      • Daniel: Should check with the Boise group, so that there is no overlap.
      • Victor: Boise group is working on the surface active excitation, they have many measurements done, Victor's student is helping out. Probably will publish separately.
      • Vuk: Gary should connect with them to determine what goes in which paper.
    • We could consider a similar paper on the 3D array. SRL might be more appropriate for this one as it has a higher impact factor (I think).
    • Other things are more in planning phases. I'm starting to make some progress on this weight drop data recorded underground, but at this point I have a lot of time in front of the computer facing me.
  • Vuk (and the UMN group):
    • A paper describing the array, as Gary suggested. It may be useful to show some data in this paper, such as the seismic noise levels, variation over time, variation with depth, maybe coherence plots. No significant data analysis, just basic things.
      • Victor: SRL is good for papers without very deep science, more speculative things.
    • Parameter estimation for the Rayleigh and Love eigenfunctions (amplitude vs depth dependence). This could be a separate paper, or a part of the above paper.
      • Victor: inversion (using velocity structure) instead of parameter estimation? Tanner: Yes, planning to do both approaches and to compare them. Will try to improve on parameter estimation that Tanner did, and compare it to Victor's approach.
      • Daniel interested in collaborating.
    • Estimating directional and modal composition of the seismic noise. This would include the radiometer method description, some simulations, and applications to real data IF we manage to convince ourselves that the method is working. It may be simplest to try this first at low frequencies (0.1 Hz), maybe compare with other existing techniques. And then work upwards in frequency until things start to break.
    • For an estimate of the modal and directional composition of the seismic noise, compute the estimate of Newtonian gravity fluctuations. Could be done using (semi)analytical approach, or maybe purely numerical simulations.
  • Jan/Michael: Wiener filter application
    • Ran the old wiener filtering code on the new data, started compiling plots and summarizing results. There are ideas about related analyses, eg estimating the body waves by filtering 4850 with surface stations. Time scale, not sure, depends on what new things we want to add.
    • Michael: Jan has code to show data in the k-f space. Ran the code on some of the data, have velocity histograms. Not sure what Jan wants to do with it.
    • Vuk: Will invite Jan to attend the next call and explain. There are overlaps with other efforts on estimating wave speed etc.
  • Daniel and Victor will discuss and add to this list too.
    • Victor: looking at teleseismic EQs and near-surface scattering. Another direction is to look at time-variable velocity changes, but did not start it yet. Either or both could be directions for papers.

• Other comments:
  • From Gary (email):
    • We have this interesting data set that relates to the fundamentals of surface wave propagation. The problem is how to avoid throwing our hands up and saying it is a mess. What have we learned at this point and what has potential? I'm a bit befuddled at the moment as at this point my view is everything I've looked at in these data tend to only uncover more complexity I don't understand. I think we need to focus on defining a key testable hypothesis we can evaluate. Maybe we should discuss that first?
      • Victor: slightly more optimistic, Tanner's results are promising.
      • Vuk: agreed.
      • Tanner: with Rayleigh measurements, looked at the EQ signals but not much information at high frequencies. So limited to large wavelengths, which don't help much with eigenfunctions. For mine blasts (from Wyoming), get the expected results up to 0.4 Hz or so, above this frequency start to see secondary effects, maybe things that Gary is talking about. Could improve the analysis technique, may it would make a difference.
      • Victor: for mine blasts, the waves are likely already affected by the surface by the time they reach Homestake. Might not be any cleaner than the surface data, at least at high frequency. But at lower frequencies we may benefit, i.e. on the scales that are large relative to the scale of drift sizes etc.
    • If I could give one thing I've learned from this experiment so far it is this. Going into this my working hypothesis was that getting away from the very heterogeneous surface weathered layer would simplify the data. This hasn't proven true at all. A lot of the reason is that we aren't that far away from the surface even at 4850 and most of our observations are still colored by (possible) near surface scattering. The underground active source data are the main exception, BUT that data was apparently contaminated by wave interaction with the drift itself due to the frequency content our “airless jackhammer” source that seemed to couple well to modes of the drift. There also may be weird resonances on the free face of the drift - a pure hypothesis. In any case, that data is far from simple either. We thus may be left with a standard geophysical problem that we need to develop a full 3D model of the structure inside the array to make progress. That is a huge task that is not realistic to complete in the time remaining on this project so a more strategic approach is needed. I am presently far from knowing what the best course is.