Routine analysis of the data produced by BSL networks begins as the waveforms are acquired by computers at UC Berkeley, and ranges from automatic processing for earthquake response to analyst review for earthquake catalogs and quality control.
Over the last 12 years, the BSL has invested in the development of the hardware and software necessary for an automated earthquake notification system (Gee et al., 1996; 2003a). The Rapid Earthquake Data Integration (REDI) project is a research program at the BSL for the rapid determination of earthquake parameters with three major objectives: to provide near real-time locations and magnitudes of northern and central California earthquakes, to provide estimates of the rupture characteristics and the distribution of ground shaking following significant earthquakes, and to develop better tools for the rapid assessment of damage and estimation of loss.
In 1996, the BSL and USGS began collaboration on a joint notification system for northern and central California earthquakes. The current system merges the programs in Menlo Park and Berkeley into a single earthquake notification system, combining data from the NCSN and the BDSN. Today, the joint BSL and USGS system forms the Northern California Earthquake Management Center (NCEMC) of the California Integrated Seismic Network (Chapter 39).
With partial support from the USGS, the BSL is currently embarking on the development and assessment of a system to warn of imminent ground shaking in the seconds after an earthquake has initiated but before strong motions begin at sites that may be damaged (Chapter 12).
The details of the Northern California processing system and the REDI project have been described in previous annual reports. In this section, we describe how the Northern California Earthquake Management Center fits within the CISN system, detail recent developments, and discuss plans for the future development.
Figure 39.3 in Chapter 39 illustrates the NCEMC as part of the the CISN communications ring. The NCEMC is a distributed center, with elements in Berkeley and in Menlo Park. The 35 mile separation between these two centers is in sharp contrast to the Southern California Management Center, where the USGS Pasadena is located across the street from the Caltech Seismological Laboratory. As described in Chapter 39, the CISN partners are connected by a dedicated T1 communications link, with the capability of falling back to the Internet. In addition to the CISN ring, the BSL and the USGS Menlo Park have a second dedicated communications link to provide bandwidth for shipping waveform data and other information between their processing systems.
Figure 45.1 provides more detail on the current system at the NCEMC. At present, two Earthworm-Earlybird systems in Menlo Park feed two ``standard" REDI processing systems at UC Berkeley. One of these systems is the production or paging system; the other is set up as a hot backup. The second system is frequently used to test new software developments before migrating them to the production environment. The Earthworm-Earlybird-REDI systems perform standard detection and location, and estimate , , and , as well as processing ground motion data. The computation of ShakeMaps is also performed on two systems, one in Menlo Park and one in Berkeley, as described below. An additional system performs finite-fault processing and the computation of higher level ShakeMaps.
The dense network and Earthworm-Earlybird processing environment of the NCSN provides rapid and accurate earthquake locations, low magnitude detection thresholds, and first-motion mechanisms for small quakes. The high dynamic range data loggers, digital telemetry, and broadband and strong-motion sensors of the BDSN along with the REDI analysis software provide reliable magnitude determination, moment tensor estimation, peak ground motions, and source rupture characteristics. Robust preliminary hypocenters are available about 25 seconds after the origin time, while preliminary coda magnitudes follow within 2-4 minutes. Estimates of local magnitude are generally available 30-120 seconds later, and other parameters, such as the peak ground acceleration and moment magnitude, follow within 1-4 minutes (Figure 45.2).
Earthquake information from the joint notification system is distributed by pager/cellphone, e-mail, and the WWW. The first two mechanisms ``push" the information to recipients, while the current Web interface requires interested parties to actively seek the information. Consequently, paging and, to a lesser extent, e-mail are the preferred methods for emergency response notification. The recenteqs site has enjoyed enormous popularity since its introduction and provides a valuable resource for information whose bandwidth exceeds the limits of wireless systems and for access to information which is useful not only in the seconds immediately after an earthquake, but in the following hours and days as well.
Figure 45.1 illustrates the current organization of the two systems. As described above, each Earthworm/Earlybird component is tied to a REDI component and the pair form a single ``joint notification system." Although this approach has functioned reasonably well over the last eight years, there are a number of potential problems associated with the separation of critical system elements by 35 miles of San Francisco Bay.
