Local Multi-Station Digital Recordings of
by
Open-File Report 82-777
Aftershocks of the January 9th, 1982
New Brunswick Earthquake
E. Cranswick,
C. Mueller,
R. Wetmiller*,
and E. Sembera
This report is preliminary and has not been
reviewed for conformity with
U. S. Geological Survey editorial standards
(and stratigraphic nomenclature).
Any use of trade names is for descriptive purposes only and does
not imply endorsement by the U. S. Geological Survey.
*
Earth Physics Branch, Division of Seismology and Geomagnetism,
Menlo Park, California 94025
1982
1 Observatory Crescent, Ottawa, Ontario, Canada K1A OY3
On January 9, 1982, at 12:53:51.8 GMT, there occurred a magnitude 5.7 mb earthquake in New Brunswick, Canada, which has been given a preliminary location by NEIS of 46.98° North Latitude, 66.66° West Longitude and a depth of 10 kilometers. The earthquake caused light damage in New Brunswick and was felt as far away as southern Connecticut U.S.A. (Schlesinger-Miller, et al., 1982). The largest aftershocks have been a magnitude 5.1 mb aftershock at 16:36 GMT on January 9, and a magnitude 5.4 mb aftershock on January 13.
Responding to a request for instruments from the Earth Physics Branch (EPB) of the Canadian Department of Energy, Mines, and Resources, Ottawa, the U. S. Geological Survey, Menlo Park, California, dispatched nine three-component digital seismographs and two staff members to New Brunswick on January 13, 1982. The American equipment and personnel arrived at Newcastle, New Brunswick on the evening of January 14. The digital instruments were deployed in the epicentral region in cooperation with the Canadian seismological field party commencing on January 15.
This report is a summary of the digital seismograph aftershock survey conducted in the epicentral region from January 15 to January 22, 1982. It includes a brief description of the Geographical Setting of the seismic activity. There is a general description of the field operations with particular emphasis on those aspects which control the reliability of the parameters - station location, transducer orientation, etc. - that define the data collected. The raw results are summarized with the intent of providing other investigators sufficient information to use the digital data for further study. Appendix A is a set of figures that display the contents of the raw data digital tape files. Appendix B consists of figures that display the data after it has been processed digitally in a fashion similar to the standard digital processing that has traditionally been applied to strong motion (SMA) film records. The preface to Appendix B describes the processing done to the New Brunswick digital records. Appendix C is addressed primarily to the engineering seismologists, and it consists of whole record acceleration spectra of some representative events.
The epicentral region is located in the northern central section of New Brunswick. Figure 1 shows a simplified geologic map of the area adapted from a geological map of New Brunswick (Potter, et al., 1979). The epicenter of the January 9 mainshock is plotted as an eight-pointed star to the immediate northwest of Tuadook Lake. The two nearest population centers are Newcastle and Chatham to the east and Plaster Rock to the west. The region has a low population density, and the principal land uses are logging and recreation.
The geologic map shows a pronounced overall northeast-southwest structural trend. Fast-northeast to west-southwest trending faults are prominent in the north of the map. The major structural unit depicted is the body of granitic intrusives of Devonian age which outcrop in the immediate epicentral region. Other outcrops of the same material can be seen to the northeast, and the same unit continues off the map to the southwest. The area of study which is delineated by the rectangle in Figure 1 is underlain by granite in the west and the center, and by metamorphosed Ordovician sediments in the east. These crystalline rocks were glaciated in the Pleistocene, and thus they are fresh and unweathered at the surface. The seismological consequence of this is that the source region is probably relatively free of any heterogeneities and velocity gradients that would be introduced by lithologic boundaries and near surface mineral alteration (Moos, 1982).
The study area rectangle in Figure 1 is shown at larger scale in Figure 2. The four digital seismograph stations that were established are shown as large black triangles, and the epicenters of the forty aftershocks digitally recorded are indicated by stars. The area can be seen to have moderate relief, with the greatest difference in elevation equal to about 250 meters. There is some suggestion of a northwest~southeast topographic grain which is principally defined by the river valleys. This lineament trend is in contrast to the overall northeast~southwest structural grain of the region previously described, and it may reflect preferential erosion along fractures in the homogeneous granite that is otherwise devoid of structure. These hypothetical fractures would possibly be related to states of stress, such as those responsible for the offsets of the Triassic dike on the east side of Figure 1, that post-date the Paleozoic states of stress that produced the regional structures.
