Berkeley Digital Seismic Network

Introduction

The Berkeley Digital Seismic Network (BDSN) is a regional network of very broadband and strong motion seismic stations spanning northern California and linked to UC Berkeley through continuous telemetry (Figure 4.1 and Table 4.1). This network is designed to monitor regional seismic activity at the magnitude 3+ level as well as to provide high quality data for research projects in regional and global broadband seismology.

The network upgrade and expansion initiated in 1991 has continued, and it has grown from the original 3 broadband stations installed in 1986-87 (BKS, SAO, MHC) to 23 stations in 2002. One station was added in the past year (HUMO).

We take particular pride in high quality installations, which involves often lengthy searches for appropriate sites away from sources of low-frequency noise as well as continuous improvements in installation procedures and careful monitoring of noise conditions at existing stations.

Future expansion of our network is contingent on the availability of funding and coordination with other institutions for the development of a denser state-of-the-art strong motion/broadband seismic network and joint earthquake notification system in this seismically hazardous region.

Figure 4.1: Map illustrating the distribution of operational (filled squares), planned (open squares), and closed (grey squares) BDSN stations in northern and central California.

BDSN Overview

Twenty-one of the BDSN sites are equipped with 3 component broadband seismometers and strong-motion accelerometers, and a 24-bit digital data acquisition system or data logger. Two additional sites (RFSB and SCCB) consist of a strong-motion accelerometer and a 24-bit digital data logger. Data from all BDSN stations are transmitted to UC Berkeley using continuous telemetry. In order to insure against data loss during utility disruptions, each site has a 3-day supply of battery power and is accessible via a dialup phone line. The combination of high-dynamic range sensors and digital data loggers ensures that the BDSN has the capability to record the full range of earthquake motion for source and structure studies. Table 4.2 lists the instrumentation at each site.

Most BDSN stations have Streckeisen three-component broadband sensors (Wielandt and Streckeisen, 1982; Wielandt and Steim, 1986). Guralp CMG-3T downhole broadband sensors contributed by LLNL are deployed in post-hole installations at BRIB and FARB. The strong-motion instruments are Kinemetrics FBA-23 or FBA-ES-T with $\pm$ 2 g dynamic range. The recording systems at all sites are either Q730, Q680, Q980 or Q4120 Quanterra data loggers, with 3, 6, 8, or 9 channel systems. The Quanterra data loggers employ FIR filters to extract data streams at a variety of sampling rates and these have been implemented as acausal filters in the BDSN. In general, the BDSN stations record continuous data at .01, 0.1, 1.0, and 20.0 samples per second and triggered data at either 80 or 100 samples per second using the Murdock, Hutt, and Halbert event detection algorithm (Murdock and Hutt, 1983) (Table 4.3). In addition to the 6-channels of seismic data, signals from thermometers and barometers are recorded at nearly every site (Figure 4.2).

Figure 4.2: Schematic diagram showing the flow of data from the sensors through the data loggers to the central acquisition facilities of the BSL. This particular example shows an STS-1, an FBA-23, thermometer, barometer, and a GPS receiver.

In parallel with the upgrade of the broadband network, a grant from the CalREN (California Research and Education Network) Foundation enabled the BSL to convert data telemetry from analog leased lines to digital frame-relay connections. The frame-relay network uses digital phone circuits that can support 56 Kbit/s to 1.5 Mbit/s throughput. Since frame-relay is a packet-switched network, a site may use a single physical circuit to communicate with multiple remote sites through the use of "permanent virtual circuits". Frame Relay Access Devices (FRADs), which replace modems in a frame-relay network, can simultaneously support multiple interfaces such as RS-232 async ports, synchronous V.35 ports, and ethernet connections. In practical terms, the upgrade to frame relay communication provides faster data telemetry between the remote sites and the BSL, remote console control of the data loggers, additional services such as ftp and telnet to the data loggers, data transmission to multiple sites, and the ability to communicate and transmit data from multiple instruments such as GPS receivers and/or multiple data loggers at a single site. Today, 20 of the BDSN sites use frame-relay telemetry for all or part of their communications system.

Table 4.1: Stations of the Berkeley Digital Seismic Network. Each BDSN station is listed with its station code, network id, location, operational dates, and site description. The latitude and longitude (in degrees) are given in the WGS84 reference frame and the elevation (in meters) is relative to the WGS84 reference ellipsoid. The elevation is either the elevation of the pier (for stations sited on the surface or in mining drifts) or the elevation of the well head (for stations sited in boreholes). The overburden is given in meters. The date indicates either the upgrade or installation time.

