Homestake Data Analysis Workshop, 10/23-25, 2015
Agenda:
Vuk's notes
Friday, 10/23
-
Seismic Waves Primer (Gary)
P and S body waves in isotropic media,
Rayleigh and Love waves,
anisotropic media basics (velocity dependence on direction, shear wave splitting, and particle motions in an isotropic medium),
free surface interaction in isotropic and anistropic media.
instrument calibration, try to get it going during the workshop
Composition of the seismic noise: radiometer and other algorithms, understanding the Rayleigh waves (amplitude dependence on depth, P-vs-S contributions) - Vuk, Tanner, Levi
Saturday, 10/24
-
Brief overview from Gary:
slides Gary has some analysis code under development to show. These are tools to visualize and measure three-component particle motions.
Overview of the active source data and any preliminary results (Gary/Ross)
What do we know about velocity (vs depth, vs direction) - this is linked to your active source study
Work on earthquakes (Daniel)
noise cross correlations (Daniel)
Event catalog (Gary)
Extending the life of the array?
papers?
Sunday, 10/25
Wiener filter study - this is pursued by Michael Caughlin and Jan Harms, and neither will be present. Maybe we can ask them to connect remotely -
pdf slides Coupling of solar wind modes into Earth modes (Ross):
pdf slides glitch-finding algorithms
Data Analysis Projects
Solar wind modes (Ross, Gary)
measure lines in the spectra at low frequencies, 40-400 uHz, using multi-taper techniques
could potentially benefit from tilt-meters or magnetometers
depending on the absolute displacements, possibly could do this with GPS stations.
could also try cross-correlations between pairs of detectors
Deep EQ in Japan (Daniel, Victor)
Estimate incident and reflected waves, resolve them as they propagate through the array.
Deep source, does not suffer from surface effects. Can get a cleaner incident signal at deeper stations.
Cross-correlations (Daniel, Victor)
cross-correlation between two stations is related to the Green's function relating the two stations, under the assumptions of isotropic source distribution and equipartition of energy.
extract the dependence of Rayleigh wave speed on frequency, which gives the dependence of shear wave speed on depth.
may be able to measure body waves directly with underground stations?
Could potentially detect the presence of Love waves.
Attempt to compare the noise correlation functions (pseudo-Green's functions) from different periods of time to see if we can resolve temporal or seasonal differences in the structure (i.e.: a difference in fluid content or snowpack weight between summer and winter).
test various pre-processing techniques to maximize signal recovery. This may be of relevance for radiometer signal pre-processing too.
How does this extend to isotropic GW background? Can you measure the speed of gravitational waves? (Vuk, …)
Visualization of the motion (Gary)
identify polarization of the waves. At high frequencies expect deviations from the standard theory due to anisotropy. At low frequencies the standard theory probably fits data better.
Can you apply these visualization techniques to seismic noise in very narrow bands? Do you see the retrograde/prograde motion? Do you see Love waves?
Active source experiments (Ross, Gary)
Measure the speed at different locations on the surface and along the drifts where the measurements were done
try to extract average speeds in other directions/locations using the active sources and the Homestake array instruments. Would need the absolute timing of the shots, which does not exist but could be estimated.
Wiener filter (Michael, Jan)
how well does the subtraction work at different depth, different frequencies, different filter lengths
how do we determine the optimal array design? Test performance of the filtering after dropping some of the stations
Question: is it possible to use the spatial patterns of signal coherence (or some other misfit metric) to map regions of higher scattering or anisotropy, and/or to determine a weighting scheme for a radiometer or beamforming code?
Related to Wiener: how do we measure correlations in an inhomogeneous seismic field? That is, how do we measure/identify local sources, local scattering etc?
Use the deep EQ in Japan to estimate the incident and reflected waves, subtract them from data and the residual should carry information about the local scattering…
Event Catalog (Gary)
Continue to identify and catalog events observed by the Homestake array
correlate them with USGS database
Glitch-finding algorithms (related to the Event Catalog above)
Apply Omicron to the Homestake data (Michael)
look at events on the time-scale of 1 sec, including local (mine) events
Try to apply STAMP to an example EQ other events on the time-scale of 100sec (Tanner, Pat?)
If the results are interesting for geo, potentially consider other applications
Radiometer (Tanner, Pat, Levi, Vuk)
correct the Rayleigh model: double exponential, new speed estimates etc
for speeds: could start with the preliminary estimates from Gary: v_s = 3300-3900 m/s, v_p = 5700 m/s
for speeds: in the long terms would like to use the estimate of v_s(depth), estimated as a power law. This would then give us all exponentials for Rayleigh and Love waves.
modify the code to use larger frequency bands, windowing, etc
develop a tool for Love waves as well (single exponential).
compare to existing directional tools