• Bemis, K., , VonHerzen R. , , and Mottl M. , 1993: Geothermal heat flux from hydrothermal plumes on the Juan de Fuca Ridge. J. Geophys. Res., 98 , B4,. 63516365.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Clifford, S., 1971: Temporal-frequency spectra for a spherical wave propagating through atmospheric turbulence. J. Opt. Soc. Amer., 61 , 12851292.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Clifford, S., , and Farmer D. , 1983: Ocean flow measurements using acoustic scintillation. J. Acoust. Soc. Amer., 74 , 18261832.

  • Delaney, J., , Robigou V. , , McDuff R. , , and Tivey M. , 1992: Geology of a vigorous hydrothermal system on the Endeavour Segment, Juan de Fuca Ridge. J. Geophys. Res., 97B , 1966319682.

    • Search Google Scholar
    • Export Citation
  • Di Iorio, D., , and Farmer D. , 1998: Separation of current and sound speed in the effective refractive index for a turbulent environment using reciprocal acoustic transmission. J. Acoust. Soc. Amer., 103 , 321329.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Di Iorio, D., , and Yüce H. , 1999: Observations of Mediterranean flow into the Black Sea. J. Geophys. Res., 104 , 30913108.

  • Farmer, D., , and Clifford S. , 1986: Space–time acoustic scintillation analysis: A new technique for probing ocean flows. IEEE J. Oceanic Eng., 11 , 4250.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Farmer, D., , Clifford S. , , and Verrall J. , 1987: Scintillation structure of a turbulent tidal flow. J. Geophys. Res., 92 , 53695382.

  • Ginster, U., , and Mottl M. , 1994: Heat flux from black smokers on the Endevour and Cleft segments, Juan de Fuca Ridge. J. Geophys. Res., 99 , B3,. 49374950.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Ishimaru, A., 1978: Wave Propagation and Scattering in Random Media. Vol. 2, Academic Press, 572 pp.

  • Lavelle, J., , Baker E. , , and Massoth G. , 1998: On the calculation of total heat, salt and tracer fluxes from ocean hydrothermal vents. Deep-Sea Res., 45B , 26192636.

    • Search Google Scholar
    • Export Citation
  • Lee, R., , and Waterman A. , 1968: Space correlations of 35 GHz transmissions over a 28 km path. Radio Sci., 3 , 135139.

  • Mackenzie, K., 1981: Nine-term equation for sound speed in the oceans. J. Acoust. Soc. Amer., 70 , 807812.

  • Monin, A., , and Ozmidov R. , 1985: Turbulence in the Ocean. D. Reidel, 247 pp.

  • Munk, W., , Worcester P. , , and Wunsch C. , 1995: Ocean Acoustic Tomography. Cambridge University Press, 433 pp.

  • Ostachev, V., 1994: Sound propagation and scattering in media with random inhomogeneities of sound speed, density and medium velocity. Waves Random Media, 4 , 403428.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Ostachev, V., 1997: Acoustics in Moving Inhomogeneous Media. Thompson Science and Professional, 259 pp.

  • Özsoy, E., , DiIorio D. , , Gregg M. , , and Backhaus J. , 2001: Mixing in the Bosphorus Strait and the Black Sea continental shelf: Observations and a model of the dense water outflow. ICES J. Mar. Sci., 31 , 99135.

    • Search Google Scholar
    • Export Citation
  • Rona, R., , Bemis K. , , Silver D. , , and Jones C. , 2002: Acoustic imaging, visualization, and quantification of buoyant hydrothermal plumes in the ocean. Mar. Geophys. Res., 23 , 147168.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Schultz, A., , Delaney J. , , and McDuff R. , 1992: On the partitioning of heat flux between diffuse and point source seafloor venting. J. Geophys. Res., 97 , B9,. 1229912314.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Tatarskii, V., 1961: Wave Propagation in a Turbulent Medium. McGraw-Hill, 285 pp. (Translated from Russian by R. A. Silverman.).

  • Tatarskii, V., 1971: The Effects of the Turbulent Atmosphere on Wave Propagation. Israel Program for Scientific Translations, 472 pp.

  • Thomson, R., , Delaney J. , , McDuff R. , , Janecky D. , , and McClain J. , 1992: Physical characteristics of the Endeavour Ridge hydrothermal plume during July 1988. Earth Planet. Sci. Lett., 111 , 141154.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Thomson, R., , Mihaly S. , , Rabinovich A. , , McDuff R. , , Veirs S. , , and Stahr F. , 2003: Constrained circulation at Endeavour ridge facilitates colonization by vent larvae. Nature, 424 , 545549.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • View in gallery

    (a) Acoustic propagation from two sources and two receivers forming parallel paths and (b) sample amplitude time series showing a time delay τ between scintillation events.

  • View in gallery

    Photo of the scintillation transmitter ready for deployment in the Bosporus Strait on the North Atlantic Treaty Organization (NATO) Research Vessel (R/V) Alliance where the transducers were horizontally spaced and a vane aligned the array parallel to the main flow.

  • View in gallery

    Component diagram of the scintillation transducer.

  • View in gallery

    Block diagram for (a) transmitter and (b) receiver.

  • View in gallery

    The Endeavour segment of the Juan de Fuca Ridge showing the topography with detailed information of the main Endeavour vent field. The Ridge Multibeam Synthesis Project provided the bottom bathymetry of the northeast Pacific ridge system; the Endeavour segment geographic information system (GIS) pages provided both the swath bathymetry data of D. Kelley, University of Washington (1996, personal communication) and the characteristics of the Main Endeavour vent field originally published by Delaney et al. (1992).

  • View in gallery

    (a) Sample acoustic scintillation data and (b) the corresponding log-amplitude cross covariance for signals along two parallel paths taken from Main Endeavour vent field. The vertical flow is calculated to be W = 0.16 m s−1 at 2102 UTC yearday 199.

  • View in gallery

    (a) The vertical velocity of vent fluid obtained from the acoustic scintillation technique, together with the log-amplitude variance and effective refractive index structure parameter throughout the measurement period. Arrows denote times when the transmitter array was moved by Jason or Alvin. (b) The power spectral density of the log-amplitude variance. The principal semidiurnal tide M2 and the principal diurnal tide K1 are also shown.

  • View in gallery

    Instrument locations in the Bosporus Canyon of the Black Sea exit region.

  • View in gallery

    (a) Sample acoustic scintillation data and (b) corresponding log-amplitude cross covariances when the Mediterranean undercurrent was strong (U = 0.6 m s−1 at 0110 UTC yearday 180) and weak (U = 0.3 m s−1 at 1354 UTC yearday 178).

  • View in gallery

    Current speed from the acoustic scintillation technique, together with current meter data, where available; the log-amplitude variance, effective refractive index structure parameter, and the turbulent kinetic energy dissipation rate throughout the measurement period.

