- Email: [email protected]
Development of a seafloor acoustic ranging system toward the seafloor cable network system
ARTICLE IN PRESS Ocean Engineering 35 (2008) 1401–1405
Contents lists available at ScienceDirect
Ocean Engineering journal homepage: www.elsevier.com/locate/oceaneng
Development of a seafloor acoustic ranging system toward the seafloor cable network system Yukihito Osada a,, Motoyuki Kido a, Hiromi Fujimoto a, Yoshiyuki Kaneda b a b
Graduate School of Science, Tohoku University, 6-6 Aramaki, Aoba-ku, Sendai 980-8578, Japan Japan Agency for Marine-Earth Science and Technology, 2-15 Natsushima-cho, Yokosuka 231-0061, Japan
a r t i c l e in fo
abstract
Article history: Received 20 December 2007 Accepted 15 July 2008 Available online 23 July 2008
We have developed a seafloor acoustic ranging system as a possible future application to monitoring seafloor crustal movement with the DONET (Development of Dense Ocean-floor Network System for Earthquake and Tsunami) cabled observatory system. In 2007 we carried out an experiment for the seafloor acoustic ranging system. We deployed two precision acoustic transponders (PXPs) with about 750 m spacing in Kumano-nada at a water depth of about 2035 m. We collected 660 ranging data in this one-day experiment. The round-trip travel time shows a variation with peak-to-peak amplitude of about 25 mm in the range. It was confirmed that most of the variation could be explained by the change in sound speed estimated from measured temperature and pressure. The remaining fluctuation in the acoustic measurements is 72 mm. & 2008 Elsevier Ltd. All rights reserved.
Keywords: Seafloor crustal movement Acoustic ranging Seafloor cable network system
1. Introduction The Philippine Sea Plate is subducting beneath the Eurasian Plate along the Nankai Trough, which has been known as a region of M8 class earthquakes, that have occurred with recurrence intervals of about 100–150 years (Ando, 1975). About 60 years have passed since the last earthquakes happened during 1944 and 1946. The Japanese Government’s Earthquake Research Committee estimated a 60% probability that the next earthquake in this area will occur within the coming 30 years. However, since most of the source region of the earthquakes at Nankai Trough is located beneath the ocean, an observation system is necessary to monitor tectonic activities in the offshore source region. The project DONET (Dense Ocean-floor Network System for Earthquake and Tsunami) aims to develop a real-time dense ocean floor network system around the Nankai Trough (Kawaguchi et al., 2007). This project is developing a cabled seafloor observatory with the focus on monitoring earthquakes, tsunamis and crustal movements on the seafloor. Each of 20 sites of observation is composed of precision seismometers, a pressure gauge and other instruments. All the sites would be connected to a submarine cable and deliver real-time data to land. We have developed a short-range seafloor ranging system as a possible future application to the DONET cable system. Direct acoustic ranging is a simple way to monitor local crustal deformation, and various groups have successfully detected Corresponding author.
E-mail address: [email protected] (Y. Osada). 0029-8018/$ - see front matter & 2008 Elsevier Ltd. All rights reserved. doi:10.1016/j.oceaneng.2008.07.007
deformation on the seafloor using such a system. Long-term monitoring of seafloor crustal deformation has revealed intermittent ridge extension at the Juan de Fuca Ridge (Chadwick et al., 1995, 1999; Chadwell et al., 1999; Chadwick and Stapp, 2002) and the East Pacific Rise (Nagaya et al., 1999). Direct acoustic ranging is different from GPS/Acoustic methods that involve indirect ranging to instruments on the seafloor from a vehicle towed by the ship or from a buoy on the surface (Spiess et al., 1998; Osada et al., 2003; Sweeney et al., 2005; Kido et al., 2006; Ikuta et al., 2008; Chadwell and Spiess, 2008). We aim to monitor the splay faults in the rupture area of the Tonankai earthquake in the Nankai subduction zone (Park et al., 2002). Slip along the active splay faults may be an important mechanism that accommodates the elastic strain caused by relative plate motion. Ito and Obara (2006) shows that VLF (very-low-frequency) earthquakes occur near the mega splay faults that branch from the plate boundary and have an estimated slip of 2–10 mm at maximum. We plan to deploy PXPs (precise acoustic transponder) across the splay faults in the Kumano-nada area to measure the horizontal crustal movement to the accuracy required to detect VLF earthquakes. In this study we report on the trial observations with a new direct acoustic ranging system and estimate the accuracy of the acoustic measurements.
2. Seafloor observation system Fig. 1 shows a schematic diagram of a prototype PXP. The PXP consists of a monopod anchor, a pressure housing in the shape of a
ARTICLE IN PRESS 1402
Y. Osada et al. / Ocean Engineering 35 (2008) 1401–1405
Fig. 2. Concept diagram for the seafloor acoustic ranging system, for detecting the horizontal seafloor crustal movements.
Fig. 1. Schematic of the seafloor acoustic ranging system. The system is equipped with a tilt meter, a compass, a pressure gauge, and a temperature gauge.