Code	Net	Latitude	Longitude	Elev (m)	Over (m)	Date	Location
ARC	BK	40.87772	-124.07738	30.1	0	1992/05 - 2001/07	HSU, Arcata
BDM	BK	37.95397	-121.86554	219.8	34.7	1998/11 -	Black Diamond Mines, Antioch
BKS	BK	37.87622	-122.23558	243.9	25.6	1988/01 -	Byerly Vault, Berkeley
BRIB	BK	37.91886	-122.15179	219.7	2.5	1995/06 -	Briones Reservation, Orinda
BRK	BK	37.87352	-122.26099	49.4	2.7	1994/03 -	Haviland Hall, Berkeley
CMB	BK	38.03455	-120.38651	697.0	2	1986/10 -	Columbia College, Columbia
CVS	BK	38.34526	-122.45840	295.1	23.2	1997/10 -	Carmenet Vineyard, Sonoma
FARB	BK	37.69782	-123.00110	-18.5	0	1997/03 -	Farallon Island
HOPS	BK	38.99349	-123.07234	299.1	3	1994/10 -	Hopland Field Stat., Hopland
HUMO	BK	42.60710	-122.95669	554.9	TBD	2002/06 -	Hull Mountain, Oregon
JCC	BK	40.81745	-124.02955	27.2	0	2001/04 -	Jacoby Creek
JRSC	BK	37.40373	-122.23868	70.5	0	1994/07 -	Jasper Ridge, Stanford
KCC	BK	37.32363	-119.31870	888.1	87.3	1995/11 -	Kaiser Creek
MHC	BK	37.34164	-121.64257	1250.4	0	1987/10 -	Lick Obs., Mt. Hamilton
MOD	BK	41.90246	-120.30295	1554.5	5	1999/10 -	Modoc Plateau
ORV	BK	39.55451	-121.50036	334.7	0	1992/07 -	Oroville
PKD	BK	35.94517	-120.54160	583.0	3	1996/08 -	Bear Valley Ranch, Parkfield
PKD1	BK	35.88940	-120.42611	431.6	0	1991/10 - 2000/09	Haliburton House, Parkfield
POTR	BK	38.20263	-121.93535	20.0	6.5	1998/02 -	Potrero Hill, Fairfield
RFSB	BK	37.91608	-122.33607	-26.7	0	2001/02 -	RFS, Richmond
SAO	BK	36.76403	-121.44722	317.2	3	1988/01 -	San Andreas Obs., Hollister
SCCB	BK	37.28773	-121.86584	101.6	0	2000/04 -	SCC Comm., Santa Clara
WDC	BK	40.57988	-122.54113	268.3	75	1992/07 -	Whiskeytown
WENL	BK	37.62211	-121.75697	138.9	30.3	1997/06 -	Wente Vineyards, Livermore
YBH	BK	41.73204	-122.71039	1059.7	60.4	1993/07 -	Yreka Blue Horn Mine, Yreka

Table 4.2: Instrumentation of the BDSN as of 06/30/2002. Every BDSN station consists of collocated broadband and strong-motion sensors, with the exception of PKD1, RFSB and SCCB which are strong-motion only, with a 24-bit Quanterra data logger and GPS timing. Additional columns indicate the installation of a thermometer/barometer package (T/B), collocated GPS receiver as part of the BARD network (GPS), and additional equipment (Other) such as warpless baseplates or electromagnetic sensors (EM). The main and alternate telemetry paths are summarized for each station. FR - frame relay circuit, R - radio, Mi - microwave, POTS - plain old telephone line, NSN - USGS NSN satellite link. An entry like R-Mi-FR indicates multiple telemetry links, in this case, radio to microwave to frame relay.

