Broadband Sensors (Under Construction)

Work accomplished thus far:

We chose to adapt the Precision Measurement Devices (PMD) 2123 seismic transducers to the seafloor. The choice of these sensors was made based on the their ruggedness, robust leveling requirements, and cost. The horizontal sensors are inherently self-leveling, up to a tilt of 5°. This applies to the vertical sensor as well, but there will be be an error since the translation measured will not be truly vertical. The sensors are electrochemical in nature, and are related to the Solion sensors used in early versions of the Sacks-Evertson dilatometer. The power requirements are low, around 42 mW. This is excellent compared to the 300 mW of the Guralp low-power sensors. The development of these sensors was supported, in part, by the National Science Foundation. The price of the sensor is just over $6,000, about twice that of the Mark Products L4-3D, but considerably lower than the price of the Guralp sensor. Figure 7 is the specification sheet for this device. Although the noise floor of these sensors is above the continental low-noise model, so is the noise on the sea floor. The particular model we are using is flat in response to velocity and has a low-frequency corner at 0.03 Hz (33 second period). We now have seven of these sensors in hand or on order.

The PMD sensor is shown in Figure 8 , which shows the PMD device next to the L4-3D sensor package presently in use. Because the initial proposal was supported at a reduced level, we adapted the existing sensor packages Barash and others (1994) for use with the PMD sensor. This package has undergone extensive testing (Trehu and Sutton, 1994) .[ trehu sutton 1994 .]

Land tests: We have tested the sensor in the laboratory as part of the MELT sensor intercomparison (Sauter and Dorman, 1997) and found that it provided about 30 dB improvement in low-frequency noise level over the Mark Products L-4. This allow access to many more earthquakes.

Sea tests: We have preliminary results from two SIO-ONR OBSs fitted with PMD 2123 broadband sensors and deployed in the San Diego Trough (SDT) at about 1km water depth at a separation of a few hundred meters.

Figure 9 shows spectra for a time segment containing background noise. The size and scale of the plot is designed to overlay the plot of seafloor and subseafloor noise spectra in the OSN pilot experiment at http://msg.whoi.edu, specifically, their figure 3e-3. Our spectra are computed using the multiple-window method with a window length of 32768 points (1024 seconds).

The differences and similarities seem to be reasonable consequences of the different environments:

(1) At frequencies below 0.08 Hz, the well-known "noise notch", the SDT data exhibit noise levels which are about 40 dB higher than at the OSN site. This is because the shallower depth of the SDT site allows infragravity waves to "feel" the seafloor (Webb, 1998). The horizontal components show noise levels about about 20 dB higher than the verticals, as do the seafloor horizontals at the OSN site. .[ webb 1998 noise under ocean .]

(2) Between 0.08 and 0.2 Hz, the noise spectra at the two sites are about the same (the Holu spectrum of McCreery and others, 1993). .[ mccreery duennebier sutton 1993 .]

(3) At higher frequencies, the SDT site is a little quieter, since it is less windy there (c.f. Dorman and others, 1993).

It is difficult for us to determine, at this point, where 1/f noise sets in.

The MOISE experiment described by Stakes and others (EOS, 1998) should provide an environment more like the SDT than did the OSN experiment, but no spectra were provided in that paper, which demonstrated the ablility of an ROV to emplace and assemble a seismic instrument on the seafloor. .[ stakes romanowicz montagner 1998 .]

The scatter between the levels of the horizontal components may be anomalously large. Figure 9a shows the ratio of the amplitude of horizontal noise to vertical noise are determined by the structure and the mode number of the dominant mode at the frequency in question (Dorman and Schreiner, unpublished manuscript).

Figures 10, 10a, 11, 11a, are amplitude spectra (as calculated by N. Kanjorski) showing P-wave corners for two earthquakes which occurred during the test deployment. The Mexico event was of magnitude 5.8 and the Nevada event was 5.1. For the Mexico event, the flat part of the spectrum extends down to a few hundredths of a Hertz.

While the comparison is not exact, Figure 12 is a similar spectrum from an outer rise event of magnitude 4.7 recorded on the LABATTS experiment using the Mark Products L4 sensors. The conclusion I offer is that the additional bandwidth provided by the PMD sensor is useful in extending the frequency range of the instruments below the microseism peak.

For the SDT test, the instrument was not configured for optimum coupling. When we are being very conservative, we fix the sensor to the OBS frame, rather than deploying it away from the body of the OBS. This increases the likelihood of obtaining data since success is not dependent on the performance of the deployment mechanism, but introduces some contamination of the signal since the package is not symmetric, the mass/area ratio is not optimal (see Duennebier and Sutton, 1995), and the instrument frame introduces spurious resonances. .[ duennebier sutton 1995 fidelity .] Nonetheless, the spectra are not greatly dissimilar to spectra from land sites, at least for the vertical component.