  • View in gallery

    Temperature/salinity profile taken at the south mooring location, together with an expanded scale of the Mediterranean water surrounding the transmitter (T) and receiver (R) depths. The sound speed is also included at this scale.

All Time Past Year Past 30 Days
Abstract Views 0 0 0
Full Text Views 28 28 6
PDF Downloads 11 11 0

A Self-Contained Acoustic Scintillation Instrument for Path-Averaged Measurements of Flow and Turbulence with Application to Hydrothermal Vent and Bottom Boundary Layer Dynamics

View More View Less
  • 1 Department of Marine Sciences, University of Georgia, Athens, Georgia
  • | 2 ASL Environmental Sciences, Sidney, British Columbia, Canada
© Get Permissions
Full access

Abstract

A self-contained acoustical scintillation instrument is described that has been used to measure flow and turbulence characteristics in two diverse oceanographic settings. This instrument is a battery-operated and internally logging acoustic propagation system that is ideally suited to monitor long-term flow and small-scale effective refractive index fluctuations. When the temperature variability dominates the acoustic scattering, as is the case of a hydrothermal vent plume, then a measure of the vertical buoyancy-driven flow, together with the root-mean-square temperature fluctuations, can be obtained. Results for vent structure Hulk of the Main Endeavour vent field of the Juan de Fuca Ridge show that the long-term (71 days) temperature fluctuations, together with the vertical flow, can be used to estimate heat flux density. Measurements also show oscillations in the log-amplitude variance that result from plume advection by the ambient tidal currents and demonstrate the need for a long time series measurement. When the turbulent velocity dominates the acoustic scattering, as is the case in some energetic bottom boundary layer flows, then the turbulent kinetic energy dissipation rate is derived, assuming isotropic and homogeneous models. The methodology and results are summarized from an application to the Bosporus Canyon of the Black Sea, to monitor the flow and turbulence associated with Mediterranean seawater inflow.

Corresponding author address: Daniela Di Iorio, Department of Marine Sciences, University of Georgia, 230 Marine Sciences Building, Athens, GA 30602. Email: daniela@uga.edu

Abstract

A self-contained acoustical scintillation instrument is described that has been used to measure flow and turbulence characteristics in two diverse oceanographic settings. This instrument is a battery-operated and internally logging acoustic propagation system that is ideally suited to monitor long-term flow and small-scale effective refractive index fluctuations. When the temperature variability dominates the acoustic scattering, as is the case of a hydrothermal vent plume, then a measure of the vertical buoyancy-driven flow, together with the root-mean-square temperature fluctuations, can be obtained. Results for vent structure Hulk of the Main Endeavour vent field of the Juan de Fuca Ridge show that the long-term (71 days) temperature fluctuations, together with the vertical flow, can be used to estimate heat flux density. Measurements also show oscillations in the log-amplitude variance that result from plume advection by the ambient tidal currents and demonstrate the need for a long time series measurement. When the turbulent velocity dominates the acoustic scattering, as is the case in some energetic bottom boundary layer flows, then the turbulent kinetic energy dissipation rate is derived, assuming isotropic and homogeneous models. The methodology and results are summarized from an application to the Bosporus Canyon of the Black Sea, to monitor the flow and turbulence associated with Mediterranean seawater inflow.

Corresponding author address: Daniela Di Iorio, Department of Marine Sciences, University of Georgia, 230 Marine Sciences Building, Athens, GA 30602. Email: daniela@uga.edu

1. Introduction

Acoustical methods are playing an increasingly important role in the measurement of small-scale processes in the ocean. Most of the applications to date are based on backscatter sonar in various forms; and the techniques of Doppler and echo sounder measurement of ocean currents, internal waves, surface waves, bubble fields, mixing processes, and turbulence, as well as biological phenomena, are in a state of rapid development. An alternative and less well developed technology exploits a bistatic system (transmitter and receiver separately located) to measure the influence of the medium on signals traveling along wholly refracted paths. Properties of the medium are then recovered through an inversion of the detected signals. This, of course, is also the goal of acoustic tomography (Munk et al. 1995), but here we are concerned with a related but somewhat different approach that has applications over shorter paths (100–300 m) with the use of a high acoustic frequency (307 kHz). We refer to this approach as acoustical scintillation. Many reports on these developments have appeared in the acoustic and oceanographic literature using fixed-bottom-mount frames that are cabled to shore. Here we describe a self-contained instrument that can be used in diverse deployments to emphasize its potential for oceanographic measurement.

The acoustic scintillation method is sensitive to spatial structure (the choice of acoustic frequency and range ensures that the scales within the inertial subrange of turbulence are sampled), and sampling is carried out fast enough (10–20 Hz) to obtain adequate temporal statistics. A crucial requirement of the scintillation method for mean flow measurements is that there is sufficient coherence between the signals detected at two spatially separated receivers, and, for turbulence measurements, that the acoustic scattering remains weak. While recognizing the importance of larger-scale phenomena (e.g., internal waves) in many environments, our discussion is primarily concerned with turbulent flows that are a dominant characteristic of many environments we have measured. In fact, the measurement of turbulence is a valuable application of the scintillation technique in boundary layer dynamics.

The advantage of the acoustical scintillation technique is that path-averaged properties over a baseline several times that of the turbulent eddies being measured are obtained. The space–time averaging properties of the scintillation method can be an advantage because geophysical turbulence is both patchy and intermittent. Another advantage of the acoustic method is that there are no probes that are deployed within the turbulent flow being measured and, thus, it is well suited for the hydrothermal vent application described here.

This paper reports the design and results of a self-contained in situ acoustic scintillation instrument that was used in two different oceanographic environments. The measurement concept and theoretical background describing acoustic propagation is described next in section 2. The design of the instrument is detailed in section 3. Section 4 describes the results from a deployment at the Main Endeavour vent field of the Juan de Fuca Ridge, and section 5 describes the results from a deployment in the bottom boundary layer of Mediterranean flow through the Bosporus Strait.

2. Measurement concept and theory

The concept of acoustic scintillation analysis has its origin in studies of wave (optical, radio, or acoustic) propagation through a turbulent atmosphere (Tatarskii 1961, 1971). Fluctuations in acoustical signals in the ocean have long been known, but the application of scintillation analysis methods has been more recent. As conceived, for example, by Lee and Waterman (1968) in the determination of the speed of the wind, a single transmitter, for example, a distant radio source, is detected at two locations, such that a radial velocity component of the perturbing medium is orthogonal to the two nearly parallel paths joining the transmitter with the two receivers. Refractive perturbations in the medium pass through the two paths, creating a correlation in the detected signals, the time lag of which is inversely proportional to the wind velocity component. Refractive variability will produce the perturbations, but a crucial requirement is that there be sufficient coherence between the signals detected at the two locations to allow for a useful measurement to be made.