The primary impediments to field operations were the extreme winter weather conditions and the limited access to the epicentral region. The only paved road in the region is Route 109 which runs between Newcastle and Plaster Rock shown in Figure 1. Unpaved logging roads and private roads normally only used in summer are the only means of reaching the study area from Route 109. On several days, heavy snowfalls and winds up to 70 kph necessitated repeated snowplowing of more than twenty kilometers of unpaved roads just in order to change records at the stations. One station was installed by transporting both equipment and personnel to the site by snowmobile.
During the study period, temperatures fell to -30° Celsius on at least two occasions. Since this is 15° C below the manufacturer's rated minimum operating temperature of the digital recorders, heated shelter was required for the instruments. The difficulties of limited access were therefore compounded by the necessity of finding buildings both near to the aftershock epicenters and well-distributed in azimuth from them. One station was placed in a private summer camp and two other stations in one-room huts. The seismographs of the fourth station were placed in a van that was left parked at the site for the duration of the station's operation. All four stations were heated by kerosene fueled space heaters. Since the heaters contained a fuel supply sufficient to operate them for a maximum of twenty-four hours, each station had to be visited once a day just to replenish the kerosene.
The study period described in this report began when the first U.S.G.S. digital seismograph station started recording on January 15, and ended on January 22 when the U.S.G.S. digital recorders were removed from the four stations established. The Canadian smoked-drum recorders which were first deployed in the epicentral region several days prior to January 15 were also removed from the field on January 22. Of the nine U.S.G.S. digital seismographs sent to New Brunswick, six were successfully deployed. Because of the relative abundance of digital seismographs compared to the paucity of appropriate station sites, two of the stations were equipped with two separate instrument systems apiece. At the dual instrument stations, one of the digital recorders was connected to a velocity transducer and the other was connected to an acceleration transducer (force balance accelerometer or FBA). The two single instrument stations were equipped with a velocity instrument and an acceleration instrument respectively. Because of the dual instrument stations, and the format requirements of the computer software used to process the digital data, the following station name/instrument code nomenclature has been used. All codes referring to the provenance of digital records consist of three characters; the first character being the letter 'C'. The second character is the station number or letter. The third character denotes the order of motion that is recorded: 'A' for acceleration and 'V' for velocity. If the three-character code refers to a dual instrument station, the third character is 'T' denoting that both acceleration and velocity are recorded there. If the three-character code refers to a specific instrument at a dual-instrument station, it is either 'A' or 'V'. Over the course of the study period, the Canadian party co-located vertical component velocity-transducer smoked-drum recorders at the four digital stations. In order to distinguish instruments, a two letter station code refers to Canadian records:
| Digital | Smoked Drum | |
| C7T | LC | |
| C8T | PC | |
| C9V | HL | |
| CBA | MR |
At all stations, the transducers themselves were placed outside whatever structure housed the recorder. None of the transducers were sited directly on bedrock outcrop. All transducers were sited on or several centimeters below frozen ground surface which had roughly .5 meters of snow cover. The thickness of the Pleistocene alluvial layer separating the transducers from bedrock at tne four stations is estimated to range from less than one meter up to a maximum of several tens of meters. These estimates are based on observations in the field and on bedrock outcrop geologic maps. At the time of this writing there has been no verification of site alluvial-layer thickness. Once installed, the transducers tended to become literally frozen into place. With the exception of the C8A transducer, the transducers became so rigidly frozen to the ground surface that they had to be forcibly pried loose with a shovel in order to be removed.