1700 -diameter glass ball for batteries and electronics, and an acoustic transducer on the top, and weighs 37 kg on land. This design was chosen to allow the instruments to be deployed in relatively rough seafloor terrain while keeping the transducer well above the local topography. Although such a design allows the PXP to sway slightly due to bottom currents, the internal tilt meter and compass can be used to correct these effects. This instrument costs about $40k each and is expected to have at least one year lifetime. Individual PXPs in pairs act as either a ‘‘master’’ or a ‘‘slave’’. The master can record the received signals of up to four slaves. To initiate a range measurement, the master sends out a command signal. The PXPs use a relatively simple detection method, triggering on the leading edge of an incoming wave when the amplitude rises above a fixed threshold. When the slave detects the master’s signal, it replies with an identical wave (after a precise 3 s delay). Finally, the master records the signal from the slave. This system automatically adjusts the receiver gain for the incoming signal and records the receiver gain. Fig. 2 shows the schematic diagram of the prototype seafloor acoustic ranging system, for detecting horizontal seafloor crustal movements. This system uses direct-path acoustic ranging between pairs of PXPs. The round-trip observations can cancel the effect of water currents in the ocean. This system has two advantages. One is the stability of sound speed near the deep seafloor. Most of the variations in the speed of sound result from variations in temperature, which is fairly stable in the ocean at mid-latitudes and at depths greater than 1000 m (Urick, 1983). The other main advantage is the low level of acoustic noise on the seafloor. Since underwater sound is generated mainly by winds and surface waves, the deep seafloor is an optimal place for acoustic observations. As acoustic ranging measures the travel time along the baseline between pairs of PXPs, the variation in sound speed has a direct effect on the measurement. The PXP records pressure and
Fig. 3. Map showing the location of the seafloor acoustic ranging system at the Kumano-nada. The grey diamonds show a pair of PXPs. The grey circles show PXPs for GPS/Acoustic observation on Kumano-nada (Kido et al., 2006). Lines are contours of water depth. Inset map shows tectonic setting and location of the experimental site in Japan. Black square indicates the location of the large map. Lines indicate the plate boundary; PH, the Philippine Sea Plate; EU, the Eurasian Plate; and NA, the North American Plate.
temperature data for the correction of the effects of the variation in the sound speed. We have not collected salinity data, since its influence on sound velocity is less when compared to temperature and pressure. For example, an increase in temperature by 0.003 1C or salinity by 0.009 % or a depth change of 0.9 m would cause an apparent shortening of the 1000 m baseline by about 1 cm. In the Kumano-nada area, Hamamoto et al. (2005) made heat flow measurements. The temperature variation shows a peak-to-peak amplitude of 0.2 1C during 300 days. Therefore it is important to observe the temperature with this system. Sound speed in mid-latitude waters shows a minimum value at a depth of around 1000 m and then increases with depth due to increasing pressure. Accordingly, a ray path of an acoustic wave
ARTICLE IN PRESS Y. Osada et al. / Ocean Engineering 35 (2008) 1401–1405
1403
tends to be bent upwards in the deep sea due to the gradient in the speed of sound. The PXP consists of a 2.5-m-tall monopod anchor. Although such a design allows the PXP to sway slightly due to bottom currents, the internal tilt meter and compass can be used to correct these effects.
3. Experiment with the seafloor acoustic ranging system We carried out an experiment with the seafloor acoustic ranging system on 5 August 2007. Two PXPs were deployed at a depth of 2035 m in the Kumano-nada area (Fig. 3). We were forced to deploy them from the surface. The baseline was about 750 m. This experiment was aimed at estimating the precision of acoustic ranging. We measured the two-way travel time with four types of transmitted signals at 10 min intervals. Concurrent measurements of attitude, temperature, and pressure were also carried out. Data comprising 660 measurements were collected during a period of 24 h. We examined four types of transmitted signals for precise acoustic ranging. Of the four types, three were M-sequence (maximum length sequence) signals and one was a sweep signal. The M-sequence signals, which are basically a pseudo-random sequence of pulses, were composed of sine waves with positive and negative phases. The sweep wave for the chirp signal was composed of sine waves whose frequency increased at a linear rate with time. Both the waves were found to be effective in reducing the effect of noise, when measuring the arrival time through cross-correlation (e.g., MacWilliams and Sloane, 1976). Three signals adopt a 10 kHz acoustic wave, which were coded with a 5th-order M-sequence with 31(251) bits. One bit of these transmitted signals consisted of carrier waves of two, four, and six waves; thus the length of the whole signal amounts to 6.2, 12.4, and 18.6 ms. In contrast, the sweep wave was composed of sine waves with increasing frequency between 8 and 14 kHz; thus the length of the whole signal amounts to 1.4415 ms. The master records the signal from the slave with a sampling interval of 10 ms, which corresponds to a one-way range resolution of 7.5 mm. We can resample the signal with an interval of 1 ms using cubic spline interpolation when we calculate the correlation. The master and the slave record the temperature and attitude at both ends of the distance measurement. Pressure is recorded only by the master. The information obtained with the compass and the tilt meter at each PXP are then used to estimate the attitude of each monopod.
