UNRELEASED PRODUCT — UNDER CONSIDERATION
Last update: 2017-08-01
The horizontal to vertical spectral ratio (HVSR) of a station’s ambient noise provides useful information on its site response characteristics. The proposed product allows HVSR computations using available power spectral density (PSD) data and creates a site response database for the TA stations in Alaska.
The horizontal to vertical spectral ratio (HVSR) of ambient noise could be used to characterize site response of seismic stations. We propose to create a site response database for the USArray TA station in Alaska to provide information on HVSR anomalies due to extreme seasonal temperature variations and/or presence of unconsolidated sediments under some TA stations. The proposed product will be based on the development of two open source Python scripts that could also be used by users to obtain HVSR for from PSDs for other stations. These scripts would:
- compute long-term PSD PDF spectra for given station-channel pairs in the form of median, lower and higher percentiles from the existing PSDs.
- compute HVSR of ambient noise for stations after removing PSDs that are outside the long-term lower and higher percentiles of PSD PDF (#1 above)
The first script:
- utilizes the available hourly PSDs to compute and store long-term PSD PDF characteristics spectra for each station-channel pairs
The second script:
- utilizes the the computed long-term PSD PDF spectra above to eliminate possibly contaminated station’s hourly PSDs to isolate the ambient noise spectra
- group the cleaned hourly PSDs to daily bins and obtain the daily median PSDs
- compute daily HVSR from the daily median PSDs
- compute and save average daily HVSR for selected time windows
Example — site characterization:
- TA stations: TCOL & I23K northwest of Fairbanks, Alaska (Figure 1)
- time window: April 2015
- HVSR type: monthly (Figure 2)
Figure 2. Three-component PSDs for stations TA.I23K (left top 3 panels) and TA.TCOL (right top 3 panels) for 2015-04. Yellow, red and orange curves show the computed long-term (2014-10 to 2015-06) station trends at 95%, 50% and 5%, gray PSDs are outside the long-term trend and green trimmed PSDs are the ambient noise PSDs. The 0.1 to 10 Hz mean and +/- one standard deviation HVSRs for 2014-04 (bottom panels) show 1-5 Hz anomaly at TCOL station site.
Example — seasonal effect:
- TA station: TOLK
- time window: 2011 – 2015
- HVSR type: average monthly
Figure 3. Average monthly HVSR computed for the months with thawed and partially thawed ground (June – November, 2011-2015) for the TA station TOLK (Toolik Lake Research Station) in Alaska showing increase in HVSR in the frequency bands of 0.4 to 1 Hz and 8 to 10 Hz with the highest HVSR levels in July-October. June and November HVSR represent transitional levels with lower HVSR.
Figure 4. Average Monthly HVSR computed for the months with frozen ground (November – May, 2011-2015) of the TA station TOLK (Toolik Lake Research Station) in Alaska. HVSR drops compared to the thawed months (Figure 3) .
PLEASE DO NOT QUOTE OR CITE THIS PAGE. THIS IS A WORK IN PROGRESS.
- DMC Data Products Team
- Created : Posted as Google Document
- Posted : Posted under Data Product Laboratory