Code	Broadband	Strong-motion	Data logger	T/B	GPS	Other	Telemetry	Dial-up link
ARC	STS-2	FBA-23	Q980				FR	X
BDM	STS-2	FBA-23	Q4120	X			FR
BKS	STS-1	FBA-23	Q980	X		Baseplates	FR	X
BRIB	CMG-3T	FBA-23	Q980		X	Vol. Strain	FR	X
BRK	STS-2	FBA-23	Q680				POTS
CMB	STS-1	FBA-23	Q980	X	X	Baseplates	FR/NSN	X
CVS	STS-2	FBA-23	Q4120	X			FR
FARB	CMG-3T	FBA-23	Q4120	X	X		R-FR/R
HOPS	STS-1	FBA-23	Q980	X	X	Baseplates	FR	X
HUMO	STS-2	FBA-ES-T	Q4120	X			NSN	X
JCC	STS-2	FBA-23	Q980	X			FR	X
JRSC	STS-2	FBA-23	Q680				FR	X
KCC	STS-1	FBA-23	Q980	X		Baseplates	R-Mi-FR	X
MHC	STS-1	FBA-23	Q980	X	X		FR	X
MOD	STS-1	FBA-ES-T	Q980	X	X	Baseplates	NSN	X
ORV	STS-1	FBA-23	Q980	X	X	Baseplates	FR	X
PKD	STS-2	FBA-23	Q980	X	X	EM	R-FR	X
PKD1		FBA-23	Q980				FR	X
POTR	STS-2	FBA-ES-T	Q4120	X	X		FR	X
RFSB		FBA-ES-T	Q730				FR
SAO	STS-1	FBA-23	Q980	X	X	Baseplates, EM	FR/NSN	X
SCCB		FBA-ES-T	Q730		X		FR
WDC	STS-2	FBA-23	Q980	X			FR/NSN	X
WENL	STS-2	FBA-23	Q4120	X			FR
YBH	STS-1	FBA-23	Q980	X	X	Baseplates	FR	X

As described in Chapter 11, data from the BDSN are acquired centrally at the BSL. These data are used in the Rapid Earthquake Data Integration System as well as in routine earthquake analysis (Chapter 12). As part of routine quality control (Chapter 11), power spectral density analyses are performed weekly and Figure 4.3 shows a summary of the results for 2001-2002. The occurrence of a significant teleseism also provides the opportunity to review station health and calibration and Figure 4.4 displays the response of the BDSN to a $M_{w}$ 7.3 earthquake in Russia-China border region.

BDSN data are archived at the Northern California Earthquake Data Center and this is described in detail in Chapter 13.

Figure 4.3: PSD noise analysis for BDSN stations, by channel, in the period range from 32-128 sec. PKD stands out in terms of its high noise level variation, which was caused by a problem in the sensor. FARB, sited on the Farallon Islands, stands out as the station with the highest average background noise level. BRIB, sited in a shallow borehole on a hillside prone to seasonal tilting, is also relatively noisy. YBH, sited in a remote and abandoned hard rock mining drift, stands out as exceptionally quiet site. The newest BDSN station, JCC, is also among the quietest of the BDSN stations.

Table 4.3: Typical data streams acquired at BDSN stations, with channel name, sampling rate, sampling mode, and the FIR filter type. C indicates continuous; T triggered; Ac acausal. The LL and BL strong-motion channels are not transmitted over the continuous telemetry but are available on the Quanterra disk system if needed.

Sensor	Channel	Rate (sps)	Mode	FIR
Broadband	UH?	0.01	C	Ac
Broadband	VH?	0.1	C	Ac
Broadband	LH?	1.0	C	Ac
Broadband	BH?	20.0	C	Ac
Broadband	HH?	80.0/100.0	T	Ac
Strong-motion	LL?	1.0	C	Ac
Strong-motion	BL?	20.0	C	Ac
Strong-motion	HL?	80.0/100.0	C	Ac
Thermometer	LKS	1.0	C	Ac
Barometer	LDS	1.0	C	Ac

Figure 4.4: BDSN Z-component broadband recording of the P, pP and sP recorded waveforms from a large deep focus teleseism the occurred in the Russia-northeast China border region ($M_{w}$ 7.3; 2002.179,17:19; depth 566 km; 75 deg NW of Berkeley). The waveforms have been bandpass filtered (0.01-0.1 Hz), deconvolved to absolute ground displacement, and ordered by distance from the epicenter. The similarity of the waveforms in the BDSN broadband records implies that the vertical components are responding normally and that the instrument responses are correct.

2001-2002 Activities

Station Maintenance

Given the remoteness of the off-campus stations, BDSN data acquisition equipment and systems have been designed, configured, and installed for both cost effectiveness and reliability. As a result, the need for regular station visits has been reduced. Most station visits are necessitated by some catastrophic failure. The 2001-2002 fiscal year was no exception.

Vandalism

At station JCC, vandals broke into the steel and concrete vault, cut the instrument power cables, and stole the station batteries. Data acquisition was interrupted while the batteries were replaced and multiple steel locking mechanisms were installed to prevent future vandalism. Local police were notified.

Local police were also notified when vandals broke into the instrument enclosure and stole the GPS receiver at SAO. Engineers from the BSL rebuilt the enclosure building and instrument mounting in a manner to minimize the likelihood of future problems, although the proximity of this site to areas of public access remains a concern.