4. Results The round-trip travel time, temperature as observed by the two PXPs, and pressure measured by the master PXP were then used to derive the range between the PXP pair. The value of the sound speed was computed from the observed pressure and temperature using the equation for the speed of sound (Chen and Millero, 1977). The ranges were corrected for instrument tilts. The time series of the observed round-trip travel times are shown in Fig. 4. Peak-to-peak variation of the travel times was 18 ms for the range measurements, which corresponds to about 25 mm (Fig. 4(a)). Fig. 4(b) shows the correlation results for the four different codes. All the received signals were stable and of high quality. Fig. 5(a) shows temperature variation observed by the master and the slave. The discrete values of the temperature are due to the resolution of the measurements. This area shows fairly stable water temperature, and the peak-to-peak variation was about 0.010 1C during this experiment. Since the temperature resolution was limited, and the water temperature was slightly different at the two devices, the temperature variation was estimated by averaging the
Fig. 4. Time series of (a) the round-trip travel time and (b) correlation. Squares show M-sequence signals with 4 waves per bit. Circles correspond to sweep signals. Triangles and diamonds show M-sequence signals with 2 and 6 waves per bit.
data with a 30-min moving window. The variation in the sound speed was estimated to be approximately 0.017 m/s during the experiment, which resulted in a variation of 18 mm in the ranges. The effect of tidal pressure variation was corrected with the observed pressure data. The peak-to-peak variation was about 1.2 dbar during this experiment. At 2035 m depth, the gradient in sound speed (inferred from measurements of 1.99 1C, 35 % at 2060 dbar) was approximately 0.008 (m/s)/dbar and the resultant range correction was up to 0.010 m. The residual variation in the round-trip travel time after the correction of these effects showed that the range across this 747 m distance was measured with a peak-to-peak range of 8 mm (Fig. 6). These results show that 70% of the variation in the ranges could be explained by variations in temperature and pressure. Next, we correct the effect of the change in attitude of the PXPs, using the data from the tilt meters and the compasses. Fig. 7 shows the time series of attitude data on each PXP. The attitude data show a change of 11 in heading and 0.11 in pitch and roll. This correction of range is up to 8 mm. The range residuals show a precision of about 2 mm, when we take into account all of the above-mentioned corrections (Fig. 8).
5. Discussion We compared the four different transmitted signals (or codes) used in this experiment to find the best suitable code. The
ARTICLE IN PRESS 1404
Y. Osada et al. / Ocean Engineering 35 (2008) 1401–1405
1.995
Temperature [degree C]
AVE Master Slave
1.99
1.985
1.98 08/05-12:00
08/06-00:00
08/06-12:00
08/06-00:00
08/06-12:00
Pressure [dbar]
2060
2059.5
2059
08/05-12:00
Time [JST] Fig. 5. Time series of temperature and pressure at each PXP. (a) Crosses and squares show the variations in temperature at the master and the slave, respectively. Black line depicts the variation of the average temperature, which was calculated by using a moving average of 30 min. (b) Time series of pressure variation at the master.
Fig. 6. Time series of ranges showing the effects of the sound speed correction. Line-plot with circles shows the sound speed-corrected ranges. Line-plot connected with crosses uses only an average sound speed. Line-plot with circles is the 5th-order M-sequence with 4 waves per 1 bit.
Fig. 7. Time series of the attitude data at each PXP (cross: master, circle: slave).
cross-correlation was more than 0.7 for all the codes used. The RMS (root mean square) deviation of the observed result with each code was 1.5 mm from the average value with the four codes. The cause of this difference may be the fluctuation of the electric mechanism that makes the transmission signals. Most received signals were coherent and had a good signal to noise ratio. Consequently, it was possible to measure the travel time with a precision of better than 1 sample (10 ms). During the deployment on 08/05 at 21:00, we recorded a change of 0.141 in the roll value. The temperature showed a similar change of 0.008 1C, as obtained from two temperature gauges. It is thought that a coherent bottom current crossed the area at that time. After 08/06 at 00:00 the two temperature gauges show slightly different changes. This was apparently a localized change, and the temperature variation was found to differ by 0.04 1C at the two PXPs. In the temperature data, about 60% of the temperature variation was coherent across the baseline of 750 m. The range residuals show a resolution of about 2 mm after corrections for the variations in temperature, pressure, and the attitude of the instruments. The good results may be partly due to the fact that we carried out the experiment under good conditions. Although the experimental site was landward of a trench, the bottom currents were weak, presumably due to the
ARTICLE IN PRESS Y. Osada et al. / Ocean Engineering 35 (2008) 1401–1405
1405
2035 m on the Kumano-nada for 24 h. The baseline length was about 750 m. We estimated the precision for acoustic ranging to be about 2 mm after correcting for the variations in the sound speed and the attitude of the instruments.
Acknowledgments This study was supported by the DONET Project of Ministry of Education, Culture, Sports, Science and Technology. The two anonymous reviewers whose comments helped improve the manuscript. The generic mapping tools (GMT) software (Wessel and Smith, 1991) was used to prepare the figures. References
Fig. 8. Time series of ranges after correcting for attitude. The line-plot is a 5thorder M-sequence with 4 waves per 1 bit.
flat-bottom topography. We hope to improve the system for achieving a precision of 1 mm or better for the range. One of the problems lies in the resolution of the temperature measurements, which currently is 0.001 1C. The temperature change around the device can be more accurately observed by introducing a temperature gauge having a better resolution and accuracy in the future. Another source of difficulty is the location of the water temperature measurements. The temperature was observed at both sides of the PXP pair for this experiment, but the temperature change along the ray path may not be accurately represented using the average temperature. In the future we plan to deploy a moored system for observing the temperature at the middle point between the PXP pair (Nagaya et al., 1999). We plan to change this system to one that could operate on a cable, which will involve changes in the power supply and the network system for the observation data. The current seafloor ranging system is a battery-operated autonomous system. The DONET system can allow up to 3 kW (3kVDC/1A) of electric power in operation. Approximately 30 W a sensor unit on average can be distributed by using the supplied electric power from the landing station through a power distribution unit, with each observation system connected to the 100 Mbps data transmission rate (Kawaguchi et al., 2007). It will be easy to connect to the LAN system for sending observation data because a LAN system for seafloor networks has been developed on other devices (Kanazawa et al., 2007). This system uses the lithium ion battery, but we will need to change to a battery that can be charged with the DONET system. On a cable, timing precision will be maintained, by connecting to the DONET cabled seafloor network, with an accurate clock signal maintained by an external clock. If the system is connected to the DONET in the future, real-time and long-term observations would become available for the monitoring of the crustal movements across a fault on the seafloor.