History

The general specifications for the ONR OBSs were established by a broadly based ad hoc committee named the Seafloor Noise Advisory Group (SNAG), in response to a navy need to understand more about the seismo-acoustic noise on the sea floor. The design of the instruments was, however, made sufficiently flexible that they are useful for a wide variety of marine or lacustrine seismic experiments, both passive and active.

The design and manufacture of the instrument was undertaken by a consortium consisting of the Scripps Institution of Oceanography, the Woods Hole Oceanographic Institution, the Massachusetts Institute of Technology, and the University of Washington. The SIO group provided the analog systems and control computer, the WHOI group undertook the mechanical systems design and recording module, the MIT group the mechanical design of the sensor package. The UW group designed the analog-to-digital converter (ADC) board.

Thirty-one of the instruments were constructed and divided between the West Coast (SIO) and East Coast (WHOI) operating bases and the instruments have been in frequent use since about 1990, on projects supported by the Office of Naval Research and by the National Science Foundation. Two have been lost. At the present time (1998) the Office of Naval Research is not actively managing the facilities.

General Description

The instruments are described in Sauter and others, 1990. The analog electronics are based on a series of earlier systems operated at SIO (Prothero, 1976), (Moore and others, 1981) all of which used seismometers of 1 Hz frequency.

The orthogonal 3-component set of seismometers (Mark Products L4-3D or PMD Broadband sensors) are mounted in gravity-leveled gimbals damped in a high-viscosity fluid (polydimethylsiloxane, 980 Pa-s viscosity). The gimbals are suspended in a pressure case of 0.35 m diameter. The properties of this sensor package have been extensively studied (Barash and others, 1994) and are well known. In use the pressure case is deployed on a pivoting arm which drops it to the seafloor about a meter away from the instrument frame (See Figure 1). This provides some isolation from mechanical noises generated by the recording devices and from vortex shedding from seafloor current acting on the mechanical frame (Trehu and others, 1994).

The external sensor package contains the preamplifier, which is controlled by the acquisition computer to provide gains ranging from unity to 512 (0 to 54 dB), and the circuits which generate a pseudo-random telegraph code for calibration (Berger, Agnew, Parker and Farrell, 1979; Sauter and Dorman, 1986). The analog amplifiers utilize shaping filters which reduce the dynamic range required to cope with the background noise on the seafloor (which has steeper spectrum than that encountered on land). The OBS uses a 16-bit ADC which provides a dynamic range of 90 dB. Immediately before the digitizer is a fast gain-ranging circuit which allows reduction of gain by a factor of 8 or 64 (36 dB) to avoid clipping on very large signals. Thus the aggregate dynamic range is 120+ dB.

The system operates on batteries and writes the digital seismic and timing information to a SCSI storage device. Currently we are using a mix of devices, since we are in the midst of an upgrade. We have been using the Hewlett-Packard 35470 DAT tape transport, which holds 2 gigabytes of data, and can accommodate three transports (6 gigabytes total) per instrument. These are being replaced by disk recorders with 9.1 gbyte capacity. The disks allow recording of 4 channels at 128 Hz for 100 days.

An external computer (either a DOS or Unix) is used to program and operate the OBS. Once data is recorded, the raw data tape is read and converted to ROSE (Latraille and Dorman, 1979) or SEG-Y format. We are now committed to provide the SEED format.

Table 1. Sample rates and recording times available with 9.1 gbyte disks.

Usage:

The ONR instruments have been in active use since their development. As evidence of this, we show the schedule of usage of the West Coast Facility. The East Coast group has had a similar level of effort and some of the experiments below have used or will use instruments from both facilities.

For past and planned instrument use, see the OBS Facility home page

To schedule use of these instruments, contact Professor Dorman at the address above. The facility staff will provide more detailed information on instrument capabilities, advice on instrument usage, and provide a budget, to be incorporated in the requestors proposal. Scheduling control ultimately rests with the Office of Naval Research.

Comparison with other U.S. OBSs:

Existing OBS instruments in the U.S. are roughly divided into two categories which I will, for brevity, call "Small" and "Large", although the distinction blurs at times. The small instruments typically use 4.5 Hz geophones or hydrophones, have somewhat limited recording capacity and endurance, and are typically used in active-source seismic experiments and for microearthquake studies with durations of a week to a month. The "Large" instruments implement features of land observatories in terms of bandwidth, dynamic range, timing accuracy, and endurance. The latter are more useful for observation of teleseisms and are more focussed on the needs of "whole earth seismologists".