BDM Power system

We experienced problems at BDM (Black Diamond Mine) this year, with the data logger and telemetry subsystem rebooting periodically. The cause of the reboots was found to be the interaction among a power supply of marginal size, a low voltage cut-off device, and battery with an internally shorted cell. When the line power to the data acquisition system was turned off, the system batteries would power the system until the battery voltage was sufficiently reduced to cause the low voltage device to turn off all power to the system. When the line power was restored, the power supply was inadequately sized to raise the voltage of the internally shorted battery high enough that the system could reboot while pulling the battery voltage lower to the point of the low voltage device turning the system off again. The cycle would thus repeat indefinitely. BSL engineers replaced all the station batteries and installed a larger power supply.

PKD

At station PKD, data are telemetered via spread spectrum radios from the vault to Carr Hill, where a frame-relay connection carries the data to Berkeley. During the past year, as additional groups have added experiments and additional radios, the radio throughput became unacceptably low. After multiple trips and collaboration with USGS and other technicians, BSL engineers were able to improve the data rate by raising the height of the antenna and changing polarization of the transceiver antennas. However, data throughput could diminish further if additional transmitters are added near Parkfield.

Also during 2001-2002, the STS-2 sensor at PKD was replaced. Routine PSD analysis indicated an increase in the noise level.

YBH Upgrades

Station YBH was previously chosen as an alternative monitoring station by both IMS and DTRA. In collaboration with the IMS, BSL installed a VSAT data link, long-period microbaragraph, and separate battery back-up. We also installed an IMS-supplied, stand-alone data validation computer and door switch. To observe long periods on the microbargraph, the sensor was installed in a copper pipe anchored to the rock wall of the adit. Foam insulation was applied to the outside of the copper pipe to further dampen the temperature fluctuations of the sensor. Additionally, a second pier was constructed at YBH for installation of an STS-2 seismometer in 2002-2003. The STS-2 will be deployed in parallel with the existing STS-1 seismometers, bringing YBH into compliance as an auxillary station of the IMS.

CMB Trailer

After many years of service, the portable trailer at CMB was retired. This structure housed the data logger and other electronics. Prior to its deployment at CMB, the trailer was in service at the Jamestown station for nearly twenty years. The combination of the structural failure, the pooling of rainwater on the roof, and the collapse of the floor led to the purchase and installation of a steel shipping container in 2001-2002. Special attention was paid to insulating, and grounding the the container prior to installation of the data acquisition and telemetry equipment.

STS-1 hinges

In November of 2001, Bob Uhrhammer reported observations of 1-sided steps on the STS-1 North component at station BKS. In January and early February of 2002, BSL staff replaced the electronics box and tested the baseplates, and concluded that rust on the sensor hinge was the source of the noise. Small rust spots were observed on another STS-1 sensor (both sensors had not been evacuated in their early history).

Since replacement hinges are not available from Streckheisen - and since as many as 20 BDSN sensors could develop this problem, BSL staff began efforts to manufacture replacement hinges. During a visit in February, Erhard Wielandt recommended replacing all 4 hinges simultaneously, using material similar to the original. BSL staff spent time developing a reproducible recipe for the hinges, including laser cutting the edges for smoothness. The first set of replacement hinges is being tested now.

New Installations

Hull Mountain, Oregon (HUMO)

In the fall of 2000, we began a search for a site to extend BDSN north of the California/Oregon border, as part of a collaboration with the USGS National Seismic Network and the Global Seismic Network of IRIS, to be located north of the midpoint between the existing sites at MOD and YBH. This station will be one ofthe sites of the 100 station NSN/GSN "backbone" in the U.S. Our experience shows that sites which are located on hard granite have less ground tilt and are quieter in the long period, so the first step in locating the new site was to consult a geological map of the area.

In the Medford, Oregon area, we contacted officials at six ranger offices within two National Forests, and the Bureau of Land Management in hopes of finding an abandoned mining adit similar to YBH and MOD. This area was historically an active mining region with many possible sites. Several of the possible adit sites were flooded, several more were located but had collapsed since mining ending, and others were deemed too far from utilities. A suitable adit was located north of Medford, 12 kilometers west of Shady Cove, and this site was permitted with the US Bureau of Land Management. By placing the seismographic equipment on public lands as opposed to private lands, the longevity of the site is assured.

Although instruments located underground clearly exhibit better signal to noise characteristics, technical challenges to securing the site, power, telemetry, and external clock present themselves as a result of this type of installation.

As found, the front of the adit was largely blocked by a small landslide. The entrance was excavated and a steel culvert placed to line the weathered rock near the entrance, prevent collapse, and to provide an anchor point for a lockable grate steel grate. Commercial power and phone lines were trenched and placed below the steel culvert. At this location, at the request of the BLM biologist, a solid door was not used in order that bats may enter and hibernate during winter months. All instrumentation -data logger, seismometers, battery back up, barometer, thermistor and telemetry equipment, is located 110 meters inside of the adit. To achieve external GPS clock signal, a high gain antenna was placed outside the adit. Low attenuation coaxial cable connects the antenna and the data logger clock.