6. Summary We have developed a seafloor acoustic ranging system, which can form a future application for the DONET cabled seafloor observatory for monitoring seafloor crustal movements. We performed an experiment using the seafloor acoustic ranging system on 5 August 2006. Two PXPs were deployed at a depth of
Ando, M., 1975. Source mechanisms and tectonic significance of historical earthquakes along the Nankai Trough, Japan. Tectonophysics 27 (2), 119–140. Chadwell, C.D., Spiess, F.N., 2008. Plate motion at the ridge-transform boundary of the south Cleft Segment of the Juan de Fuca Ridge from GPS–Acoustic data. Journal of Geophysical Research 113, B04415. Chadwell, C.D., Hildebrand, J.A., Spiess, F.N., Morton, J.L., Normark, W.R., Reiss, C.A., 1999. No spreading across the southern Juan de Fuca ridge axial cleft during 1994–1996. Geophysical Research Letters 26 (16), 2525–2538. Chadwick, W.W., Stapp, M., 2002. A deep-sea observatory experiment using acoustic extensometers: precise horizontal distance measurements across a mid-ocean ridge. IEEE Journal of Oceanic Engineering 27 (2), 193–201. Chadwick Jr., W.W., Milburn, H.B., Embley, R.W., 1995. Acoustic extensometer: measuring mid-ocean spreading. Sea Technology 36 (4), 33–38. Chadwick Jr., W.W., Embley, R.W., Milburn, H.B., Meinig, C., Stapp, M., 1999. Evidence for deformation associated with the 1998 eruption of Axial Volcano, Juan de Fuca Ridge, from acoustic extensometer measurements. Geophysical Research Letters 26 (23), 3441–3444. Chen, C.-T., Millero, F.J., 1977. Speed of sound in seawater at high pressures. Journal of the Acoustical Society of America 62 (5), 1129–1135. Hamamoto, H., Yamano, M., Goto, S., 2005. Heat flow measurement in shallow seas through long-term temperature monitoring. Geophysical Research Letters 32 (21), L21311, doi:10.1029/2005GL024138. Ikuta, R., Tadokoro, K., Ando, M., Okuda, T., Sugimoto, S., Takatani, K., Yada, K., Besana, G.M., 2008. A new GPS–acoustic method for measuring ocean floor crustal deformation: application to the Nankai Trough. Journal of Geophysical Research 113, B02401. Ito, Y., Obara, E., 2006. Very low earthquakes with accretionary prisms are very low stress-drop earthquakes. Geophysical Research Letters 33, L09302. Kanazawa, T., Shinohara, M., Sakai, S., Sano, O., Utada, H., Shiobara, H., Morita, Y., Yamada, T., Yamazaki, K., 2007. A new low cost ocean bottom cabled seismometers. In: Proceeding of UT07/SSC07, vol. 1, Tokyo, Japan, pp. 362–366. Kawaguchi, K., Araki, E., Kaneda, Y., 2007. A design concept of seafloor observatory network for earthquakes and tsunami. In: Proceeding of UT07/SSC07, vol. 1, Tokyo, Japan, pp. 176–179. Kido, M., Fujimoto, H., Miura, S., Osada, Y., Tsuka, K., Tabei, T., 2006. Seafloor displacement at Kumano-nada cause by the 2004 off Kii Peninsula earthquake, detected through repeated GPS/Acoustic surveys. Earth, Planets, and Space 58 (7), 911–915. MacWilliams, F.J., Sloane, N.J.A., 1976. Pseudo-random sequences and arrays. Proceedings of the IEEE 64 (12), 1715–1729. Nagaya, Y., Urabe, T., Yabuki, T., 1999. Crustal deformation observation at the crest of the EPR 18.5S with the seafloor acoustic ranging method. EOS Trans. Amer. Geophys. Union 80 (46), F1076 Fall Meeting Supplement. Osada, Y., Fujimoto, H., Miura, S., Sweeney, A., Kanazawa, T., Nakao, S., Sakai, S., Hildebrand, J.A., Chadwell, C.D., 2003. Estimation and correction for the effect of sound velocity variation on GPS/Acoustic seafloor positioning: an experiment off Hawaii Island. Earth, Planets, and Space 55, E17–E20. Park, P.O., Tsuru, T., Kodaira, S., Cummins, P.R., Kaneda, Y., 2002. Splay fault branching along the Nankai subduction zone. Science 297 (5584), 1157–1160. Spiess, F.N., Chadwell, C.D., Hildebrand, J.A., Young, L.E., Purcell Jr., G.H., Dragert, H., 1998. Precise GPS/Acoustic positioning of seafloor reference points for tectonic studies. Physics of the Earth and Planetary Interiors 108 (2), 102–112. Sweeney, A.D., Chadwell, C.D., Hildebrand, J.A., Spiess, F.N., 2005. Centimeter-level positioning of seafloor acoustic transponders from a deeply-towed interrogator. Marine Geodesy 28, 39–70. Urick, R.J., 1983. Principles of Underwater Sound, third ed. McGraw-Hill, New York, NY. Wessel, P., Smith, W.H.F., 1991. Free software helps map and display data. EOS Trans. Amer. Geophys. Union 72, 441, 445–446.