The digital recorders (described in the next section) were generally operated in the mode of recording three orthogonal components of motion. The three individual component-transducers are packaged together as one unit. The vertical orientations of the transducers were established by means of a bubble level. The azimuths of the horizontal directions were set with a Brunton compass assuming a local magnetic declination of 23° west of North. At the time of station removal , these orientations were rechecked. Vertical orientations appeared to have changed little if at all and are believed to be correct within the radius of error defined by the bull's eye circle of the bubble level. Horizontal azimuths are believed to be true to within less than 5° error of their recorded values. The peculiar azimuth given for transducer C9V is a result of oversight during the haste of installation.
The digital seismographs used in this study were Sprengnether DR-100 self-triggering three channel digital recorders connected to Sprengnether S-6000 velocity transducers and Sprengnether SA-3000 FBA's. Generally the instruments were operated in the mode of serially sampling three components of motion at a sample rate of 200.32 samples/second/component. For one day, January 16-17, the instrument C9V was accidentally operated in the mode of sampling only the vertical component at 600.96 samples/second. Five of the instruments deployed had 5 pole anti-aliasing filters with corner frequencies at 50 Hz. By chance, instrument C9V had a 5 pole anti-aliasing filter with a corner frequency at 70 Hz. The velocity transducers had a natural period of 2 Hz and a damping coefficient of .6 critical. The FBA's had a natural period of 85 Hz and a damping coefficient of .55 critical. The dynamic range of the recorders is ± 2048 digital counts. The internal trigger used an short term average/long term average algorithm to detect events. The instruments had .74 seconds of pre-event memory and were set to record for 5 seconds after trigger. The records produced were generally of about 6 seconds duration. See Table 2 and Table 3 for summaries of the instrument constants. The instruments were powered by a 12 volt car battery. See Fletcher (1982) for further description of the instruments.
Because of the cold operating conditions, and the jolting that accompanied transport to the field area, there was the possibility that the actual instrument constants during the study period differed from their rated values. As a test of the fidelity of the overall system response, the following comparison was made. Since there were two separate instruments at stations C7T and C8T, it was possible to compare separate records of ground motion recorded in the same place at the same time. The top trace in Figure 3 is the N-S (transverse in terms of source-receiver) component of horizontal acceleration recorded by instrument C7A of the S-arrival of the event at 13:32:59, January 17. The middle trace is the top trace integrated to velocity, and the bottom trace is the direct velocity recording made by instrument C7V during the same time period as the traces above. All three traces have been high passed through a zero-phase-shift filter with a corner at 1.6 Hz and have had their sampling augmented in the frequency domain (Cranswick and Spudich, 1982). to produce an apparent sample rate of 1600 samples/second. The middle and the bottom trace are plotted to amplitude normalized scales to facilitate waveform comparison. It can be seen that whereas the two lower traces are not identical in detail, the overall similarity of the waveforms is striking. The corresponding absolute amplitudes for the total six seconds of record of the integrated acceleration and velocity agree to within 10%. This congruence of recorded signals suggests that the overall system response can be trusted to accurately reflect ground motion in a predictable fashion. See the preface to the Processed Seismogram appendix for further discussion of instrument response.
The instruments were subject to two known malfunctions: one analog and one digital. At trigger time, the tape drive motor is turned on, which places a high transient current drain on the system power supply. This in turn introduces a transient bias into the analog electronics which manifests itself as a decaying step in the signal baseline beginning .74 seconds into the resulting record. The initial amplitude of this step is about 2 digital counts and is not generally apparent in the raw recorded time-series. See the appendix for further details. The data also suffered from intermittent tape drop out problems which varied in degree and frequency of occurrence from instrument to instrument, and which were exacerbated by the very cold ambient environment that developed when kerosene space heaters exhausted their fuel. The approximately five records which have the worst drop outs have yet to be fully processed, pending further developments in computer software.
The internal clocks of the DR-100 recorders were initially set and subsequently checked against a portable master clock. The master clock had an ovenheated crystal, a nominal precision of .1 milliseconds, and was calibrated to CHU radio time and the Canadian field master clock each day. Following the conclusion of the study, a least squares line was fit to the daily clock corrections to the masterdock. This line had a slope, the masterdock drift, of 5.001 milliseconds/day with a standard deviation of 1.01 milliseconds.
