Modeling of the Byerly's False S Phase

David Dolenc (U of Minnesota) and Doug Dreger


Byerly's False S phase was first observed more than 70 years ago (Byerly, 1937) for earthquakes off the coast of Northern California. We used the recent dense station coverage in the region (USArray, Northern California Seismic Network, Berkeley Digital Seismic Network, Mendocino Array) to study the False S observations. We identified offshore events that produced the False S phase, relocated them using a double-difference algorithm, and inverted for the seismic moment tensor. With the location and source parameters constrained, we modeled the broadband waveforms using 2D and 3D velocity structures to find the origin of the False S arrivals.

History of False S

Byerly (1937) observed False S for the events located in the Mendocino Triple Junction (MTJ) at the stations to the south in the $2^\circ$-$9^\circ$ distance range. The speed of False S and the fact that the phase died out at distance led Byerly to believe that False S is a P wave that propagated through the sedimentary layer. However, Byerly noted that False S was not observed at Ferndale, the closest station to the MTJ at that time, and concluded that his proposed origin of the False S was open to objection. Cameron (1961) determined that False S is a longitudinal wave that arrives at the observing stations with a small angle of emergence. He also concluded that False S could be a P wave that propagated in the upper sedimentary layer. Oliver and Major (1960) suggested that False S results from the amplitude variation in the leaking-mode propagation within the near-surface wave guide. Maulik (1964) tried to explain the propagation of the phase through the sedimentary layer and concluded that a similar ``false'' phase should follow the true S wave, corresponding to S-wave propagation in the sedimentary layer. However, no such phase was observed. Ghosh (1964) suggested that False S resulted from the Stoneley wave propagating along the inclined continental margin. But Auld et al. (1969) showed that the phase was not observed at an ocean-bottom seismometer and ruled out the Stoneley-wave hypothesis.


Our analysis of the offshore Mendocino events showed that (1) False S is observed for some offshore events along the Mendocino Escarpment (E of $127^\circ$W) and to the north (up to $41^\circ$N), (2) for events that generate False S, the phase is always observed on the near-coast stations north of the MTJ (up to station L02A), (3) events that generate False S always generate False S at the near shore station JCC, (4) False S is never observed at the closest stations KCT and KMPB, (5) strong False S is observed when the S phase is strong, (6) for strong False S events, the phase can be observed to the south, all the way to the San Francisco Bay region, and (7) no ``false'' phase after the true S wave is observed. The results support our starting hypothesis that False S is an SP phase that propagates as an S wave through the crustal layer above the Gorda plate and is refracted as a P wave back to the surface when it hits the thicker part of the subducting plate. This model can explain (1) the time delay between P and False S, (2) why a strong False S is observed when the S phase is strong, (3) why False S is never observed at the closest stations, (4) why the phase dies out at long distances, and (5) why False S phase has a small angle of emergence. We modeled the False S observations by developing a network of 2D structures to explain False S on all the stations leading to 3D modeling. The results showed that the origin of the False S phase could be explained by simple 2D modeling. The finite-difference modeling so far could not reproduce the False S observations. We are further testing the 2D and 3D models using finite-difference code to complete this work.


This project was funded by NSF grant EAR-0745803. We thank R. Allen, E. Humphreys, and A. Levander for sharing the Mendocino Array data with us.


Auld, B. et al., Seismicity off the coast of northern California determined from ocean bottom seismic measurements, Bull. Seism. Soc. Am., 59, 2001-2015, 1969.

Byerly, P., Earthquakes off the coast of northern California, Bull. Seism. Soc. Am., 27, 73-96, 1937.

Cameron, J. B., Earthquakes in the northern California coastal region (Part II), Bull. Seism. Soc. Am., 51, 337-354, 1961.

Ghosh, M. L., On a possible explanation of the PF-phase, Bull. Seism. Soc. Am., 54, 1417-1427, 1964.

Maulik, T. N., On the PF and other phases in the early part of seismogram, Pure and Applied Geophysics, 58, 49-62, 1964.

Oliver, J. and M. Major, Leaking modes and the PL phase, Bull. Seism. Soc. Am., 50, 165-180, 1960.

Zelt, C. A. and R. B. Smith, Seismic traveltime inversion for 2-D crustal velocity structure, Geophys. J. Int., 108, 16-34, 1992.

Figure 2.33: Left: Locations of the offshore events included in the study. Filled black circles are 2005-2008 events for which False S was observed on at least some stations. Open black circles are 2005-2008 events for which False S was not observed. Right: Velocity waveforms for the four numbered events. Shown is the vertical JCC component that had instrument response deconvolved.
\epsfig{file=dolenc10_1_1.eps, width=18cm}\end{center}\end{figure*}

Figure 2.34: Example of 2D ray-tracing (RAYINVR; Zelt and Smith, 1992) used to model travel times and amplitudes. Left: (A) P-wave modeled as a head wave that propagated just below the 5 km interface. We looked for the phase that arrives at the station as a P wave, does not reflect more than once, and arrives with the largest possible delay relative to P. The resulting phase SP is shown on the right. (B) Gorda plate is modeled as a 2-layer slab with a $6^\circ$ dip. The phase that has the largest delay relative to the P-wave and satisfies the above conditions is shown as an SP phase that is reflected at the bottom of the Gorda plate. (C) Gorda plate has non-uniform thickness. S wave propagates through the GIL7 crustal layer above the Gorda plate, and when it hits the thicker part of the Gorda plate, it is refracted back to the surface as a P wave. This type of model can explain the observed time delay between P and False S. Right: Synthetic seismograms for the 3 models are compared to the observations at JCC. Only model C can produce a phase that has travel-time comparable to that of the observed False S.
\epsfig{file=dolenc10_1_2.eps, width=14.5cm}\end{center}\end{figure*}

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