The application of this concept to the ocean was first demonstrated in a simple set of measurements that were acquired with a sound source and pair of hydrophones towed through the water (Clifford and Farmer 1983; Farmer and Clifford 1986). Moreover, in many interesting coastal environments it is possible to fix the source-and-receiver system to the seafloor so as to provide path stability, but this is not essential and will be shown with the system described in this paper. The application of space–time coherence of the fluctuating field distinguishes this approach from most other inversions of propagation data. Analysis of these signals can provide information on the path-averaged current that is resolved perpendicular to the path and turbulent refractive index intensity and related phenomena. Our survey includes the slowly varying properties, in addition to the fluctuating properties, and is illustrated with results acquired in two very different environments.

In its simplest configuration (Fig. 1a), two sources transmit an acoustic pulse and are received at two receivers. Both transducers are perpendicular to the acoustic path and aligned with the direction of the main flow being measured. Effective refractive index perturbations passing through the parallel paths create fluctuating signals, the time-lagged cross correlation of which can be solved to determine the flow speed. The parallel path configuration is easy to use, and path-dependent weighting is essentially uniform.

More generally, we consider the Helmholtz equation for wave propagation in a moving random medium,
i1520-0426-22-10-1602-e1
where p(r) is the acoustic pressure field, n(r) is the refractive index of the medium resulting from sound speed changes from temperature and salinity, k is the acoustic wavenumber, and co is the mean sound speed; u/co is considered to be first order, and orders of u2/c2 have been neglected. This equation is a simplified form of the full linearized sound pressure equation that is obtained by Ostachev (1994), because we neglect gradients of oceanographic properties.

To solve the wave equation in (1), we make use of the weak scattering theory of Rytov outlined in Tatarskii (1971). The transformation Ψ(r) = ln p(r) = Ψ0 + Ψ1 is expressed in terms of a mean and fluctuating part, where Ψ0 corresponds to spherical wave propagation in a medium with no motion or refractive index fluctuations. The perturbation n = no + η ∼1 + η represents the refractive index from temperature and salinity variability in terms of a mean and fluctuating component.

The first-order equation becomes
i1520-0426-22-10-1602-e2
For Ψ0 = ln(Q/r eikr), the right-hand side of (2) is simplified to give −2ηeffk2, where the effective refractive index is
i1520-0426-22-10-1602-e3
where the turbulent velocity (u) is resolved along the acoustic path having unit vector n. By solving (2), the statistics for the perturbed field [Ψ1 = χ + iS, where χ = ln(A/A0) and S = ϕϕ0 are the log-amplitude and phase fluctuations, respectively] can be obtained, provided that we work with the statistics for the effective refractive index fluctuations.
The space–time cross correlation of the received log-amplitude signal, as a function of horizontal hydrophone separation (rx) and time lag (τ), may be expressed as an integral over both the acoustic path (0 < y < L) and the two-dimensional transverse refractive index wavenumber [κ = (K2x + K2z)1/2] (Tatarskii 1971; Clifford 1971; Farmer et al. 1987). Assuming isotropic and homogeneous turbulence in the plane perpendicular to the direction of acoustic propagation, the log-amplitude cross-correlation function for two parallel paths is
i1520-0426-22-10-1602-e4
where k is the acoustic wavenumber, Φeff is the three-dimensional spectrum of the effective refractive index fluctuations evaluated at Ky = 0, L is the acoustic propagation distance, and U is the current velocity perpendicular to the propagation direction, but parallel to the hydrophone separation rx. Diffraction effects are described by the sin2 term.
This equation shows that the mean current speed (U) perpendicular to the measuring baseline, which is approximately along channel, can be determined by the scintillation drift, as discussed by Farmer et al. (1987). From parallel paths (two transmitters, two receivers), and the delay to the peak of the cross correlation, the flow speed is
i1520-0426-22-10-1602-e5
where τ is the delay to the maximum.
The three-dimensional spectrum for the effective refractive index fluctuations, assuming isotropic and homogeneous turbulence (Tatarskii 1971), is
i1520-0426-22-10-1602-e6
The level of the effective refractive index fluctuations is defined by (Ostachev 1997)
i1520-0426-22-10-1602-e7
and is expressed in terms of the refractive index fluctuations arising from temperature and/or salinity variability (scalars) and those arising from the current variability (vectors). By implementing reciprocal transmission in a medium where both scalar and vector variability dominate the flow, Di Iorio and Farmer (1998) experimentally verified this mathematical relationship.
The log-amplitude variance allows measurement of C2neff through the equation
i1520-0426-22-10-1602-e8
where k is the acoustic wavenumber and L is the acoustic pathlength. Using a single transmitter and receiver (rx = 0 m), the variance measurement that is obtained is from the refractive index variability weighted at the center of the acoustic path (Farmer et al. 1987). Diffraction effects from the Fresnel scale (λL) at midpath masks the focusing ability from smaller scales along the path (Tatarskii 1971). Note that the acoustic amplitude variability cannot distinguish between finescale variability from scalars and that from current, but, as will be shown in two different oceanic applications, either the scalar or vector variations will dominate the acoustic scattering.

3. Instrument description

The scintillation instrument for flow and turbulence has been previously described in F. Rowe and D. Lemon (1992, unpublished manuscript). It consists of a transmitter and receiver as two separate, independent modules, each contained in its own pressure case and powered independently by its own battery pack. Figure 2 shows the exterior view of the transmitter module, which is externally identical to the receiver module. In this Bosporus Strait deployment the two-transducer array is mounted horizontally above the end cap; a vane, together with swivels, forced the transducer array parallel with the mean flow. For the hydrothermal vent measurement program the transducers were separated vertically to measure the buoyancy-driven flow. Because the system was intended to be deployed in a taut-line mooring, the pressure cases were enclosed within a load-bearing frame that was included as an in-line portion of the mooring.

a. Scintillation transducers

Special consideration was given to the design and fabrication of the acoustic transducers for both the transmitter and receiver. The following characteristics were achieved: a center frequency of 307 kHz with a 3-dB bandwidth of at least 30 kHz, operational at depths of 2500 m, omnidirectional in the horizontal plane with a 10° vertical beamwidth (3.5-cm toroid height), and modular so that the separation between pairs of transducers could be adjusted easily. The toroidal transducer design was chosen to provide horizontal omnidirectionality, thereby preventing any sensitivity to twisting by the mooring and removing any requirement for array alignment.

The main part of the transducer is a toroidal EC-97 piezoelectric ceramic. EC-97 material was chosen for its low sensitivity to large changes in hydrostatic pressure, which tends to shift the center frequency. The toroidal shape was chosen to obtain a uniform response in the horizontal plane, while maintaining some directivity in the vertical.

