Electronic Supplement to
Detection of Seismic Signals Using Seismometers and Strainmeters

by Andrew J. Barbour and Duncan Carr Agnew

Seismic Noise, Statistical Variation, and LSM Signal Detection

There are two tables here. The first gives station coordinates for the Plate Boundary Observatory (PBO) stations and broadband seismometers in southern California (which we use in this study). The second table provides seismic noise levels for these stations during 2009.

There are seven figures here. The first is a map of stations. The second shows noise levels for borehole strainmeters (used to form Figure 1A in the main article). The third shows variation in borehole seismic noise. The fourth is similar to the second, except it shows noise levels for surface broadband seismometers (Figure 1B/C in the main article). The fifth is similar to the third, except is shows noise variation for surface broadband seismometers. The sixth shows relative signal-to-noise ratios for the seismometers we analyze. The comparisons are between the surface broadband seismometers in the Anza network and the other seismometer types. The seventh is a figure similar to Figure 2 in the main article, except the detection curve is based on the noise model for the longbase laser strainmeters (LSM); the curve in this figure does not extend above 0.5 Hz, the Nyquist frequency of LSM data.


Tables

Table S1. Station coordinates. Here we give stations' coordinates (sorted by latitude) and their surface elevation (in meters). Note that "PFOI" is the IDA-IRIS broadband seismometer; whereas, "PFO" is the Anza broadband seismometer.

Table S2. Seismic noise for all stations, channels, and frequencies. Here we give noise levels for the seismometers used in this study. Column 1 is the station name (as in Table S1). Column 2, "Ch", is the seismometer channel: 1,2,N,E are horizontal channels and Z is vertical. The PBO stations have not necessary been oriented to true North; hence, the horizontal channels are labeled 1 and 2. Columns 4 - 12 are noise levels for data in 2009, in decibels relative to acceleration [ (m/s^2)^2/Hz ], for the following empirical cumulative probabilities: 0.025, 0.05, 0.1, 0.25, 0.5, 0.75, 0.9, 0.95, and 0.975. The 0.5 probability represents the median, or typical noise levels.

Figures

Figure S1. Map of stations. Here we show the stations used in this study, major fault traces (black lines), major roads (gray lines), regions with alluvium coverage (filled gray regions).

Figure S2. Borehole seismometer seismic noise. Here we show seismic noise levels (medians) above 10 mHz, for borehole seismometers. Filled regions show variation across all instruments, whereas thick red lines show levels for individual channels.

Figure S3. Borehole seismometer seismic noise variation. Here we show the variation of borehole seismic noise (pdfs) relative to median values. The contoured plots show power variation relative to median levels. Specifically, shaded contours are 50%, 80%, and 90% variation, and dashed lines are 95% variation.

Figure S4. Surface broadband seismometer seismic noise. Here we show seismic noise levels (medians) above 1 mHz, for surface broadband seismometers; this is similar to Figure S2.

Figure S5. Surface broadband seismometer seismic noise variation. Here we show the variation of surface broadband seismic noise (pdfs) relative to median values; the contouring scheme is the same as for Figure S3.

Figure S6. Relative signal-to-noise ratios for seismometers. Seismometer performance as a function of frequency, in units of decibels, for the different seismometer types analyzed here. Comparisons shown are for short-period borehole geophones of the PBO network (thick lines), and surface very-broadband (VBB) instruments of the IDA-IRIS network (thin lines); each comparison is relative to median noise levels for the vertical components of surface broadband (BB) instruments in the Anza network. Dashed lines show comparisons using vertical components, and solid lines show comparisons using horizontal components.

Figure S7. LSM signal detection. Here we show unity signal-detection curves [D(f)==1] using longbase laser strainmeter (LSM) noise levels. The gray region highlights where a seismometer will have a greater sensitivity to seismic waves, and the black dotted-line shows the composite dispersion spectrum observed in the earth (see main article). The upper frequency limit of the curves is at the Nyquist frequency of the LSM data (0.5 Hz).

[ Back ]