Recognizing this, we are redesigning the Northern California operations so that identical, complete systems operate independently at the USGS and UC Berkeley. In FY01/02, specifications were established and the details required for design were determined. In the interim, however, much of the development effort focused on statewide CISN activities, and specific plans for the ``next generation" Northern California system were put on hold. The enforced wait provided the opportunity for some ideas to mature and the current plans for the NCEMC are somewhat different from those envisioned in 2001.
The current design draws strongly on the experience in Southern California for the development of TriNet (Figure 45.3), with modifications to allow for local differences (such as very different forms of data acquisition and variability in network distribution). In addition, the BSL and the USGS want to minimize use of proprietary software in the system. The TriNet software used three forms of proprietary software: Talerian Smart Sockets (TSS) for inter-module communication via a ``publish and subscribe" method, RogueWave software for database communication, and Oracle as the database management system. As part of the development of the Northern California Earthquake Data Center, the USGS and BSL have worked extensively with Oracle databases and extending this to the real-time system is not viewed as a major issue. However, we did take the opportunity to review options for replacing Smart Sockets and RogueWave with Southern California, resulting in joint agreement on replacement packages and shared development effort.
In the last four years, BSL staff, particularly Pete Lombard, have become extremely familiar with portions of the TriNet software. We have continued to adapt the software for Northern California, making adjustments and modifications along the way. For example, Pete Lombard has adapted the TriNet magnitude module to northern California, where it is running on a test system. Pete made a number of suggestions on how to improve the performance of the magnitude module and has worked closely with Caltech and the USGS/Pasadena on modifications.
The BSL and the USGS Menlo Park have developed and tested a design to exchange ``reduced amplitude timeseries." One of the important innovations of the TriNet software development is the concept of continuous processing (Kanamori et al., 1999). Waveform data are constantly processed to produce Wood Anderson synthetic amplitudes and peak ground motions. A program called rad produces a reduced timeseries, sampled every 5 secs, and stores it in a memory area called an ``Amplitude Data Area" or ADA. Other modules can access the ADA to retrieve amplitudes to calculate magnitude and ShakeMaps as needed. The BSL and the USGS Menlo Park have collaborated to establish the tools for the ADA-based exchange. As part of the software development in northern California, a number of modules have been developed.
During 2005-2006, progress has continued toward the retirement of CUSP - the system used by the USGS Menlo Park to time earthquakes. CUSP was initially developed in Southern California during the late 1970s - early 1980s and has been used for a number of years in Northern California. However, the CUSP system is becoming increasingly outdated.
The NCEMC has implemented a plan to retire CUSP, using some components of the Southern California system. The primary responsibility for the necessary programming and development rest on the shoulders of BSL staff. They have implemented the RequestCardGenerator (a module that decides which channels to archive, given a particular earthquake), a waveform archiving module, and iiggle (the earthquake timing interface). The NCEMC and SCMC collaborated on modifications to jiggle for use in Northern California such as the computation of . The test system is operating, and USGS timers have begun to assess jiggle.
Also during the past year, Northern and Southern California developers spent a day in Pasadena discussing issues of joint interest.
The REDI system has routinely produced automatic estimates of moment magnitude () for many years. However, wary of complications caused by the publication of multiple magnitudes, these estimates were not routinely used as the ``official" magnitude until after the 05/14/2002 Gilroy earthquake ( 4.9, 5.1).
In a past annual report, we discussed the question of when to report . As currently implemented, solutions that meet a minimum quality criterion are automatically reported (a variance reduction of 40% or higher). This criterion appears to work very well and screens out events contaminated by teleseisms. Over the last few years, nearly all events over 4.5 have met this criterion, as have a number of events in the M3.5-4.5 range. As part of the effort to establish a statewide magnitude reporting hierarchy, we have looked more closely at the estimates of (Gee et al., 2003b; 2004) and the comparison between and .
Two methods of determining regional moment tensor (RMT) solutions are part of the REDI system - the complete waveform modeling technique (CW) of Dreger and Romanowicz (1994) and the surface wave inversion (SW) of Romanowicz et al. (1993). In the past year processing for the SW algorithm was discontinued, however CW moment tensors continue to be calculated, reviewed and reported. Comparison between the results of the CW method and other regional moment tensor studies in northern California and the western United States show excellent agreement in the estimate of seismic moment and . Over 128 events, the average difference in is 0.002 magnitude units.