Contents lists available at ScienceDirect
Ocean Engineering journal homepage: www.elsevier.com/locate/oceaneng
Development of a seafloor acoustic ranging system toward the seafloor cable network system Yukihito Osada a,, Motoyuki Kido a, Hiromi Fujimoto a, Yoshiyuki Kaneda b a b
Graduate School of Science, Tohoku University, 6-6 Aramaki, Aoba-ku, Sendai 980-8578, Japan Japan Agency for Marine-Earth Science and Technology, 2-15 Natsushima-cho, Yokosuka 231-0061, Japan
a r t i c l e in fo
abstract
Article history: Received 20 December 2007 Accepted 15 July 2008 Available online 23 July 2008
We have developed a seafloor acoustic ranging system as a possible future application to monitoring seafloor crustal movement with the DONET (Development of Dense Ocean-floor Network System for Earthquake and Tsunami) cabled observatory system. In 2007 we carried out an experiment for the seafloor acoustic ranging system. We deployed two precision acoustic transponders (PXPs) with about 750 m spacing in Kumano-nada at a water depth of about 2035 m. We collected 660 ranging data in this one-day experiment. The round-trip travel time shows a variation with peak-to-peak amplitude of about 25 mm in the range. It was confirmed that most of the variation could be explained by the change in sound speed estimated from measured temperature and pressure. The remaining fluctuation in the acoustic measurements is 72 mm. & 2008 Elsevier Ltd. All rights reserved.
Keywords: Seafloor crustal movement Acoustic ranging Seafloor cable network system
1. Introduction The Philippine Sea Plate is subducting beneath the Eurasian Plate along the Nankai Trough, which has been known as a region of M8 class earthquakes, that have occurred with recurrence intervals of about 100–150 years (Ando, 1975). About 60 years have passed since the last earthquakes happened during 1944 and 1946. The Japanese Government’s Earthquake Research Committee estimated a 60% probability that the next earthquake in this area will occur within the coming 30 years. However, since most of the source region of the earthquakes at Nankai Trough is located beneath the ocean, an observation system is necessary to monitor tectonic activities in the offshore source region. The project DONET (Dense Ocean-floor Network System for Earthquake and Tsunami) aims to develop a real-time dense ocean floor network system around the Nankai Trough (Kawaguchi et al., 2007). This project is developing a cabled seafloor observatory with the focus on monitoring earthquakes, tsunamis and crustal movements on the seafloor. Each of 20 sites of observation is composed of precision seismometers, a pressure gauge and other instruments. All the sites would be connected to a submarine cable and deliver real-time data to land. We have developed a short-range seafloor ranging system as a possible future application to the DONET cable system. Direct acoustic ranging is a simple way to monitor local crustal deformation, and various groups have successfully detected Corresponding author.
E-mail address: [email protected] (Y. Osada). 0029-8018/$ - see front matter & 2008 Elsevier Ltd. All rights reserved. doi:10.1016/j.oceaneng.2008.07.007
deformation on the seafloor using such a system. Long-term monitoring of seafloor crustal deformation has revealed intermittent ridge extension at the Juan de Fuca Ridge (Chadwick et al., 1995, 1999; Chadwell et al., 1999; Chadwick and Stapp, 2002) and the East Pacific Rise (Nagaya et al., 1999). Direct acoustic ranging is different from GPS/Acoustic methods that involve indirect ranging to instruments on the seafloor from a vehicle towed by the ship or from a buoy on the surface (Spiess et al., 1998; Osada et al., 2003; Sweeney et al., 2005; Kido et al., 2006; Ikuta et al., 2008; Chadwell and Spiess, 2008). We aim to monitor the splay faults in the rupture area of the Tonankai earthquake in the Nankai subduction zone (Park et al., 2002). Slip along the active splay faults may be an important mechanism that accommodates the elastic strain caused by relative plate motion. Ito and Obara (2006) shows that VLF (very-low-frequency) earthquakes occur near the mega splay faults that branch from the plate boundary and have an estimated slip of 2–10 mm at maximum. We plan to deploy PXPs (precise acoustic transponder) across the splay faults in the Kumano-nada area to measure the horizontal crustal movement to the accuracy required to detect VLF earthquakes. In this study we report on the trial observations with a new direct acoustic ranging system and estimate the accuracy of the acoustic measurements.
2. Seafloor observation system Fig. 1 shows a schematic diagram of a prototype PXP. The PXP consists of a monopod anchor, a pressure housing in the shape of a
ARTICLE IN PRESS 1402
Y. Osada et al. / Ocean Engineering 35 (2008) 1401–1405
Fig. 2. Concept diagram for the seafloor acoustic ranging system, for detecting the horizontal seafloor crustal movements.
Fig. 1. Schematic of the seafloor acoustic ranging system. The system is equipped with a tilt meter, a compass, a pressure gauge, and a temperature gauge.