Figure 3 shows the piezoelectric ceramic mounted on an epoxy cup, covered with a thin layer of urethane. A smaller toroid of lead shot–filled urethane is mounted on the cup inside the ceramic as a sound barrier. It absorbs part of the acoustic energy radiated by the inner wall of the ceramic to prevent the formation of unwanted interference patterns. A hole in the middle of the cup permits the mounting of the transducers on a support rod; another hole permits the passage of the cable when the transducers are mounted vertically. A screw on the side of the cup is used to secure the transducer at the desired position along the rod. The transducers connect to the electronics pressure cases with the desired length of cable and a three-pin underwater connector.

b. Transmitter unit

The transmitter was designed to run automatically, without a processor or any other intelligence built into it. The transmitter operating parameters are selected via dip-switch settings. This switch bank controls the pulse width, pulse separation (or delay), and pulse cycle (or transmission/ping rate), which are directly sent to the timing controller (as shown in Fig. 4a). The timing controller contains the system master clock and generates the transmit waveforms for the power amplifier. It also controls the operation of the power amplifier and the multiplexer, which switches the power amplifier output between the two projectors. The master clock is generated by dividing a 2.46-MHz crystal by 2 to produce the master clock frequency of 1.23 MHz. The clock signal is then divided by 4 to produce the carrier frequency of 307 kHz.

The transmitter is fitted with a single-power amplifier, whose output is switched to the transducers via a field effect transistor (FET) switch multiplexer and tuning network. There is no control over the transmitter power level, which is determined by the power amplifier supply voltage. Once the timing switches have been set and the system has been turned on, the transmitter will operate until it is switched off or the battery is exhausted. In operation, the average current consumption is 20 mA. The system is powered by batter packs consisting of either 24 alkaline D cells connected in series to form a 36-V supply, or of 12 lithium double-D cells to form a 38-V supply. In either case, two packs can be connected in parallel, resulting in a 13- or 50-A-h supply, respectively. Alkaline batteries will yield a transmitter endurance of approximately 27 days, as compared to 104 days for lithium batteries.

c. Receiver unit

Figure 4b shows the main modules of the scintillation receiver. The main computer is a single-board MC68000 microprocessor computer. This computer performs all of the required computations and control functions in the system, including communications with an external computer through an RS-232 interface, control of the acoustic signal sampling and detection, data compression and recording, and power consumption and optimization. Raw amplitude, time of arrival, and phase data, together with system time, are recorded onto a 33-Mbyte recorder that is implemented with highly reliable flash Erasable Programmable Read-Only Memory (EPROM) chips. Power is supplied by up to two packs of lithium or alkaline batteries in the same configuration as those that are used in the transmitter. To conserve battery power and extend the data collection time, the system has the capability to power down between sampling intervals. Either an alarm in the real-time clock or the arrival of a character through the RS-232 interface will wake the system.

1) Hardware

The tuning circuitry adjusts the frequency response of the transducer so that it is centered at the operating frequency of 307 kHz, with the required bandwidth of 30 kHz. The hard-limiter receiver is a two-stage amplifier. The first stage provides a fixed 20-dB gain and is optimized for low noise. This signal is further processed for the phase measurement. It is filtered with a fourth-order LC bandpass filter centered on the carrier to remove higher harmonics of the carrier frequency. In-phase (I) and quadrature (Q) components are then obtained by sampling a 1/4-carrier cycle apart every four cycles (150-kHz digitization rate). The phase is then defined as the arctan (Q/I).

The second-stage amplifier is implemented with an intermediate frequency (IF) frequency modulation (FM) receiver with a dynamic range of 80–90 dB, which provides a hard-limited signal output with a voltage output that is proportional to the logarithm of the input signal amplitude. A two-pole RC low-pass filter is then applied to remove the carrier frequency. To increase the amplitude resolution prior to digitization, a fixed voltage is subtracted from the log-amplitude signal and then the difference is amplified by a factor of 5. This scheme permitted a 0.08-dB resolution in the sampling of the signal amplitude. The offset voltage is periodically adjusted (from the stored statistics of the amplitude) to compensate for long-term fluctuations. Digitization is carried out with an 8-bit analog/digital (A/D) converter with one sample every two cycles (150-kHz digitization rate). This sampling rate is essentially twice that of the phase. Over the width of the arrival peak the phase is slowly varying and, thus, is permitted a reduced sampling rate.

A timing circuit provides sampling windows that can be moved in time under control of the computer, which would permit acquiring samples only around the times of arrival of the pulses. Two sampling windows are provided that correspond to the signal from the two transmitters. The computer controls the spacing between the windows that is set by the transmitter pulse separation and the pulse cycle, as well as their size determined by the pulse width (recall Fig. 4a). Within each window are sampling strobes for triggering the A/D conversions. Samples within a window are stored automatically in a bank of first-in/first-out (FIFO) memory, which has sufficient capacity to store up to 512 samples of amplitude and phase for both receiver channels. Once full, an interrupt is signaled to the computer to begin processing the signals for the peak amplitude, phase, and arrival time.

2) Software

The operating software consists of a number of processes that run in a sequential manner and can operate in manual or stand-alone mode. The processes are synchronized by the real-time operating system (RTOS), which gives highest priority to signal locking and data acquisition. The software was written in C with some Motorola MC68000 assembler language. The operating software consists of four main processes: communications (manual mode), control (stand-alone mode), signal locking and data acquisition, and data storage.

In manual mode the system is under user control with commands from the RS-232 port for setting system parameters and testing system functions. In the stand-alone mode of operation, the system is run by the control process. The control process is capable of continuous data acquisition or can be programmed to go to sleep to save power and wake up at specific time intervals to collect the acoustic data for a short period of time. As well as saving power it allows subsampling of the data over a longer period. When the system wakes up from sleep mode or on power up, it reads the system parameters that were set by the user, checks the clock, and compares with the data acquisition start time. If the time has been reached the process tells the signal-locking and data acquisition process to lock on the acoustic signal. If the acquisition time has not been reached the control programs the real-time clock to wake up 5 min later and then puts the system back to sleep.

Detection of the acoustic signals is accomplished by the variable gain amplifier, which adjusts the offset and gain applied to the signal before digitization, as well as the position and width of the two sampling windows. For a given pulse cycle (transmission rate) a dead zone of half the pulse cycle exists where the signal is within the noise level. The dead zone is detected when 15 consecutive windows of 500 samples (a 10-Hz pulse cycle digitized at 150 kHz requires a sampling window of 15 000 samples) have been detected without a signal. Once the 15 windows have been detected the window is shifted in time by 500 samples until a signal is detected. This signal will be the first acoustic pulse in the transmission. If a signal is found before 15 consecutive windows have been found then the counter is reset to zero and scanning is continued. If the dead zone is not found after shifting the window through out the pulse cycle (i.e., more than 30 windows), then the gain setting may be too low and the system raises the gain offset and then tries again to locate the dead zone.

