Stability of Local and Regional Coda: Application to the Wells, Nevada Sequence

Kevin Mayeda


For this year, we have studied the near-regional Lg-coda from the $M_w$ 6.0 Wells, Nevada earthquake of February 21, 2008 as well as a local $M_w$ 4.2 event from the San Francisco Bay area that exhibited a clear case of source directivty. The Wells event sequence was unique because the U.S. Array broadband seismic deployment had recorded this event along with its aftershocks, thus providing excellent station coverage and redundancy. Figure 2.26 below shows that the local to near-regional coda is very stable in comparison to the direct Lg. We have formed amplitude ratios, in this case 3.5 Hz, between the mainshock and a selected aftershock. The coda velocity and envelope shape functions were found to be virtually identical from station to station, in spite of significant geologic variation for paths traveling east versus those to the north and north-west. Though this event did not exhibit strong directivity (D. Dreger, pers. comm., 2009), we still see a radiation pattern in the direct Lg ratios, or perhaps random variation between the two sources. However, for the coda, we see significantly less variation, confirming that the coda is not sensitive to the source mechanism.

Figure 2.26: Left figure above shows source ratios for Lg at 3.5 Hz for 12 stations surrounding the Wells mainshock region, roughly 200 km in epicentral distance. We observe large variations even though path and site response are removed by forming the source amplitude ratio. In sharp contrast, the right figure shows the coda-derived source ratios for the same event pair and frequency. As found in other studies, the coda is significantly less variable.
\begin{figure}\centering\epsfig{file=Mayeda_fig1_Direct_vs_Coda.eps, width=8.5cm}\end{figure}

Figure 2.27: Left figure above shows spectral ratios taken between the Wells mainshock and a co-located aftershock. Solid lines represent coda ratios at the 12 stations and colored dots are the direct Lg source ratios. All stations are roughly 200 km in distance. The figure on the right shows the average and standard deviation for both coda and Lg. For the coda, red is the mean and blue is the +/- 1 standard deviation and for Lg, green is the mean and light blue is the +/- 1 standard deviation. See Mayeda et al. (2007) for methodology details.
\begin{figure*}\centering\epsfig{file=Mayeda_fig2_ratios.eps, width=14cm}\end{figure*}

For our second example, we studied an $M_w$ 4.2 event that occurred along the Hayward fault in the San Francisco Bay area on July 20, 2007. This event was very interesting because there were seismic stations equidistant from the mainshock along the strike of the Hayward fault. The mainshock exhibited over a factor of 10 larger amplitude in the north-westerly direction (along strike) relative to the south-east direction. Figure 2.28 shows that station CVS to the north-west has much larger direct wave amplitudes than station MHC to the south-east; however, the coda envelopes are exactly the same after a few tens of seconds. In addition, an aftershock for the same stations and frequency bands does not show any directivity.


We have documented the coda's property of insensitivity to both the source radiation pattern and directivity. The examples shown in Figures 2.26, 2.27, and 2.28 are strong evidence that the coda's averaging properties also applies to the source heterogeneity, not just path heterogeneity. The recent 2008 $M_w$ 5.8 Wells, Nevada earthquake was well recorded by the U.S. Array, and we show that the local-to-near-regional coda is virtually insensitive to the source radiation pattern and directivity effects. In addition, we demonstrate that the coda wavefield becomes homogenized a few tens of seconds past the expanding direct-wave front.

Figure 2.28: We show a clear case of source directivity (left column) that results in roughly a factor or 10 larger amplitude for the direct waves at station CVS, located to the north of the event and along strike relative to station MHC located to the south. However, the coda levels at both stations are identical. An aftershock shows no evidence of directivity (right column).
\begin{figure}\centering\epsfig{file=Mayeda_fig3_Montclair_Envelopes.eps, width=8.5cm}\end{figure}


Mayeda, K., L. Malagnini, W.R. Walter, A new spectral ratio method using narrow band coda envelopes: Evidence for non-self-similarity in the Hector Mine sequence, Geophys. Res. Lett., doi:10.1029/2007GL030041, 2007.

Berkeley Seismological Laboratory
215 McCone Hall, UC Berkeley, Berkeley, CA 94720-4760
Questions or comments? Send e-mail:
© 2007, The Regents of the University of California