1700 -diameter glass ball for batteries and electronics, and an acoustic transducer on the top, and weighs 37 kg on land. This design was chosen to allow the instruments to be deployed in relatively rough seafloor terrain while keeping the transducer well above the local topography. Although such a design allows the PXP to sway slightly due to bottom currents, the internal tilt meter and compass can be used to correct these effects. This instrument costs about $40k each and is expected to have at least one year lifetime. Individual PXPs in pairs act as either a ‘‘master’’ or a ‘‘slave’’. The master can record the received signals of up to four slaves. To initiate a range measurement, the master sends out a command signal. The PXPs use a relatively simple detection method, triggering on the leading edge of an incoming wave when the amplitude rises above a fixed threshold. When the slave detects the master’s signal, it replies with an identical wave (after a precise 3 s delay). Finally, the master records the signal from the slave. This system automatically adjusts the receiver gain for the incoming signal and records the receiver gain. Fig. 2 shows the schematic diagram of the prototype seafloor acoustic ranging system, for detecting horizontal seafloor crustal movements. This system uses direct-path acoustic ranging between pairs of PXPs. The round-trip observations can cancel the effect of water currents in the ocean. This system has two advantages. One is the stability of sound speed near the deep seafloor. Most of the variations in the speed of sound result from variations in temperature, which is fairly stable in the ocean at mid-latitudes and at depths greater than 1000 m (Urick, 1983). The other main advantage is the low level of acoustic noise on the seafloor. Since underwater sound is generated mainly by winds and surface waves, the deep seafloor is an optimal place for acoustic observations. As acoustic ranging measures the travel time along the baseline between pairs of PXPs, the variation in sound speed has a direct effect on the measurement. The PXP records pressure and
Fig. 3. Map showing the location of the seafloor acoustic ranging system at the Kumano-nada. The grey diamonds show a pair of PXPs. The grey circles show PXPs for GPS/Acoustic observation on Kumano-nada (Kido et al., 2006). Lines are contours of water depth. Inset map shows tectonic setting and location of the experimental site in Japan. Black square indicates the location of the large map. Lines indicate the plate boundary; PH, the Philippine Sea Plate; EU, the Eurasian Plate; and NA, the North American Plate.
temperature data for the correction of the effects of the variation in the sound speed. We have not collected salinity data, since its influence on sound velocity is less when compared to temperature and pressure. For example, an increase in temperature by 0.003 1C or salinity by 0.009 % or a depth change of 0.9 m would cause an apparent shortening of the 1000 m baseline by about 1 cm. In the Kumano-nada area, Hamamoto et al. (2005) made heat flow measurements. The temperature variation shows a peak-to-peak amplitude of 0.2 1C during 300 days. Therefore it is important to observe the temperature with this system. Sound speed in mid-latitude waters shows a minimum value at a depth of around 1000 m and then increases with depth due to increasing pressure. Accordingly, a ray path of an acoustic wave
ARTICLE IN PRESS Y. Osada et al. / Ocean Engineering 35 (2008) 1401–1405
1403
tends to be bent upwards in the deep sea due to the gradient in the speed of sound. The PXP consists of a 2.5-m-tall monopod anchor. Although such a design allows the PXP to sway slightly due to bottom currents, the internal tilt meter and compass can be used to correct these effects.
3. Experiment with the seafloor acoustic ranging system We carried out an experiment with the seafloor acoustic ranging system on 5 August 2007. Two PXPs were deployed at a depth of 2035 m in the Kumano-nada area (Fig. 3). We were forced to deploy them from the surface. The baseline was about 750 m. This experiment was aimed at estimating the precision of acoustic ranging. We measured the two-way travel time with four types of transmitted signals at 10 min intervals. Concurrent measurements of attitude, temperature, and pressure were also carried out. Data comprising 660 measurements were collected during a period of 24 h. We examined four types of transmitted signals for precise acoustic ranging. Of the four types, three were M-sequence (maximum length sequence) signals and one was a sweep signal. The M-sequence signals, which are basically a pseudo-random sequence of pulses, were composed of sine waves with positive and negative phases. The sweep wave for the chirp signal was composed of sine waves whose frequency increased at a linear rate with time. Both the waves were found to be effective in reducing the effect of noise, when measuring the arrival time through cross-correlation (e.g., MacWilliams and Sloane, 1976). Three signals adopt a 10 kHz acoustic wave, which were coded with a 5th-order M-sequence with 31(251) bits. One bit of these transmitted signals consisted of carrier waves of two, four, and six waves; thus the length of the whole signal amounts to 6.2, 12.4, and 18.6 ms. In contrast, the sweep wave was composed of sine waves with increasing frequency between 8 and 14 kHz; thus the length of the whole signal amounts to 1.4415 ms. The master records the signal from the slave with a sampling interval of 10 ms, which corresponds to a one-way range resolution of 7.5 mm. We can resample the signal with an interval of 1 ms using cubic spline interpolation when we calculate the correlation. The master and the slave record the temperature and attitude at both ends of the distance measurement. Pressure is recorded only by the master. The information obtained with the compass and the tilt meter at each PXP are then used to estimate the attitude of each monopod.
4. Results The round-trip travel time, temperature as observed by the two PXPs, and pressure measured by the master PXP were then used to derive the range between the PXP pair. The value of the sound speed was computed from the observed pressure and temperature using the equation for the speed of sound (Chen and Millero, 1977). The ranges were corrected for instrument tilts. The time series of the observed round-trip travel times are shown in Fig. 4. Peak-to-peak variation of the travel times was 18 ms for the range measurements, which corresponds to about 25 mm (Fig. 4(a)). Fig. 4(b) shows the correlation results for the four different codes. All the received signals were stable and of high quality. Fig. 5(a) shows temperature variation observed by the master and the slave. The discrete values of the temperature are due to the resolution of the measurements. This area shows fairly stable water temperature, and the peak-to-peak variation was about 0.010 1C during this experiment. Since the temperature resolution was limited, and the water temperature was slightly different at the two devices, the temperature variation was estimated by averaging the
Fig. 4. Time series of (a) the round-trip travel time and (b) correlation. Squares show M-sequence signals with 4 waves per bit. Circles correspond to sweep signals. Triangles and diamonds show M-sequence signals with 2 and 6 waves per bit.