With the exception of instrument C8A, the drift rate of the internal clocks of the DR-100 recorders was on the order of 10 milliseconds/day. Individual recorder clocks settled down to a constant drift rate after transient behavior lasting about 1 to 2 days, which is presumably related to the time constant of thermal equilibration of the clock crystals with the ambient temperature.
Final clock corrections to the event trigger times were made by first correcting the individual recorder clock corrections for master clock drift. The appropriate instrument clock correction was then derived by linear interpolation between the immediate prior and post calibration points. This procedure is believed to yield corrected event trigger times that are in general accurate to within ± 5 milliseconds of absolute time. Clock corrections are listed in Table 3. Spot comparisons of the two arrival times calculated for events recorded at dual instrument stations show this agreement, though there is some discrepancy, not yet resolved, between the times of instrument LC and those of the digital instruments of C7T.
The usual procedures followed in an aftershock survey were complicated by the Logistics problems described above. Because it was never certain that the instruments could be re-visited within the next 24 hours, a conservative approach was taken to setting instrument gains, trigger sensitivities, etc. The instruments recorded on digital cassettes, each of which had a capacity of twenty-five to thirty events at the five second post-trigger duration setting. In order to ensure that there would always be tape available to record large events even if cassettes were not changed each day, and to ensure that the passing of the thirty-six ton snowplow used to clear the access roads did not cause too many false triggers, trigger sensititities were set low. In addition, gains were set relatively low on the FBA instruments to prevent them from clipping the signals of any large events.
On the first two days of the study, instruments were deployed as near to the teleseismically determined epicenter as was logistically possible. The first records made by these instruments were played back in the field on an analog strip chart recorder, and the arrival times of aftershocks were picked. The arrival times were transmitted orally by telephone to U.S.G.S., Menlo Park, Californa, where a computer program was used to calculate approximate aftershock locations. These locations made it clear that the seismic activity lay to the west of the established stations. Examination of maps of the study area revealed that there was only one road that had any westerly component and was at the same time passable to the snow-plow. Since there were no buildings at all along this route, a rented van equipped with a kerosene space heater was left at the site to house the instruments.
In the last two days of the study, after weather conditions had relented somewhat and a routine of field operations had become established, the gains were raised on the velocity transducer equipped digital instruments in an attempt to record some of the aftershocks of the smaller magnitudes that were previously missed. The daily gain settings are listed in Table 3.
In the one week period of study, a total of forty aftershocks were recorded by at least one digital instrument. These recordings probably include all aftershocks greater than local magnitude 1 mbLg. At this time, the Canadian EPB has not officially assigned magnitudes to any of these aftershocks with the single exception of the largest event which occurred at 13:33:56.2 GMT, January 17. This aftershock was recorded by at least four regional Canadian seismograph stations, the furthest one away being SCH at a range of 870 km. This event has been assigned a magnitude of 3.5 mbLg.
The forty recorded events have been located by the Canadian EPB using an in-house computer program written by R. Wetmiller. The velocity model assumed was a homogeneous half-space with a 6.2 km/s P-wave velocity, a 3.57 km/s S-wave velocity, and corrections made for station elevation. Arrival times at the digital stations have been supplemented with these taken from Canadian EPB smoked-drum recorders in the epicentral region. Figure 4 shows the locations of what have been referred to as the Canadian smoked-drum instruments. These include smoked-drum recorders operated by Massachusetts Institute of Technology and by Lamont-Doherty Geological Observatory whose records have been included in the Canadian EPB data analysis. The locations of the recorded events can be seen in Figure 2, with the epicenter of the magnitude 3.5 event plotted as a star twice the size of the stars representing other aftershocks. The origin times, epicenters, and depths are listed in Table 4 and the epicenters are plotted by their listed number on a large scale map in Figure 5. These locations are preliminary pending the release of the final locations by the Canadian EPB in September, 1982. The depths listed in Table 4 are in km. below sea level. The headings "NS" and "NP" refer to the number of stations and the number of phases respectively used in locating the different events. Because of the preliminary nature of the locations, the estimates of spatial location error are not included. It is sufficient to say that with the exception of the first event, the spatial errors are significantly less than 1 kilometer. The root mean square (RMS) values of the travel time residuals have been included as an approximate indication of the accuracy of the locations given the small number of stations used.
