1 / 26

PBO Borehole Strainmeter Data in the Period Band 4-16 Days

PBO Borehole Strainmeter Data in the Period Band 4-16 Days. Evelyn Roeloffs, USGS Sept-Oct 2007. Motivation. Investigate relative gains of the 4 gauges Are these frequency-dependent? Are they stable with time? Are similar to the values measured on the bench?

kemp
Télécharger la présentation

PBO Borehole Strainmeter Data in the Period Band 4-16 Days

An Image/Link below is provided (as is) to download presentation Download Policy: Content on the Website is provided to you AS IS for your information and personal use and may not be sold / licensed / shared on other websites without getting consent from its author. Content is provided to you AS IS for your information and personal use only. Download presentation by click this link. While downloading, if for some reason you are not able to download a presentation, the publisher may have deleted the file from their server. During download, if you can't get a presentation, the file might be deleted by the publisher.

E N D

Presentation Transcript

  1. PBO Borehole Strainmeter Data in the Period Band 4-16 Days Evelyn Roeloffs, USGS Sept-Oct 2007

  2. Motivation • Investigate relative gains of the 4 gauges • Are these frequency-dependent? • Are they stable with time? • Are similar to the values measured on the bench? • Investigate strain induced by atmospheric pressure • Is there frequency dependence? • Is there azimuthal dependence?

  3. Features in 4-16 day band • No tidal signal expected • Long-term trends removed • Expect mainly atmospheric pressure unless some other form of strain is being imposed • Atmospheric pressure has equilibrated over the ocean • Distribution of loading should be uniform with frequency in this band

  4. Steps in the analysis, 1 • Select a stretch of data, 60-90 days • 10-minute bottle data for strain, decimated to 30 minutes and linearized • 30-minute barometer bottle data • Should not have long gaps • Interpolate over the short gaps using cubic splines • Detrend by subtracting cubic splines fit to averages over apx 37 days • Maybe not the best way to detrend, but seems to work

  5. Steps in the analysis, 2 • Bandpass into 4-16 day band, and 4 sub-bands: • 4-6 days,7-9 days,10-12 days,13-15 days • I used an 8-pole Butterworth filter, passed through the data in both directions to get zero net phase shift • Filter the 4 strain channels and the atmospheric pressure • For each band plot • Strain versus atmospheric pressure • Timeseries • For each band regress the strain against atmospheric pressure

  6. Subset of stations examined • Analysis is not meaningful for stations with • Too short a record (< apx 100 days) • Data dominated by non-tectonic signals in the subtidal band (e.g., lots of offsets) • Very frequent pumping-induced signals (B001,B006, B082,B089) • Leaves 16 to be analyzed as of early October 2007 • This presentation contains examples of the various types of behavior

  7. Types of behavior in 4-16 day band • Very similar waveforms on all channels most of the time, closely tracks atmospheric pressure • Amplitudes very close among gauges • B004,P403(before July 2007),B028 • Amplitudes obviously different between gauges • B005(big difference), B018, B073 (small difference) • Slightly dissimilar waveforms on all channels most of the time • All coastal stations: B009,B010,B011,B012,B022 • Other stations: B007 and B084 • Very dissimilar waveforms on all channels much of the time • Eg, B035 and B036; B027 (early in record, though)

  8. Example: B004, Hoko Falls • Entire band 4-16 days • 4 channels track pressure • Similar amplitudes all 4 channels • “Consistency sum” relatively small

  9. B004, Hoko Falls: Sub-bands, Apr & May 2007 • Elliptical paths indicate phase lags • At these periods, seems unlikely to be an instrument effect • Relative gains of gauges vary between bands

  10. B004, Hoko Falls: Sub-bands, June & July 2007 • Gain differences not same as Apr&May • Relative amplitudes vary between bands

  11. Example: B005, Shores 1 • Entire band 4-16 days • 4 channels track pressure • Different amplitudes • Consistency sum is large • Suggests differences in amplitudes due to different gauge gains

  12. B005, Shores 1: Sub-bands, Apr & May 2007 • Phase lags seem smaller than at B004

  13. B005, Shores 1: Sub-bands, June & July 2007 • Gain differences similar to Apr&May

  14. Error bars are 2 x standard error of the slope most are smaller than the symbols, possibly errors are underestimated From data shown, time variation and frequency dependence are comparable Need to analyze all the data! B004 & B005: Atmospheric Pressure Response vs Period

  15. Example: B073, Varian 1 • Entire band 4-16 days • 4 channels track pressure • Slightly different amplitudes

  16. B073, Varian 1: Sub-bands, Aug & Sept 2007 • Gains nearly equal, little phase lag or frequency dependence

  17. B073, Varian 1: Sub-bands, June & July 2007 • Gain differences similar to Apr&May

  18. Example: B009, PGC 1 • Entire band 4-16 days • 4 channels track pressure approximately but not in detail • All coastal stations seem to behave this way • Csum is “small”, so probably an actual strain

  19. B009, PGC1: Sub-bands, June & July 2007 • This behavior would seem likely to impede accurate determination of atmospheric loading effects • Are long-period signals in the tide gauge data?

  20. These two sites about 100 m apart behave similarly in that they do not track atmospheric pressure much of the time B035 and B036 Grants Pass OR

  21. B035, Grants Pass1: Sub-bands, Aug-Sep 2007 • There seems to be a strain signal in the 13-15 day band that does not come from the atmospheric pressure • It could be a shear strain signal because CH1 and CH3 are mirror images

  22. B036, Grants Pass2: Sub-bands, Aug-Sep 2007 • In 13-15 day band there is a signal similar to B035 • As at B035, CH1 and CH3 are nearly opposite • Behavior in all bands resembles B035

  23. B035, Grants Pass1: Sub-bands, Feb-April 2007 • All 4 channels track atmospheric pressure well at this time • Barometer now has lots of energy in 13-15 day band, and strain channels track it • No B036 data for this time period

  24. Time series plots show a similar signal for B035 and B036 from about 12 July to 1 August on the EN-shear Both stations have longer-period variations in the data that are not “self-consistent” Probably due to recency of installation at B036 (8 May 2007) B035 and B036, Grants Pass OR

  25. Findings • B004 (and some other strainmeters) exhibit time- and/or frequency dependent atmospheric pressure responses, as well as phase lags of strain behind atmospheric pressure at periods longer than 4 days • Coastal stations have other long-period signals in the data at all times examined • Grants Pass strainmeters seem to be detecting strain of non-atmospheric origin in the 4-16 day period band much of the time

  26. Need to analyze the entire available data sets Try to distinguish between time dependence and frequency dependence Use cross-spectral analysis for longer data sets that appear stationary Need to examine the consistency sums: What is the proper scaling (ie, what is “small” vs “large) Use this to separate azimuthal dependence of atmospheric pressure effects from effect of different gauge gains Determine source of non-atmospheric signals in time series from coastal stations Investigate the data from Grants Pass Automate plots of the regression analysis results Future Work, etc.

More Related