data with a 30-min moving window. The variation in the sound speed was estimated to be approximately 0.017 m/s during the experiment, which resulted in a variation of 18 mm in the ranges. The effect of tidal pressure variation was corrected with the observed pressure data. The peak-to-peak variation was about 1.2 dbar during this experiment. At 2035 m depth, the gradient in sound speed (inferred from measurements of 1.99 1C, 35 % at 2060 dbar) was approximately 0.008 (m/s)/dbar and the resultant range correction was up to 0.010 m. The residual variation in the round-trip travel time after the correction of these effects showed that the range across this 747 m distance was measured with a peak-to-peak range of 8 mm (Fig. 6). These results show that 70% of the variation in the ranges could be explained by variations in temperature and pressure. Next, we correct the effect of the change in attitude of the PXPs, using the data from the tilt meters and the compasses. Fig. 7 shows the time series of attitude data on each PXP. The attitude data show a change of 11 in heading and 0.11 in pitch and roll. This correction of range is up to 8 mm. The range residuals show a precision of about 2 mm, when we take into account all of the above-mentioned corrections (Fig. 8).
5. Discussion We compared the four different transmitted signals (or codes) used in this experiment to find the best suitable code. The
ARTICLE IN PRESS 1404
Y. Osada et al. / Ocean Engineering 35 (2008) 1401–1405
1.995
Temperature [degree C]
AVE Master Slave
1.99
1.985
1.98 08/05-12:00
08/06-00:00
08/06-12:00
08/06-00:00
08/06-12:00
Pressure [dbar]
2060
2059.5
2059
08/05-12:00
Time [JST] Fig. 5. Time series of temperature and pressure at each PXP. (a) Crosses and squares show the variations in temperature at the master and the slave, respectively. Black line depicts the variation of the average temperature, which was calculated by using a moving average of 30 min. (b) Time series of pressure variation at the master.
Fig. 6. Time series of ranges showing the effects of the sound speed correction. Line-plot with circles shows the sound speed-corrected ranges. Line-plot connected with crosses uses only an average sound speed. Line-plot with circles is the 5th-order M-sequence with 4 waves per 1 bit.
Fig. 7. Time series of the attitude data at each PXP (cross: master, circle: slave).
cross-correlation was more than 0.7 for all the codes used. The RMS (root mean square) deviation of the observed result with each code was 1.5 mm from the average value with the four codes. The cause of this difference may be the fluctuation of the electric mechanism that makes the transmission signals. Most received signals were coherent and had a good signal to noise ratio. Consequently, it was possible to measure the travel time with a precision of better than 1 sample (10 ms). During the deployment on 08/05 at 21:00, we recorded a change of 0.141 in the roll value. The temperature showed a similar change of 0.008 1C, as obtained from two temperature gauges. It is thought that a coherent bottom current crossed the area at that time. After 08/06 at 00:00 the two temperature gauges show slightly different changes. This was apparently a localized change, and the temperature variation was found to differ by 0.04 1C at the two PXPs. In the temperature data, about 60% of the temperature variation was coherent across the baseline of 750 m. The range residuals show a resolution of about 2 mm after corrections for the variations in temperature, pressure, and the attitude of the instruments. The good results may be partly due to the fact that we carried out the experiment under good conditions. Although the experimental site was landward of a trench, the bottom currents were weak, presumably due to the
ARTICLE IN PRESS Y. Osada et al. / Ocean Engineering 35 (2008) 1401–1405
1405
2035 m on the Kumano-nada for 24 h. The baseline length was about 750 m. We estimated the precision for acoustic ranging to be about 2 mm after correcting for the variations in the sound speed and the attitude of the instruments.
Acknowledgments This study was supported by the DONET Project of Ministry of Education, Culture, Sports, Science and Technology. The two anonymous reviewers whose comments helped improve the manuscript. The generic mapping tools (GMT) software (Wessel and Smith, 1991) was used to prepare the figures. References
Fig. 8. Time series of ranges after correcting for attitude. The line-plot is a 5thorder M-sequence with 4 waves per 1 bit.
flat-bottom topography. We hope to improve the system for achieving a precision of 1 mm or better for the range. One of the problems lies in the resolution of the temperature measurements, which currently is 0.001 1C. The temperature change around the device can be more accurately observed by introducing a temperature gauge having a better resolution and accuracy in the future. Another source of difficulty is the location of the water temperature measurements. The temperature was observed at both sides of the PXP pair for this experiment, but the temperature change along the ray path may not be accurately represented using the average temperature. In the future we plan to deploy a moored system for observing the temperature at the middle point between the PXP pair (Nagaya et al., 1999). We plan to change this system to one that could operate on a cable, which will involve changes in the power supply and the network system for the observation data. The current seafloor ranging system is a battery-operated autonomous system. The DONET system can allow up to 3 kW (3kVDC/1A) of electric power in operation. Approximately 30 W a sensor unit on average can be distributed by using the supplied electric power from the landing station through a power distribution unit, with each observation system connected to the 100 Mbps data transmission rate (Kawaguchi et al., 2007). It will be easy to connect to the LAN system for sending observation data because a LAN system for seafloor networks has been developed on other devices (Kanazawa et al., 2007). This system uses the lithium ion battery, but we will need to change to a battery that can be charged with the DONET system. On a cable, timing precision will be maintained, by connecting to the DONET cabled seafloor network, with an accurate clock signal maintained by an external clock. If the system is connected to the DONET in the future, real-time and long-term observations would become available for the monitoring of the crustal movements across a fault on the seafloor.