During the study period, there were many small events that did not trigger any of the digital instruments. These small events were visible on the smoked-drum recorders co-located with the digital instruments. Most of the larger events, i.e., those that triggered the digital instruments, had S-P times approximately equal to one second. The smaller events seen on the smoked-drum records were an order of magnitude more numerous than the larger events and had S-P times that ranged down to less than .1 seconds. This implies that the locations of the smaller events were distributed throughout the source volume as a whole. They might be visualized as being the response of an isotropic solid, the source volume, to the strain relaxation caused by the stress release of the mainshock and major aftershocks. Conversely, the 4 to 6 kilometer depth of the larger events and their relative areal concentration suggests that the larger events are related to the mainshock fault plane in particular.
Peak accelerations of approximately .08 g were recorded by instrument C7A for both the P-wave arrival on the vertical component and the S-wave arrival on the N-S horizontal component of the magnitude 3.5 aftershock (Event no. 8 in Table 4). Instrument C7V recorded a peak velocity of 1.25 cm/s on the N-S horizontal component for the S-wave arrival of the same event. All the digital records made of the magnitude 3.5 aftershock are shown in Figures 6a-6f. A small foreshock (Event no. 7 in Table 4) that occurred 57 seconds prior to the magnitude 3.5 event and which had roughly the same location and which was recorded by the same instruments is also included in these figures for comparison. The epicenters of the events are plotted as numbered stars, and the corresponding recorded traces are plotted to the right and just above and below the respective stations. The numeral "1" refers to the magnitude 3.5 event and the numeral "2" refers to the foreshock. The time scale and the amplitude scale are indicated by the time-amplitude axis plotted in the upper left hand corner of each figure. Only the vertical component of velocity at Station C9V appears in these figures because that the only component recorded at that station at the time of these earthquakes. The horizontal components of the other stations have been rotated into the radial and transverse directions (in terms of source-receiver).
The most striking feature of these figures is the contrast of the recordings of the main 3.5 event made at Station C7T with those made at Station C8T. The differences in waveforms are most evident in the transverse components of the respective stations; the traces at C7T are far more impulsive and more compact in nature than those at C8T. Both of these stations are approximately the same distance from the epicenter, and the radial directions at the two stations are within 30° of each other. However, examination of Figure 2 will show that Station C7T is sited in an area of low relief while Station C8T is sited in an area of high relief. The observed differences in waveforms could be caused by differences in surface scattering at the two stations (Boore et al., 1981).
Comparison of the recordings of the two events shows that the finite dimension of the source influences the waveforms of the larger earthquake. The P-wave arrivals of the foreshock on the vertical components are much simpler and consist of fewer cycles than the P-wave arrival of the magnitude 3.5 event on the same components. Homogeneity of the propagation medium is evident in the absence of phases after P and before S. This is particularly true of the foreshock waveforms recorded on instrument C7V which resemble the impulse response of a homogeneous half space.
Preliminary source paramenter calculations using a Brune model (Brune 1970, 1971) have been made for twenty-five of the digitally recorded aftershocks. Calculated moments ranged from 1020 to less than 1017 dyne-cm (Cranswick and Mueller, 1982). A complete study of the source parameters of all the digitally recorded events is currently in progress (Mueller and Cranswick, 1982).
Table 5 is a key to the digital recordings made of the aftershocks listed in Table 4. The raw data files of all the digital records indicated in Table 5 are included on the digital data tape and plotted in Appendix A as a reference. The figures in Appendix A are whole record plots, and so each has a time scale that corresponds to the record duration. The amplitude scales are normalized to the peak amplitude which permits the reader to ascertain the peak value of a record by inspecting the amplitude-scale units label. Some of the figures, reflecting the state of the digital data, still have time-marks in the traces which remain in the time-series because of processing difficulties caused by instrument malfunction due to the low temperature recording conditions.
This is a valuable data set because it vastly increases the number of digital strong motion records of Eastern North American seismicity. It includes, to the best of our knowledge, the largest event in the Northeast ever recorded by digital instruments.
