Magnitude change in UC Berkeley Earthquake Catalog

January 23rd, 1996

January 23, 1996

The UC Berkeley Seismological Laboratory recently recomputed the station adjustments used to calculate local magnitude. Using a data set of over 1000 events, Dr. Robert Uhrhammer has estimated new station adjustments for the stations of the Berkeley Digital Seismic Network. These adjustments have been applied to all synthesized Wood-Anderson readings and new magnitude estimates generated.

As of January 23, 1996, the magnitudes in the UC Berkeley catalog computed with synthesized Wood-Anderson readings reflect these new station adjustments for events from 1988-present.

Historical Context

Richter (1935, 1958) introduced the concept of an instrumental earthquake magnitude scale and defined "local magnitude" (ML) in the following way:

ML = log A - log A0 + dML

where A is the maximum trace amplitude on the photographical recorded Wood-Anderson (WA) torsion seismograph, A0 is a distance-dependent attenuation factor, and dML is a station-dependent magnitude adjustment. The main use of ML has been to provide a simple and quantitative measure of earthquake size.

Over the years, the Berkeley Seismological Laboratory has operated Wood-Anderson seismographs at a number of sites in northern California. Station adjustments were determined for each site in order to maintain continuity of magnitude. From 1962 until 1992, UC Berkeley operated Wood-Anderson torsion seismometers at ARC (Arcata, California), BKS (Berkeley, California), MHC (Mt. Hamilton, California), and MIN (Mineral, California) and utilized the photographic records to determine earthquake magnitude.

Beginning in 1990, the Berkeley Seismological Laboratory initiated a major upgrade of their instrumentation. The Berkeley Digital Seismic Network is a modern array of broadband seismometers and strong motion instruments with 24-bit digital data recording. In addition to the advantages of digital recording over photographic records, these instruments record ground motion over a broader range of the seismic spectrum than the Wood Anderson seismographs.

The Berkeley Seismological Laboratory collocated BDSN instruments with the Wood-Anderson seismographs in order to calibrate the effect of the new instrumentation on local magnitude. For a period of time, the Berkeley Seismological Laboratory archived "synthetic" Wood-Anderson (WAS) amplitude readings from its broadband seismometers in addition to the traditional Wood-Andersion amplitude readings taken from photographic records. In order to synthesize the Wood Anderson response, we have employed the calibration of Uhrhammer and Collins (1990). The collocation of the digital equipment with the original Wood Anderson instruments has allowed us to compare and calibrate the photographic and synthesized values.

On January 15, 1993, the Wood Anderson torsion seismographs at ARC, BKS, and MHC were turned off, followed by the instruments at MIN one year later.

Computation of New Station Adjustments

Our primary concern is to maintain continuity in the determination of ML despite the evolution from analog to digital instrumentation. A related goal is to retain the simplicity of Richter's original methodology. Here, we briefly describe the proceedures used to determine the revised station adjustments (Uhrhammer et al., 1996).

The first step in the analysis was to verify the historical values of dML that were used in the routine determination of ML at UC Berkeley. Analysis was preformed on readings from the photographic records for earthquakes which were recorded at all four station with a minimum amplitude of 0.3mm. In order to avoid biasing ML, the sum of the revised dML was constrained to be constant (+0.2) - the value of the historical adjustments.

Local Magnitude Station Adjustments
Comparision of historical and revised
Station Location dML (Historical) dML (Revised)
ARC Arcata, CA +0.2 +0.209 +/- 0.028
BKS Berkeley, CA +0.0 -0.035 +/- 0.017
MHC Mt. Hamilton, CA +0.1 +0.128 +/- 0.018
MIN Mineral, CA +0.1 -0.107 +/- 0.026

To the nearest tenth of a magnitude unit, the two values of dML are not significantly different, confirming the validity of the historical adjustments.

