Interpretation of Seismograms GEOL695

Spring 1998

	After discussions with John Anderson and those who attended the
organizational meeting Jan. 21 I propose the following proceedures and
format for the class.

1.  All seismology students will be required to audit the class and
participate in weekly record reading sessions.

2.  This year we will emphasize retrieving seismograms from internet
sources.  The first few sessions will emphasize proceedures for doing
this, and will feature presentations by Glenn Biasi, Gene Ichinose,
Arturo, and others? who have some experience doing this.  

3.  Starting about mid-February  students will be assigned responsibility
to collect data for the earthquakes of the week, and at least one other
interesting earthquake if there are none ocurring during the week, and
preparing to lead discussion and analysis of the seismograms on the
subsequent Wednesday.  At that time Jim Brune and others will give a half
hour lecture on interpretation of seismograms at the beginning of the

4.  Copies of notes will be made for some of the lectures by Jim Brune.
The notes have a list of references on interpretation of seismograms.
These references will be on file in the Seismology office for students to
check out.

Near-Real-Time Waveform Retrieval Overview and Links

These web sites provide useful information for waveform data shopping. Data can be obtained by either applying for a research Unix shell account or by anonymous web access.

Local Short Period Seismic Networks

Regional Broadband Networks

Strong Motion Networks

Global Networks

Choosing Common Download Parameters

When sitting down to construct a data set, one should consider gathering this information:

Example of retrieving near-real-time teleseismic data

The Standard for the Exchange of Earthquake Data (SEED)

The SEED format is used by many data centers. A SEED file can be very large (several to hundreds of megabytes) and contains waveform data, trigger information, and instrument responses.

Programs and Examples of Teleseismic Data

Here is an example of November 8, 1997 M 7.9 earthquake in Tibet.

Using Local-Regional Pn and Pg Phases for Estimating the Thickness and P-Wave Velocity of the Crust and Mantle

Here we use Pg and Pn arrivals at Caltech-USGS Southern California Seismic Network stations from the Ml 5.8 1995 Ridgecrest earthquake to determine the thickness of the crust in southern California. These stations are very short period (> 1Hz) and high gain, so the waveforms are usually clipped and worthless except for the phase information of the first arrivals (Pg). The timing of Pg is used for automatic network earthquake locations called CUSP (Caltech-USGS Seismic Processing). Geol695 people can try this, the dataset is in rumble:/wgb/geol695/1995_Ridgecrest

This diagram shows how one can estimate the apparent P-wave velocity (Vapp) of the crust (Vapp1) and mantle (Vapp2) by measuring the moveout of the Pn and Pg phases (dt/dx=Vapp) on the X-T plot. The true material velocity is equivalent to the apparent velocity only if the velocity gradient is laterally homogeneous (dVapp/dx=0). The cross over distance (Xcross) can also be measure from an X-T plot. Xcross is the distance where the refracted arrivals from the crust-mantle interface begin to arrive as the Pn phase. The crustal thickness (z) is then estimated from Xcross, Vapp1 and Vapp2 using the above equation with the assumption that we have a horizontally layered earth. If the crust-mantle interface dips away from the epicenter, then Vapp for Pn will be slower than the true velocity. In this case, we are looking at Pg and Pn arrivals at stations south of Ridgecrest, CA (Delta=0km) essentially across the Mojave desert to Baja California (Delta=400km). The greatest density of stations are from the Los Angeles Basin (Delta=200 km). We would need an earthquake in northern Baja California recorded by the same stations to get a reverse shot in order to rule out crust-mantle interface dip.

This is the CUSP data from the Ridgecrest earthquake plotted using a 8 km/sec reduction velocity. Notice the point at distance delta=170km and reduced time (tau=6.5sec). There is a change in Pg moveout. This is where the Pn begins to arrive. Remember tau is reduced time and not actual two way travel time. Before delta=170km is the Pg and beyond 170 km is the Pn phase arrivals. The Pn plots a little faster than 8km/sec because it has a slight negative moveout (the slope dt/dx is negative). It turns out that plotting at a reduction velocity of 8.2km/sec gives a zero slope for Pn and a revised crustal thickness of z=33.5km. Kanamori et al. (19??) was the original one who did this for southern California and found that the crust is usually around z=32 km thick.

The time when Pn arrives appears highly variable. This may be due to arrivals at different azimuths, where the crustal thicknesses varies in southern California, from like fast cold crustal roots under mountain ranges, or slow hot asthenosphere intruding into the lower crust. Caution... The lateral travel path of Pn results in the fact that Pn is a very emmergent phase and difficult to pick accurately relative to Pg. The crust-mantle interface apparently does vary in depth. This is called Moho Topography) and the depth to Moho varies from z=10 in the Salton Trough to > 40 km under the Peninsular Ranges.

Tutorial on Passcal's Quick Look (PQL)

Tutorial on Seismic Analysis Code

| What's New | Resources | Email | Questions | Suggestions | Statistics | Phones | Finger | FTP |

Author: G. A. Ichinose
UNR Seismology Lab Homepage