We wish to thank the officials of Provincial Government of New Brunswick for giving us their full cooperation in the field program. We wish to thank the people of Newcastle for providing us a place to which we could return from the cold. Two organizations were absolutely essential for field operations: the Department of Highways and the Department of Natural Resources of New Brunswick. Without the snow-plow crew from the Newcastle office of the Highway Department, we could have never reached the study area in the first place, and without the officers of the McGraw Brook Ranger Station, we would have never returned from it. Ralph Hudson, both as a representative of the New Brunswick Emergency Measures organization, and as a very enthusiastic and interested private citizen, left no stone unturned in responding to whatever challenge befell our Logistics. We would like to thank the Northeastern Seismic Network groups at both MIT and Lamont-Doherty for providing us with their data. The American authors of this report wish to thank the authorities of the Canadian Federal Government for their assistance, and the Earth Physics Branch in particular for their invitation to us to participate in this aftershock study. Edward Cranswick wishes to thank his wife, Sandra, for providing the beautiful figures that served as the carrot before the horse of this report. This report was promptly and carefully reviewed by Joe Andrews, and we are grateful to him.
Boore, D., Harmsen, S., and Harding, S., 1981,
Wave scattering from a step change in surface topography:
Bulletin of the Seismological Society of America,
v. 71, no. 1, pp. 117-125.
Brune, J., 1970,
Tectonic stress and the spectra of seismic shear waves from earthquakes:
Journal of Geophysical Research,
v. 75, pp. 4997-5009.
Brune, J., 1971,
Correction,
Journal of Geophysical Research,
v. 76, p. 5002.
Cranswick, E., and Mueller, C., 1982,
Source parameters of aftershocks of the January 9, New Brunswick earthquake [abs.]:
EOS,
v. 63, no. 18, p. 383.
Cranswick, E., and Spudich, P., 1982,
The history of rupture of the October 15, 1979 Imperial Valley earthquake as determined from apparent phase velocities measured at the Differential Array
(in preparation).
Fletcher, J. B., 1982,
A comparison between the tectonic stress measured in situ and stress parameters from seismic waves at Monticello, South Carolina: A site of induced seismicity:
Journal of Geophysical Research
(in press).
Moos, D., and Zoback, M., 1982,
In situ studies of seismic velocity in fractured crystalline rocks:
Journal of Geophysical Research
(in press).
Mueller, C., and Cranswick, E., 1982,
Comparison of the source parameters of earthquakes in New Brunswick, Canada with those of earthquakes in Mammoth Lakes, California
(in preparation).
Potter, R., Hamilton, J., and Davies, J., 1979,
Geological map of New Brunswick, Department of Natural Resources, Geological Survey of Canada.
Schlesinger-Miller, E., Barstowe, N., and Kafka, A. 1982,
Intensity surveys for three eastern North American earthquakes in January 1982 [abs.]:
Seismological Society of America, Spring Meeting, April, 1982, Anaheim, California.
Front Cover Map: New Brunswick
Figure 1: Geological map of Northern Central New Brunswick
Figure 2: Topographic map of study area
Figure 3: Test of instrument reliability
Figure 4: Smoked-drum recorder location map
Figure 5: Aftershock location map
Figure 6: Digital records made of the magnitude 3.5 aftershock
Figure 6c: Transverse velocity
Figure 6d: Vertical acceleration
Table 1: Station descriptions and histories
Table 3: Station-instrument histories
Table 5: Aftershock trigger table
Appendix A: Raw Data Whole Record Seismograms
Appendix B: Seismogram Processing
The GEOS Home Page is at /GEOS/geos.html
Figures
Tables
Appendix
The USGS Home Page is at http://www.usgs.gov
The USGS Geologic Division Home Page is at http://geology.usgs.gov
The USGS National Strong Motion Program Home Page is at http://earthquake.usgs.gov/monitoring/nsmp
The URL of this page is /GEOS/NBK82/OFR_82-777/New_Brunswick.html
e-mail: glassmoyer@usgs.gov
or address:
Gary Glassmoyer
345 Middlefield Road M/S 977
Menlo Park, California 94025