In the next step, we compared the maximum trace amplitudes of the Wood Anderson photographic records with the synthesized recordings from the BDSN. No significant or systematic variations were observed, and the values of dML in the table above were then adopted for the BDSN stations at these sites.

The dML for the rest of the BDSN stations was then determined using approximately 1000 earthquakes which occurred between 1992 and 1995.

Local Magnitude Station Adjustments
Newly adopted values
Station Location dML
ARC Arcata, CA +0.209 +/- 0.028
BKS Berkeley, CA -0.035 +/- 0.017
CMB Colombia College, CA +0.240 +/- 0.009
HOPS Hopland Field Station, CA +0.324 +/- 0.032
JRSC Jasper Ridge, CA +0.139 +/- 0.016
MHC Mt. Hamilton, CA +0.128 +/- 0.018
MIN Mineral, CA -0.107 +/- 0.026
ORV Oroville, CA +0.428 +/- 0.011
PKD1 Parkfield, CA -0.198 +/- 0.013
SAO San Andreas Geophysical Observatory, CA +0.314 +/- 0.009
STAN Stanford, CA -0.233 +/- 0.017
WDC Whiskeytown Dam, CA +0.484 +/- 0.012
YBH Yreka Blue Hill, CA +0.499 +/- 0.015

These adjustments vary significantly from one station to another and generally reflect the local geology. For example, the stations in the Sierra generally have very large, positive adjustments.

As new stations are added to the BDSN, they are utilized in the ML estimation without a station adjustment. After a period of 6 months to 1 year, we estimate a station adjustment and recomputed the magnitudes. At the present, KCC is the only station without an adjustment.

Change in UC Berkeley Catalogue Magnitudes

As of January 23, 1996, the new station adjustments have been used to recompute all of the "network average" estimates of local magnitude reported in the UC Berkeley catalog, and to recompute the "traditional" local magnitudes based on synthetic Wood Anderson records. These new station adjustments will also be used in all future local magnitude calculations.

The following magnitudes are reported in the UC Berkeley catalog:

"Traditional" ML magnitudes:

  1. For events prior to January 17, 1993, only WA amplitude readings are used for determining "traditional" ML magnitude. There has been no change to these magnitude values.
  2. For events from January 17, 1993 to the present, WAS records from the four original Wood Anderson stations are used for determining "traditional" ML magnitude using the new station adjustment values.

"Network Average" ML magnitudes:

  1. For all events from May 13, 1988 to the present, the "network average" magnitude for all events has been recomputed from the average of all BDSN stations which provided WAS readings using the new station adjustment values.

In general, these adjustments have not significantly changed the reported magnitudes, although those events with a limited number of observations are the most sensitive. To demonstrate the effect of the recalculation, this figure compares the new and old "network average" magnitude for all events affected by the change in station adjustments. The top figure shows the ML difference (new - old) is plotted against the new ML and the bottom figure is a histogram of the ML differences. The horizontal lineations in the top figure are primarily due to the inclusion of small earthquakes where the ML was determined by a single station and the location of the line reflects the difference in the new and old station ML adjustments for that station. The histogram is skewed towards positive differences between the new and old station ML adjustments and the individual spikes are also due to the inclusion of small earthquakes which were recorded by only one station.

Because of the effort we have put into calibrating the Wood-Anderson synthetics, we feel that this definition will provide continuity of magnitude in our catalog.

References

Richter, C. F., An Instrumental Earthquake Magnitude Scale, Bull. Seis. Soc. Am., 25,1-32, 1935.

Richter, C. F., Elementary Seismology, Freeman and Co., New York, 768 pp., 1958.

Uhrhammer, R. A., and E. R. Collins, Synthesis of Wood-Anderson Seismograms from Broadband Digital Records, Bull. Seis. Soc. Am., 80, 702-716, 1990.

Uhrhammer, R. A., S. J. Loper, and B. Romanowicz, Determination of Local Magnitude using BDSN Broadband Records, Bull. Seis. Soc. Am., submitted, 1996.