At this time, connection to the National Seismic Network VSAT is being established. Because the equipment is underground and the site is located within a mature forest, it was necessary to locate the VSAT dish approximately 300 meters away from the data logger in a location with a view of the southern sky (where the NSN satellite is located). To achieve digital telemetry over this distance, a fiber optic link will connect the data logger with the VSAT hardware. Again, all data and power lines will be trenched and buried.

While the VSAT system is being completed, the station is accessible via a dial-up phone line. Preliminary data analysis indicates that the background noise level at HUMO is lower than that observed at WDC. Figure 4.5 shows raw P waveforms at HUMO and WDC from a recent teleseism.

In this collaborative effort, the USGS/NSN provided the STS-2 seismometer, the BSL supplied the Episensors and the Quanterra data logger, and IRIS provided support for the installation expenses.

Figure 4.5: A comparison of the raw P-wave Z-component waveform recorded by the STS-2 broadband seismometer sited at the newest BDSN station at Hull Mountain, OR (HUMO) with the corresponding waveform recorded by the STS-2 broadband seismometer sited at Whiskeytown Dam, CA (WDC) 228 km S of HUMO. The P-waves were generated by a large deep-focus teleseism that occurred in the Russia-northeast China border region ($M_{w}$ 7.3; 2002.179,17:19; depth 566 km; 72 degrees NW of HUMO). Both stations are sited in remote abandoned hard rock mining drifts. The highly similar first $\sim$15 seconds of the P-wave waveforms shows that the two stations are responding nearly identically to the teleseismic signal and the differences later in the P-wave coda are due primarily to differences in the crustal structure at the two sites.

New Site Development

In the past year, two new sites north of the San Francisco Bay were located and permitted for development as broadband observatories. In Lake County, the BSL will build an obsevatory on the property of the Homestake Mining Company's McLaughlin Mine. Further north, the BSL will build an observatory at the Alder Springs Conservation Camp in the Mendocino National Forest.

McLaughlin Mine

The McLaughlin Mine site is on property owned and formerly operated by Homestake Mining Company as a surface gold mine. Although local geological maps indicate the area to be volcanic in origin, the mining operation revealed the geology to be extremely varied and complex. With the conclusion of mining operations, the property will be managed as a UC-Davis reserve for research. The seismographic vault will be one of the first research projects on the new reserve. The site will be located approximately 20 kilometers east of the town of Lower Lake, California. The area where the vault will be constructed consists largely of Franciscan sandstone. At this time, a contractor has been hired to excavate and construct a vault from a steel container similiar to those found at stations JCC, PKD, and HOPS. The BSL has registered the name HOME for this site.

Alder Springs

Located approximately 35 kilometers west of the central valley town of Williams, at the Alder Springs site a short period observatory is operated by the California Department of Water Resources. Rocks are mostly serpentine in nature. Again, a seismographic vault similar to those at JCC, PKD, and HOPS will be built. The BSL vault will house the Department of Water Resources equipment presently installed in a fiberglass enclosure. This site has been named GASB by the BSL.

Acknowledgements

Under Barbara Romanowicz's general supervision, Lind Gee and Doug Neuhauser oversee the BDSN data acquisition operations and Bill Karavas is head of the field engineering team. John Friday, Dave Rapkin, Cathy Thomas, and Bob Uhrhammer contribute to the operation of the BDSN. Bill Karavas, Bob Uhrhammer, and Lind Gee contributed to the preparation of this chapter.

Support for the installation of HUMO was provided by the USGS/NSN and IRIS. The California Governor's Office of Emergency Services provided funding toward the development of sites HOME and GASB as part of the CISN.

References

Murdock, J., and C. Hutt, A new event detector designed for the Seismic Research Observatories, USGS Open-File-Report 83-0785, 39 pp., 1983.

Wielandt, E., and J. Steim, A digital very broad band seismograph, Ann. Geophys., 4, 227-232, 1986.

Wielandt, E., and G. Streckeisen, The leaf spring seismometer: design and performance, Bull. Seis. Soc. Am., 72, 2349-2367, 1982.

Berkeley Seismological Laboratory
215 McCone Hall, UC Berkeley, Berkeley, CA 94 720-4760
Questions or comments? Send e-mail: www@seismo.berkeley.edu
© 2002, The Regents of the University of California.