Once the first pulse of the transmission has been located the sample window is narrowed to 60 samples over each signal. A quadratic fit of the signal amplitude is performed to compute the maximum peak amplitude and the position of the peak to give arrival time. The phase value is obtained at the nearest integral arrival time of the phase signal. The mean amplitude and arrival time statistics are monitored so that both the gain and window positions can be adjusted, thus, remaining locked on the acoustic signal. Because the parallel path configuration gives uniform weight on the flow measurement, only those signals were recorded (i.e., signals from T1 to R1 and signals from T2 to R2 in Fig. 4. This also maximizes data storage.

The data storage process consists of a data compression algorithm that was developed specifically to maximize the quantity of raw data stored. The compression algorithm is based on calculating successive differences on a sequence of data and then subtracting the minimum difference from each data value. The initial starting value and the minimum difference are stored in a 26-bit header. A sequence of samples is complete when either a maximum of 127 samples has been used or when a difference value requiring more than four bits is found. This algorithm achieved a compression ratio of approximately (127 × 4 + 26)/127 × 8 bit = 52.5%. After compression, the storage process writes to flash EPROM recording cards. The storage process also has the ability to transmit compressed or uncompressed data over the RS-232 line to a separate personal computer (PC) for testing or real-time applications.

4. Endeavour hydrothermal vent field

Special considerations were made in designing the moorings for deployment in the Main Endeavour vent field of the Juan de Fuca Ridge, because they needed to be small and light enough so that they could be maneuvered into position by Deep Submergence Vehicle (DSV) Alvin. All of the pressure housings were tested 10 times at a pressure of 1.5 times that encountered during the dive, and syntactic foam was used for flotation. Galvanized rings were included in the mooring line above and below the acoustic release and above the anchor as attachment points for Alvin’s manipulator. The net weight of the mooring in water was 36 kg, which was well within Alvin’s handling capabilities.

The system was deployed from Canadian Coast Guard Ship (CCGS) Parizeau during a cruise to the Juan de Fuca Ridge in July 1991. The moorings were deployed by free fall to the drop locations listed in Table 1 and located by Alvin and maneuvered into place on the positions shown in Fig. 5. The deployment period was from mid-July to the beginning of October 1991—a total of 71 days. During this time, the transmitter was snagged twice by remotely operated vehicle (ROV) Jason over the period between 9 and 12 August and was moved. During this time the system failed several times in locking and tracking the received acoustic signals. On 19 September Alvin tried to move it back close to its original position.

Figure 5 shows detailed information of the Main Endeavour vent field as described by Delaney et al. (1992). The active black smoker vent, identified as “Hulk,” or 8F, is the northernmost structure of the Main Endeavour vent field and is where the scintillation transmitter and receiver formed a 95-m acoustic path through the buoyant plume 25 m above the vent orifice. The acoustic path (38°T) was somewhat parallel to the valley sidewall where the semidiurnal lunar (M2) tidal currents are aligned approximately along ridge (20°T) because of topographic steering by the valley walls (Thomson et al. 2003). As a result, any displacement of the plume resulting from tidal currents would occur mostly along the acoustic path. Acoustic imaging of black smoker–type buoyant plumes on the east Pacific rise (EPR) (Rona et al. 2002) showed a centerline offset of 5 m at 25 m above the chimney for a horizontal current of 5 cm s−1.

The system was programmed for burst sampling with 6-min sampling at 10 Hz each hour. The hourly repetition was frequent enough to resolve tidal frequency variations that were caused by ambient currents, and the 6-min sampling was expected to be long enough to average over the high-frequency turbulent refractive index. Table 1 summarizes the scintillation experimental parameters.

Sample acoustic scintillation data are presented in Fig. 6 with the corresponding log-amplitude cross-covariance function, which shows a vertical velocity of 16 cm s−1. The phase difference between the two parallel paths shows high-frequency oscillations that are probably associated with mooring vibrational motion at a dominant frequency of 0.6 Hz. These phase oscillations are approximately 1 rad (5.2-μs travel time variation), and clearly have no effect on the amplitude fluctuations. Mooring strumming was reduced by fitting a fairing consisting of nylon tubing (zipped over the line) with 12.5-cm-long by 12.5-mm-wide nylon streamers attached. This fairing acted to break up vortices that were shed from the cable, and, thus, prevented the build up of significant strum amplitude.

The current speed is calculated from (5), because the time delay of the log-amplitude cross covariance is measured using 3-min (1800 samples) segments of data. During the 6-min burst two measurements of flow and log-amplitude variance are obtained and, thus, averaged together. The hourly results are shown Fig. 7a. Over the 71-day measurement period the vertical velocity remains fairly constant with a mean of 15.6 cm s−1 and a standard deviation of 4.0 cm s−1. These values are consistent with the observations of Bemis et al. (1993) that were made at similar heights using an array of thermistors and flowmeters. The arrows that are marked on the figure indicate times when the ROV Jason snagged the transmitter array and when the DSV Alvin had tried to reposition the mooring.

The log-amplitude variance that is shown, scaled to also give the level of the effective refractive index fluctuations, is within the weak scattering regime (σ2χ ≪ 0.25) and shows much variability. The level of the effective refractive index fluctuations C2neff from (8) is a factor of 20 greater than those observed for boundary layer turbulent flow discussed subsequently. Using the empirical equation of sound velocity from Mackenzie (1981), we can approximate the contributions of temperature (T), salinity (S), and current speed (w) to the refractive index variance of the flow:
i1520-0426-22-10-1602-e9
i1520-0426-22-10-1602-e10
i1520-0426-22-10-1602-e11
where co = 1495 m s−1 is the path-averaged sound speed, σT ∼ 0.1°C and σS ∼ 0.01 psu are the anomalies observed by Thomson et al. (1992), and σw ∼0.04 m s−1 is based on our measurements of the buoyancy-driven flow. On the basis of these numbers, we expect temperature variability to be the main source of forward acoustic scatter.
The level of the effective refractive index fluctuations that are dominated by temperature can be related to the refractive index variance by temperature fluctuations and the outer scale of turbulence (Lo) (see Ishimaru 1978, e.g.). If the outer scale is approximated as twice the standard deviation of the plume radius, which has a Gaussian temperature distribution (Bemis et al. 1993)
i1520-0426-22-10-1602-e12
where r is the horizontal radius of the plume, z is the height above the vent, the constant 0.12 is the expansion rate of the plume, and Tc is in terms of the heat flux at the centerline, then the effective structure parameter is
i1520-0426-22-10-1602-e13
The root-mean-square temperature fluctuations that are evaluated in this way are indeed of the order of 0.1°C at 25 m above the vent orifice, which is consistent with the maximum plume anomalies relative to the background waters that were observed by Thomson et al. (1992). If σTθpθb represents the plume-induced thermal anomaly (difference between plume and background waters at the same level), then according to Lavelle et al. (1998) when σT is integrated over a volume, all of the heat released is accounted for.