Additional Information

If you have further questions about this change in magnitude calculation, please send email to: info@seismo.berkeley.edu

Change in NCSN Earthquake Catalog

August 1st, 1995

August 1st, 1995

The USGS Northern California Seismic Network (NCSN) has implemented a new policy regarding the archival of automatically timed (RTP) earthquake data in their catalog at the NCEDC. The RTP is an automated system that automatically picks seismic phases, and is used for rapid earthquake response.

Formerly, all locations that were derived from only RTP data were reviewed by analysts, but otherwise no restrictions were applied to the RTP data archived at the NCEDC.

As of August 1, 1995, RTP earthquake data will no longer be reviewed by analysts. To minimize the archival of poor quality RTP information, earthquake locations that are derived soley from RTP phase readings meet ALL of the following criteria to be included in the NCSN catalog:

  • The location must be determined from at least 5 P-readings.
  • The location must have an RMS < 0.5 seconds.
  • The location must have an azimuthal gap < 235 degrees.
  • The location must fall within the following polygon:

36:41.08 -117:47.58
37:45.00 -118:15.00
37:48.30 -120:09.30
42:18.00 -121:30.00
42:18.00 -125:30.00
40:00.00 -125:30.00
34:41.67 -121:22.78

This change may affect the completeness of the catalog at the lower magnitude levels. There has been NO change in the policy or procedures for including or reviewing events that acquired by the CUSP data acquisition system. Events aqcuired by the CUSP data acquisition system will continue to be reviewed by analysts.

Questions about the NCSN catalog located at the NCEDC can be directed to Lind Gee at the USGS, lgee@usgs.gov

Magnitude change in UC Berkeley Earthquake Catalog

February 8th, 1995

February 8, 1995

The UC Berkeley Seismological Laboratory recently discovered a bug in the program used to compute earthquake magnitudes in the UCB catalog. Under certain conditions some station corrections were not always applied to the magnitude estimate. This problem existed for events in the UCB catalog from 1984 through 1995.

As of February 8, 1995, we have corrected this bug, and have re-computed the magnitudes for all events in the 1984-1995 catalogs. The average bias in the magnitude estimate due to this bug was 0.1 magnitude units.

However, at the time we corrected this bug, we decided to changed the definition of the magnitudes provided in the catalog in order to more clearly define the type of information used to calculate the magnitude.

For the last 5 years, the Seismological Laboratory has archived "synthetic" Wood-Anderson (WAS) amplitude readings from its broadband seismometers in addition to the traditional Wood-Andersion (WA) amplitude readings taken from photographic records. Prior to February 8, 1995, both WA and WAS readings for the stations ARC, BKS, MHC, and MIN were used to compute local magnitude (Ml).

In revising the definition of magnitudes in our catalog, we are attempting to provide a clear-cut break between the use of the synthetic records and the traditional readings. We are now distinguishing between these two types of amplitude readings.

Ml magnitude calculations (effective Feb 8, 1995):

  1. For events prior to 1/16/93, only WA amplitude readings are used for determining Ml magnitude.
  2. For all events after that 1/16/93, only WAS records are used for determining Ml magnitude. In both cases, the Ml is computed from the four original stations.

"Network Average" Ml calculations:

  1. The magnitude referred to as "network average" is the local magnitude computed from the average of all BDSN stations which provided WAS readings. This definition has not changed.

Because of the effort we have put into calibrating the Wood-Anderson synthetics, we feel that this definition will provide continuity of magnitude in our catalog.

This change in the definition of magnitude means that there is considerable scatter in the comparison of old and revised numbers. The scatter is more pronounced for smaller events (M < 3.0).

If you have further questions about this change in magnitude calculation, please send email to: info@seismo.berkeley.edu


Contact us: ncedcinfo@ncedc.org
NCEDC blog (News and Info)
Site Map | Privacy
Copyright © 2016 UC Regents