The largest fleet of "Small" instruments are those at UTIG. These are encased in glass spheres, and are easily handled by one person. This fleet numbers 20, with 15 additional units operated collaboratively by ORSTOM. The other "Small" instruments are those of the USGS, whose fleet numbers about 7.

The "Large" instruments provide many features of land seismic observatories, relatively high dynamic range -up to 126 dB-, excellent clock stability-1 ms per day drift. They all use 1 Hz inertial seismic sensors and can see earth noise down to about 0.05 Hz. Large events have provided surface wave signals down to 0.01 Hz (Blackman Orcutt, Forsyth, 1995). This class of instruments can be equipped with hydrophones useful down to a millihertz. Endurance of 4 months on the seafloor with three months of continuous recording- sampling 4 channels at 32 Hz has been demonstrated on the LABATTS- Lau Basin experiment. The operators of these instruments have committed to record continuously for six months on the MELT experiment.

The largest fleet of "Large" instruments is that constructed in 1989 with ONR funding and operated at two facilities, one at SIO operated by LeRoy Dorman and the other at WHOI operated by Mike Purdy. This fleet is comprised of 29 instruments. Next in order of population are the SIO/MPL instruments operated by Spahr Webb, which began their lives as hydrophone instruments but have been fitted with 1 Hz seismic sensors. This fleet numbers about 16. Third in rank is the SIO/IGPP fleet operated by John Orcutt, which numbers 5 instruments.

Table 2 compares the characteristics of the "Large" OBSs.

Table 2. Characteristics of U.S. Ocean Bottom Seismographs.

The schedule indicates a high level of activity supported by both NSF and ONR. Availability of broader-band sensors would enhance all of the experiments in which the instruments have been used and all those which are currently scheduled.

Evolution of usage of OBS systems

It is likely that the "Large" instruments will be used for the longer deployments (six months to a year) required for teleseismic studies and that the "Small" instruments will be used for smaller-scale structural studies, along with Ocean-Bottom Hydrophones (OBHs) in operation at WHOI and SIO/IGPP (currently under construction under NSF funding).

References:

Agnew, D. C., J. Berger, W. E. Farrell, J. F. Gilbert, G. Masters, and D. Miller,
Project IDA: a decade in review, EOS, Transactions of the AGU, 67, 203-212, 1986.

Barash, T. W., C. G. Doll, J. A. Collins, G. H. Sutton and S. C. Solomon,
Quantitative Evaluation of a Passively Leveled Ocean Bottom Seismometer,
Mar Geophys. Res., 16, 347-363, 1994.

Berger, J, D. C. Agnew, R. L. Parker, and W. E. Farrell, Seismic Calibration: 2.
Cross-spectral calibration using random binary signals.
Bull. Seismol. Soc. Am., 69, 271-288, 1979.

Forsyth, D. W. and A. D. Chave, Experiment Investigates Magma in the
Mantle Beneath Mid-Ocean Ridges, EOS, Transactions, AGU, 46, 537-540, 1994.

R. S. Jacobson, L. M.. Dorman, G. M. Purdy, A. Schultz, S. Solomon,
Ocean Bottom Seismometer Facilities Available, EOS, Transactions AGU, 72, pp506,515, 1991.

Latraille, S, and L. M. Dorman, A standard format for the storage and
exchange of natural and explosive source seismic data: the ROSE format,
Mar. Geophys. Res., 6, 99-105, 1983.

Montagner, J. P., Romanowicz, and J. F. Karczewski, A first Step Toward
an Oceanic Geophysical Observatory, EOS, 75, 150-151, 1994.

Prothero, W. A., A Digital Event-Recording Ocean Bottom Seismometer
Capsule, Mar. Geophys. Res., 3, 119-141, 1976.

Moore, R. D., L. M. Dorman, C-Y Huang, D. L. Berliner, An ocean bottom,
microprocessor-based seismometer, Mar. Geophys. Res., 4, 451-477,
1981.

Sauter, A. W. and L. M. Dorman, Instrument calibration of Ocean Bottom Seismometers,
Mar. Geophys. Res., 8, 265-275, 1986.

Sauter, A. W., J. Hallinan R. Currier, T. Barash, B. Wooding, A.
Schultz, L. M. Dorman, Proc MTS conf on Marine Instrumentation, 99-104,
1990.

Webb, S. C., W. C. Crawford, and J. A. Hildebrand, Long Period
Seismometer deployed at OSN-1, Seismic Waves (Ocean Seismic Network
Newsletter), 3, 4-6, 1994.

Last Revised: 26 August, 1998 lmd

This document was translated by troff2html v0.21 on March 6, 1995.