6. Summary We have developed a seafloor acoustic ranging system, which can form a future application for the DONET cabled seafloor observatory for monitoring seafloor crustal movements. We performed an experiment using the seafloor acoustic ranging system on 5 August 2006. Two PXPs were deployed at a depth of
Ando, M., 1975. Source mechanisms and tectonic significance of historical earthquakes along the Nankai Trough, Japan. Tectonophysics 27 (2), 119–140. Chadwell, C.D., Spiess, F.N., 2008. Plate motion at the ridge-transform boundary of the south Cleft Segment of the Juan de Fuca Ridge from GPS–Acoustic data. Journal of Geophysical Research 113, B04415. Chadwell, C.D., Hildebrand, J.A., Spiess, F.N., Morton, J.L., Normark, W.R., Reiss, C.A., 1999. No spreading across the southern Juan de Fuca ridge axial cleft during 1994–1996. Geophysical Research Letters 26 (16), 2525–2538. Chadwick, W.W., Stapp, M., 2002. A deep-sea observatory experiment using acoustic extensometers: precise horizontal distance measurements across a mid-ocean ridge. IEEE Journal of Oceanic Engineering 27 (2), 193–201. Chadwick Jr., W.W., Milburn, H.B., Embley, R.W., 1995. Acoustic extensometer: measuring mid-ocean spreading. Sea Technology 36 (4), 33–38. Chadwick Jr., W.W., Embley, R.W., Milburn, H.B., Meinig, C., Stapp, M., 1999. Evidence for deformation associated with the 1998 eruption of Axial Volcano, Juan de Fuca Ridge, from acoustic extensometer measurements. Geophysical Research Letters 26 (23), 3441–3444. Chen, C.-T., Millero, F.J., 1977. Speed of sound in seawater at high pressures. Journal of the Acoustical Society of America 62 (5), 1129–1135. Hamamoto, H., Yamano, M., Goto, S., 2005. Heat flow measurement in shallow seas through long-term temperature monitoring. Geophysical Research Letters 32 (21), L21311, doi:10.1029/2005GL024138. Ikuta, R., Tadokoro, K., Ando, M., Okuda, T., Sugimoto, S., Takatani, K., Yada, K., Besana, G.M., 2008. A new GPS–acoustic method for measuring ocean floor crustal deformation: application to the Nankai Trough. Journal of Geophysical Research 113, B02401. Ito, Y., Obara, E., 2006. Very low earthquakes with accretionary prisms are very low stress-drop earthquakes. Geophysical Research Letters 33, L09302. Kanazawa, T., Shinohara, M., Sakai, S., Sano, O., Utada, H., Shiobara, H., Morita, Y., Yamada, T., Yamazaki, K., 2007. A new low cost ocean bottom cabled seismometers. In: Proceeding of UT07/SSC07, vol. 1, Tokyo, Japan, pp. 362–366. Kawaguchi, K., Araki, E., Kaneda, Y., 2007. A design concept of seafloor observatory network for earthquakes and tsunami. In: Proceeding of UT07/SSC07, vol. 1, Tokyo, Japan, pp. 176–179. Kido, M., Fujimoto, H., Miura, S., Osada, Y., Tsuka, K., Tabei, T., 2006. Seafloor displacement at Kumano-nada cause by the 2004 off Kii Peninsula earthquake, detected through repeated GPS/Acoustic surveys. Earth, Planets, and Space 58 (7), 911–915. MacWilliams, F.J., Sloane, N.J.A., 1976. Pseudo-random sequences and arrays. Proceedings of the IEEE 64 (12), 1715–1729. Nagaya, Y., Urabe, T., Yabuki, T., 1999. Crustal deformation observation at the crest of the EPR 18.5S with the seafloor acoustic ranging method. EOS Trans. Amer. Geophys. Union 80 (46), F1076 Fall Meeting Supplement. Osada, Y., Fujimoto, H., Miura, S., Sweeney, A., Kanazawa, T., Nakao, S., Sakai, S., Hildebrand, J.A., Chadwell, C.D., 2003. Estimation and correction for the effect of sound velocity variation on GPS/Acoustic seafloor positioning: an experiment off Hawaii Island. Earth, Planets, and Space 55, E17–E20. Park, P.O., Tsuru, T., Kodaira, S., Cummins, P.R., Kaneda, Y., 2002. Splay fault branching along the Nankai subduction zone. Science 297 (5584), 1157–1160. Spiess, F.N., Chadwell, C.D., Hildebrand, J.A., Young, L.E., Purcell Jr., G.H., Dragert, H., 1998. Precise GPS/Acoustic positioning of seafloor reference points for tectonic studies. Physics of the Earth and Planetary Interiors 108 (2), 102–112. Sweeney, A.D., Chadwell, C.D., Hildebrand, J.A., Spiess, F.N., 2005. Centimeter-level positioning of seafloor acoustic transponders from a deeply-towed interrogator. Marine Geodesy 28, 39–70. Urick, R.J., 1983. Principles of Underwater Sound, third ed. McGraw-Hill, New York, NY. Wessel, P., Smith, W.H.F., 1991. Free software helps map and display data. EOS Trans. Amer. Geophys. Union 72, 441, 445–446.