Given that the vertical velocity is a path-averaged measurement with essentially uniform weighting along the path, the space- and time-averaged vertical heat flux density is approximated as FT = 〈w〉 〈σTρcp = 0.06 MW m−2, where ρcp = 4.3 × 106 J °C−1 m−3 for seawater. Heat flux density measurements of 0.13 MW m−2 at vent structure 8F of the Main Endeavour field have been obtained by Ginster and Mottl (1994). Because our σT is influenced by the horizontal variations in plume temperatures (i.e., θb may not the background waters), an underestimate of the heat flux is expected.

Because the log-amplitude signal shows much variability, a power spectral density was computed and is shown in Fig. 7b. A dominant peak at the M2 tidal frequency, together with a weaker peak at the principal diurnal (K1) tidal frequency, is clearly shown. The spectral content of the vertical flow (not shown) has no tidal variations and appears to have a nearly white distribution, similar to that obtained by Schultz et al. (1992) for diffuse venting. The interaction of the buoyant plume with the tides remains a complex issue and is still not fully understood because of a lack of long-term measurements in the horizontal and vertical flow of the buoyant plume, together with temperature fluctuations. Here we present two possible explanations for the observed tidal variability in the acoustic scattering and, hence, temperature fluctuations.

First, the tidal currents cause a horizontal deflection of the hydrothermal plume along the valley (which is approximately along the acoustic path), and so the plume centerline presumably moves around relative to the acoustic midpath point. The log-amplitude variance is most sensitive to fluctuations at the center of the acoustic path (as discussed previously), and the low log-amplitude variances that occur may correspond to the plume moving out of the acoustic beam and/or may correspond to the edge of the plume being sampled acoustically. The velocity measurement is essentially a uniform path-averaged measurement, and so the plume moving along the acoustic path does not affect the measurement, even though the velocity is also approximated as having a horizontal Gaussian distribution (Bemis et al. 1993). Second, the horizontal currents cause variations in the entrainment rate of ambient fluids into the buoyant plume. Stronger horizontal flow causes greater entrainment and, hence, stronger acoustic scatter; weaker flows cause weaker acoustic variability presumably because of the lower entrainment rates. The amount of entrained fluid, however, may not cause a significant change in vertical flow as measured by the scintillation system.

5. Bosporus Strait bottom boundary dynamics

The scintillation system described was also deployed within the Bosporus canyon, which consists of Mediterranean seawater flowing into the Black Sea (Di Iorio and Yüce 1999; Özsoy et al. 2001). This location is where the Strait of Istanbul (Bosporus) meets the Black Sea continental shelf (see Fig. 8). The transmitter array was deployed on the eastern side, and the receiver array on the western side was separated by 369 m at a depth of 62.5 and 65.5 m, respectively. Other moored instrumentation is shown, but only the south mooring will be discussed because it supports the acoustic scintillation current measurement.

The acoustic transducers were horizontally separated by 0.2 m. To keep the transducer array aligned in the direction of flow, a vane and swivel were attached to the mooring (cf. Fig. 2). Because both the transmitter and receiver modules do not have compasses, and because the diverging paths were not logged, the mooring orientation and angle of arrival, respectively, were not measured. The vane allowed measurement of the current flow perpendicular to the acoustic axis and measurement of turbulent structures as they were advected past the acoustic paths. Table 2 summarizes the experimental parameters that were used during this 1996 deployment, which differ from a 1995 deployment described in Di Iorio and Yüce (1999).

To calculate the amplitude, phase, and travel time of the acoustic pulse, it is crucial that there is path separation at the receiver. With a source at 62.5 m and a receiver at 65.5-m depth only two paths are possible, given the directivity of the transducers—a direct path and a bottom-reflected path. Because the time separation between the two paths is 0.45 ms for a distance of 369 m, and with the chosen pulse width of 0.1 ms, path separation is achieved. At times the Mediterranean flow is arrested and blockage occurs because of the 60-m sill [acoustic Doppler current profiler (ADCP) location]. When this occurs, the interface would drop below the transmitter and receiver depth and acoustic alignment and path separation may not be possible. For the data described here no blockage events were observed, and the Mediterranean flow was continuous during this 5.5-day deployment.

Sample acoustic scintillation data (approximately 2 min) are shown in Fig. 9a during a time of strong (U = 0.6 m s−1) and weak (U = 0.3 m s−1) flow. Much high-frequency variability in the amplitude exists during strong flow because of the turbulent nature of the flow. The travel time difference between the two parallel paths shows a periodic oscillation as a result of mooring rotational motion affecting the pathlength. When the current weakens the acoustic amplitude variability diminishes as a result of reduced turbulent levels; the travel time difference oscillations are also reduced. The relative pathlength change that results in a 0.1-ms phase oscillation is of the order of 0.15 m, which has a negligible effect on amplitude variations.

The statistics for the normalized log-amplitude χ = ln(A/〈A〉) (where A is the acoustic amplitude and 〈 〉 denotes a time average) allow measurement of oceanographic parameters, as described by Di Iorio and Yüce (1999) during a 1995 measurement campaign. For example, Fig. 9b shows the time-lagged cross covariance between the two log-amplitude signals along parallel paths. The time lag that is observed is a result of the translation of turbulent structures parallel to the transmitter–receiver array having length rx = 0.2 m. The oscillation frequency for the cross covariance is f = U/λL, which corresponds to Fresnel length scales (1.3 m) that are advected by the mean flow.

The current speed is calculated using (5), because the time delay of the log-amplitude cross covariance is measured using 1 min (600 samples) of log-amplitude data. Any directional orientation of the arrays such that the transmitter and receiver array are not parallel would result in a decrease in path separation rx in Eq. (5), and, thus, our measurement with rx = 0.2 m could overestimate the current speed. The mean travel time difference shown in Fig. 9, however, is approximately 0 ms, which implies that the transmitter and receiver arrays are, on average, parallel, giving the mean flow that is resolved along the transducer array direction, which is assumed to be predominantly along channel. The flow result is shown Fig. 10, together with available along-channel current meter measurements (resolved along 30°T) at a depth of 64 m for comparison. After 1.5 days the meter failed because of clogging of the rotor and vane. Over the 5.5-day measurement period much variability exists in the flow. In fact, the acoustic measurement sensitivity can be improved by averaging over 5 min. The log-amplitude variance that is shown is within the weak scattering regime (σ2χ ≪ 0.25). The measurement of C2neff from (8) is also shown in Fig. 10.