As we transition toward statewide reporting of earthquake information, a comparison of magnitudes calculated for southern and northern becomes important. We have collected a set of events recorded well by digital broadband and and strong motion stations of the Northern California (NC), Berkeley (BK) and Southern California (CI) networks and are assessing the computation of local magnitude for each station.
In fiscal year 2005-2006, more than 30,000 earthquakes were detected and located by the automatic systems in northern California. This compares with over 38,800 in 2004-2005, and 12,000 in 2003-2004. Many of the large number of events in 2004-2005 are aftershocks of the 2004 Parkfield earthquakes. The number of events continues to remain high, because we are now receiving data from a network of seismometers in the Geysers, a region with a high level of small magnitude seismicity. Of the more than 30,000 events, over 170 had preliminary magnitudes greater than 3. Fourteen events had greater than 4. The largest event recorded by the system occurred on 12 May 2006 with 4.6. This earthquake near the Geysers, California was actually two events within 70 seconds of each other.
As described in the 2003-2004 Annual Report, the BSL staff are no longer reading BDSN records for local and regional earthquakes (as of March 2004). This decision was in part intended to reduce duplication of effort between Berkeley and Menlo Park.
The BSL continues to focus on the unique contributions that can be made from the broadband network. From July 2005 through June 2006, BSL analysts reviewed nearly 30 earthquakes in northern California and adjoining areas of magnitude 3.5 and higher. Reviewed moment tensor solutions were obtained for 11 events (through 6/30/2006). Figure 45.4 and Table 45.1 display the earthquakes located in the BSL catalog and the moment tensor solutions.
The density and distribution of broadband seismic stations located in Northern and Central California increased during the past year, with two new BK stations and additional broadband seismic stations installed as the USArray transportable array completes its station coverage in the region. One design goal of the transportable network is to complement the existing BDSN broadband stations and cover the region with an average interstation spacing of 70 km. Our motivation for characterizing the seismic background noise PSD level observed at the transportable stations is, in part, that we would like to occupy the best sites after the transportable array moves out of the region, in order to improve the coverage of the BDSN network.
We have been characterizing the vertical and horizontal seismic background noise levels observed in Northern and Central California via Power Spectral Density (PSD) analysis of the seismic background signals recorded by the BDSN broadband seismic stations (BK network), NCSN broadband stations (NC network) and by the USArray broadband seismic stations (TA network). Two frequency bands, the 1-5 Hz short-period (SP) band and the 30-60 second long-period (LP) band, are analyzed to characterize the background noise in the seismic bands of interest, particularly for the study of the seismic signals from local and regional seismic events.
The PSD algorithm uses a statistical approach to robustly estimate the background noise PSD. The PSD estimates are reported in dB relative to 1 (m/s2)2/Hz. The input time series is parsed into eight (possibly overlapping) time series and each of the resulting time series are appropriately windowed prior to calculating their PSD estimates. For short time series, less than 1.5 hours in length, the time series are detrended and sine tapered while for longer time series the dominant semi-diurnal gravitational tide signal is also removed to avoid biasing the long-period PSD estimates. The PSD estimates are smoothed and reported at twenty logarithmically spaced intervals per decade in period.
Owing to the statistical nature of the PSD algorithm, the time series processed must contain at least 65,635 (216) contiguous samples. Shorter time series are not processed and a warning is issued. The PSD algorithm can process data with a wide variety of sampling rates (from sps to sps). A typical usage with broadband data is for the time series to contain one day of continuous LH (1 sps) data (86,400 samples), say. Since the sensor transfer function representation in the SEED data volume for a typical inertial seismometer does not include the static component of the response, the background noise PSD estimates for periods longer than approximately an hour will be biased high; and hence, they will be unreliable.
The PSD code distribution along with examples of its usage are available via the Web at http://seismo.berkeley.edu/algorithms/.