Independent measurements of the finescale temperature/salinity structure were not obtained during this experimental program, but previous measurements in this area (Di Iorio and Yüce 1999), however, showed that turbulent velocity dominated the acoustic scattering because the refractive index variations from the temperature/salinity were at most 10% of the effective refractive index fluctuations. In this previous study the Mediterranean inflow was, on average, 60 cm s−1 during the fine-structure measurements which is similar to the maximum flow conditions that are observed on yearday 180 in this study. Also, previous measurements over a 24-h period showed a freshening of the Mediterranean water resulting from entrainment of overlying Black Sea intermediate water. The increased separation of the isohalines (and isotherms) at the interfacial zone, however, remained above 60 m and is highly dependent on what happens upstream within the Bosporus Strait. Figure 11 shows a profile of the temperature and salinity together with an expanded view of the data surrounding the depth of the transmitter (T) and receiver (R), including the sound speed with no pressure effects.

The sound speed variations about a 1-m bin-averaged profile are, at most, σc ∼0.03 m s−1, which corresponds to 〈n2c〉 ≈ (σc/c0)2 ≈ 4.0 × 10−10, whereas the observed velocity fluctuations are of the order of σu = 0.06 m s−1, which corresponds to 〈n2u〉 ≈ 1.6 × 10−9 for a given mean sound speed c0 = 1504.8 m s−1. Based on these numbers we can assume that temperature/salinity variations will have a weaker effect on acoustic forward scatter. Also, because the log-amplitude variance and, hence, the level of the effective refractive index variability varies according to the increasing and decreasing current, it can be assumed that velocity variations dominate the acoustic scattering [C2neff = (11/6)C2nυ], leading to an estimate of the turbulent kinetic energy dissipation rate determined via
i1520-0426-22-10-1602-e14
Acoustic measurements of ε in Fig. 10 range from 4 × 10−6 to 1 × 10−4 W kg−1. Assuming that the measurements are made in a constant stress layer, the bottom drag coefficient is determined from (Monin and Ozmidov 1985)
i1520-0426-22-10-1602-e15
where κ = 0.4 is von Kármán’s constant and z = 8 m distance from the bottom boundary. Measurements range from 6 × 10−3 to 13 × 10−3. These values are consistent with the measurements made during the 1995 experimental program.

6. Conclusions and future applications

A self-contained acoustic scintillation instrument is described that can be used to study hydrothermal vent flow characteristics and bottom boundary layer dynamics over an extended period of time. For the hydrothermal vent flow the system described is perfectly suited to sample the flow characteristics within the plume without having any sensors in the plume itself because the acoustic amplitude variability is path averaged with greatest sensitivity to eddies at the center of the acoustic path. In this environment, the temperature fluctuations will dominate the acoustic scattering leading to estimates of the temperature variance, which are periodic with the ambient tidal currents resulting from either plume advection or changes in entrainment of ambient fluids into the buoyant plume. Given the mean vertical flow, together with the temperature anomaly for the plume, an estimate of the heat flux is obtained. For energetic bottom boundary layer dynamics the turbulent velocity almost always dominates the acoustic scattering, and so the acoustic scintillation measurement can approximate the turbulent kinetic energy dissipation rate that is averaged along the acoustic path.

The measurement concept relies upon the flow field itself to advect turbulent eddies across the acoustic path, and requires persistence of the turbulent structures to get a useful coherence in the measured signals at two hydrophones for flow measurements. The turbulent-scale sensitivity is dependent on the acoustic frequency and pathlength, which ranged from 0.7 (hydrothermal vent) to 1.3 (bottom boundary layer) m, and their position along the acoustic path. With appropriate models for the effective refractive index fluctuations (whether dominated by turbulent velocity or temperature variations) the amplitude variability can be inverted to obtain the small-scale eddy characteristics.

A long time series measurement can be made at a reduced cost in terms of ship and/or submersible time and personnel. This moored system can be deployed for up to 1 month using alkaline cells (3 times longer with lithium batteries) and the burst sampling mode. Such a long time series measurement is ideally suited for seafloor and coastal observatory initiatives.

Acknowledgments

The authors wish to thank the technical staff of RD Instruments between 1989 and 1992 that was involved in the development of this prototype scintillation instrument through an SBIR grant by ONR. We are grateful to Drs. J. R. Delaney and R. E. MacDuff for generously providing dive time on the DSV Alvin to deploy the instruments on Endeavour Ridge, and to them and Dr. V. Robigou for their advice and assistance through out the hydrothermal vent project. Special thanks are due to Dr. R. E. Thomson for his advice and encouragement and for making ship time available to deploy and recover the instrumentation on CCGS Parizeau. Thanks to Dr. K. Payne for producing the Endeavour vent field figure. Also, special thanks to Dr. D. M. Farmer who donated the instrument to the first author for the Bosporus Strait research. SACLANT Undersea Research Centre supported the Bosporus research while Di Iorio was employed with them. Special thanks to the technical staff of SACLANTCEN and, in particular, Dr. T. Akal, engineer P. Guerrini, and R. Della Maggiore who assisted with the mooring design, deployment, and recovery along with the ship crew on the NATO R/V Alliance.

REFERENCES

  • Bemis, K., , VonHerzen R. , , and Mottl M. , 1993: Geothermal heat flux from hydrothermal plumes on the Juan de Fuca Ridge. J. Geophys. Res., 98 , B4,. 63516365.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Clifford, S., 1971: Temporal-frequency spectra for a spherical wave propagating through atmospheric turbulence. J. Opt. Soc. Amer., 61 , 12851292.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Clifford, S., , and Farmer D. , 1983: Ocean flow measurements using acoustic scintillation. J. Acoust. Soc. Amer., 74 , 18261832.

  • Delaney, J., , Robigou V. , , McDuff R. , , and Tivey M. , 1992: Geology of a vigorous hydrothermal system on the Endeavour Segment, Juan de Fuca Ridge. J. Geophys. Res., 97B , 1966319682.

    • Search Google Scholar
    • Export Citation
  • Di Iorio, D., , and Farmer D. , 1998: Separation of current and sound speed in the effective refractive index for a turbulent environment using reciprocal acoustic transmission. J. Acoust. Soc. Amer., 103 , 321329.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Di Iorio, D., , and Yüce H. , 1999: Observations of Mediterranean flow into the Black Sea. J. Geophys. Res., 104 , 30913108.

  • Farmer, D., , and Clifford S. , 1986: Space–time acoustic scintillation analysis: A new technique for probing ocean flows. IEEE J. Oceanic Eng., 11 , 4250.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Farmer, D., , Clifford S. , , and Verrall J. , 1987: Scintillation structure of a turbulent tidal flow. J. Geophys. Res., 92 , 53695382.

  • Ginster, U., , and Mottl M. , 1994: Heat flux from black smokers on the Endevour and Cleft segments, Juan de Fuca Ridge. J. Geophys. Res., 99 , B3,. 49374950.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Ishimaru, A., 1978: Wave Propagation and Scattering in Random Media. Vol. 2, Academic Press, 572 pp.