We acquired SEED data volumes containing a network day (2006.022) of 1 Hz three- component broadband data from the broadband stations of the NC network, the TA transportable network and from the BK permanent network. This data was used to characterize the seismic background noise in the long period (LP) band between 30-60 seconds. We also acquired SEED data volumes containing two network hours (2006.022.0000-0200) of 40 Hz three- component broadband data from the same stations of all three networks. (day 2006.022.0000-0200). This data was used to characterize the seismic background noise in the short period (SP) band, between 1-5 Hz. This study was done to assess the broadband NC stations and to help identify TA sites that would be good candidates for upgrading to permanent BDSN sites once TA moves away from Northern California.
The vertical-component LP (30-60 second period) seismic background noise PSD is shown in Figure 45.5A and the corresponding horizontal- component noise is shown in Figure 45.5B. The spread between the quietest and noisiest stations on the vertical-component is 20 dB. Stations of both the BK and TA are relatively quiet. The spread between the quietest and noisiest stations on the horizontal-components in the LP band is 30 dB with a larger percentage of noisy stations. The stations with the lowest noise levels are situated on bedrock or in hard rock mines and are far from cultural noise sources. Stations with high PSD noise levels on the horizontal components may either be on thick alluvium or have high ambient cultural noise levels.
The vertical-component SP (1-5 Hz) seismic background noise PSD is shown in Figure 45.5C and the corresponding horizontal-component noise is shown in Figure 45.5D. Here the spread between the quietest and noisiest stations on the vertical-component is larger than in the LP band, 50 dB. The spread between the quietest and noisiest stations on the horizontal-component is likewise 50 dB. Again, the stations with high horizontal-component PSD noise levels in the SP band are generally either sites with thick alluvial deposits or high ambient cultural noise levels.
Based on this analysis as well as criteria such as ownership, and access to power and telemetry, seven stations have been selected for upgrade to permanent sites within the next 18 months. These are: V04C (RAMR), HAST, O01C and P01C, as well as HATC, HELL and SUTB.
Peggy Hellweg oversees the REDI system and directs the routine analysis. Peter Lombard and Doug Neuhauser contribute to the development of software. Rick McKenzie, Doug Dreger, David Dolenc, Ayhi Kim, Ved Lekic, Mark Panning, Junkee Rhie, Dennise Templeton, and Akiko Toh contribute to the routine analysis. Peggy Hellweg, Doug Neuhauser and Bob Uhrhammer contributed to the writing of this chapter.
Partial support for the development and maintenance of the REDI system is provided by the USGS.
The facilities of the IRIS Data Management System, and specifically the IRIS Data Management Center, were used by Bob Uhrhammer for access to the TA network (USArray) waveform and metadata required in the noise comparison study.
Dreger, D., and B. Romanowicz, Source characteristics of events in the San Francisco Bay region, USGS Open File Report 94-176, 301-309, 1994.
Gee, L., J. Polet, R. Uhrhammer, and K. Hutton, Earthquake Magnitudes in California, Seism. Res. Lett., 75(2), 272, 2004.
Gee, L., D. Neuhauser, D. Dreger, M. Pasyanos, R. Uhrhammer, and B. Romanowicz, The Rapid Earthquake Data Integration Project, Handbook of Earthquake and Engineering Seismology, IASPEI, 1261-1273, 2003a.
Gee, L., D. Dreger, G. Wurman, Y, Gung, B. Uhrhammer, and B. Romanowicz, A Decade of Regional Moment Tensor Analysis at UC Berkeley, Eos Trans. AGU, 84(46) , Fall Meet. Suppl., Abstract S52C-0148, 2003b.
Gee, L., D. Neuhauser, D. Dreger, M. Pasyanos, B. Romanowicz, and R. Uhrhammer, The Rapid Earthquake Data Integration System, Bull. Seis. Soc. Am., 86, 936-945,1996.
Pasyanos, M., D. Dreger, and B. Romanowicz, Toward real-time estimation of regional moment tensors, Bull. Seis. Soc. Am., 86, 1255-1269, 1996.
Romanowicz, B., D. Dreger, M. Pasyanos, and R. Uhrhammer, Monitoring of strain release in central and northern California using broadband data, Geophys. Res. Lett., 20, 1643-1646, 1993.
Berkeley Seismological Laboratory
215 McCone Hall, UC Berkeley, Berkeley, CA 94720-4760
Questions or comments? Send e-mail: firstname.lastname@example.org
© 2006, The Regents of the University of California