  • Lavelle, J., , Baker E. , , and Massoth G. , 1998: On the calculation of total heat, salt and tracer fluxes from ocean hydrothermal vents. Deep-Sea Res., 45B , 26192636.

    • Search Google Scholar
    • Export Citation
  • Lee, R., , and Waterman A. , 1968: Space correlations of 35 GHz transmissions over a 28 km path. Radio Sci., 3 , 135139.

  • Mackenzie, K., 1981: Nine-term equation for sound speed in the oceans. J. Acoust. Soc. Amer., 70 , 807812.

  • Monin, A., , and Ozmidov R. , 1985: Turbulence in the Ocean. D. Reidel, 247 pp.

  • Munk, W., , Worcester P. , , and Wunsch C. , 1995: Ocean Acoustic Tomography. Cambridge University Press, 433 pp.

  • Ostachev, V., 1994: Sound propagation and scattering in media with random inhomogeneities of sound speed, density and medium velocity. Waves Random Media, 4 , 403428.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Ostachev, V., 1997: Acoustics in Moving Inhomogeneous Media. Thompson Science and Professional, 259 pp.

  • Özsoy, E., , DiIorio D. , , Gregg M. , , and Backhaus J. , 2001: Mixing in the Bosphorus Strait and the Black Sea continental shelf: Observations and a model of the dense water outflow. ICES J. Mar. Sci., 31 , 99135.

    • Search Google Scholar
    • Export Citation
  • Rona, R., , Bemis K. , , Silver D. , , and Jones C. , 2002: Acoustic imaging, visualization, and quantification of buoyant hydrothermal plumes in the ocean. Mar. Geophys. Res., 23 , 147168.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Schultz, A., , Delaney J. , , and McDuff R. , 1992: On the partitioning of heat flux between diffuse and point source seafloor venting. J. Geophys. Res., 97 , B9,. 1229912314.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Tatarskii, V., 1961: Wave Propagation in a Turbulent Medium. McGraw-Hill, 285 pp. (Translated from Russian by R. A. Silverman.).

  • Tatarskii, V., 1971: The Effects of the Turbulent Atmosphere on Wave Propagation. Israel Program for Scientific Translations, 472 pp.

  • Thomson, R., , Delaney J. , , McDuff R. , , Janecky D. , , and McClain J. , 1992: Physical characteristics of the Endeavour Ridge hydrothermal plume during July 1988. Earth Planet. Sci. Lett., 111 , 141154.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Thomson, R., , Mihaly S. , , Rabinovich A. , , McDuff R. , , Veirs S. , , and Stahr F. , 2003: Constrained circulation at Endeavour ridge facilitates colonization by vent larvae. Nature, 424 , 545549.

    • Crossref
    • Search Google Scholar
    • Export Citation

Fig. 1.
Fig. 1.

(a) Acoustic propagation from two sources and two receivers forming parallel paths and (b) sample amplitude time series showing a time delay τ between scintillation events.

Citation: Journal of Atmospheric and Oceanic Technology 22, 10; 10.1175/JTECH1799.1

Fig. 2.
Fig. 2.

Photo of the scintillation transmitter ready for deployment in the Bosporus Strait on the North Atlantic Treaty Organization (NATO) Research Vessel (R/V) Alliance where the transducers were horizontally spaced and a vane aligned the array parallel to the main flow.

Citation: Journal of Atmospheric and Oceanic Technology 22, 10; 10.1175/JTECH1799.1

Fig. 3.
Fig. 3.

Component diagram of the scintillation transducer.

Citation: Journal of Atmospheric and Oceanic Technology 22, 10; 10.1175/JTECH1799.1

Fig. 4.
Fig. 4.

Block diagram for (a) transmitter and (b) receiver.

Citation: Journal of Atmospheric and Oceanic Technology 22, 10; 10.1175/JTECH1799.1

Fig. 5.
Fig. 5.

The Endeavour segment of the Juan de Fuca Ridge showing the topography with detailed information of the main Endeavour vent field. The Ridge Multibeam Synthesis Project provided the bottom bathymetry of the northeast Pacific ridge system; the Endeavour segment geographic information system (GIS) pages provided both the swath bathymetry data of D. Kelley, University of Washington (1996, personal communication) and the characteristics of the Main Endeavour vent field originally published by Delaney et al. (1992).

Citation: Journal of Atmospheric and Oceanic Technology 22, 10; 10.1175/JTECH1799.1

Fig. 6.
Fig. 6.

(a) Sample acoustic scintillation data and (b) the corresponding log-amplitude cross covariance for signals along two parallel paths taken from Main Endeavour vent field. The vertical flow is calculated to be W = 0.16 m s−1 at 2102 UTC yearday 199.

Citation: Journal of Atmospheric and Oceanic Technology 22, 10; 10.1175/JTECH1799.1

Fig. 7.
Fig. 7.

(a) The vertical velocity of vent fluid obtained from the acoustic scintillation technique, together with the log-amplitude variance and effective refractive index structure parameter throughout the measurement period. Arrows denote times when the transmitter array was moved by Jason or Alvin. (b) The power spectral density of the log-amplitude variance. The principal semidiurnal tide M2 and the principal diurnal tide K1 are also shown.

Citation: Journal of Atmospheric and Oceanic Technology 22, 10; 10.1175/JTECH1799.1

Fig. 8.
Fig. 8.

Instrument locations in the Bosporus Canyon of the Black Sea exit region.

Citation: Journal of Atmospheric and Oceanic Technology 22, 10; 10.1175/JTECH1799.1

Fig. 9.
Fig. 9.

(a) Sample acoustic scintillation data and (b) corresponding log-amplitude cross covariances when the Mediterranean undercurrent was strong (U = 0.6 m s−1 at 0110 UTC yearday 180) and weak (U = 0.3 m s−1 at 1354 UTC yearday 178).

Citation: Journal of Atmospheric and Oceanic Technology 22, 10; 10.1175/JTECH1799.1

Fig. 10.
Fig. 10.

Current speed from the acoustic scintillation technique, together with current meter data, where available; the log-amplitude variance, effective refractive index structure parameter, and the turbulent kinetic energy dissipation rate throughout the measurement period.

Citation: Journal of Atmospheric and Oceanic Technology 22, 10; 10.1175/JTECH1799.1

Fig. 11.
Fig. 11.

Temperature/salinity profile taken at the south mooring location, together with an expanded scale of the Mediterranean water surrounding the transmitter (T) and receiver (R) depths. The sound speed is also included at this scale.

Citation: Journal of Atmospheric and Oceanic Technology 22, 10; 10.1175/JTECH1799.1

Table 1.

Acoustic scintillation instrument parameters for the Endeavour hydrothermal vent study.

Table 1.
Table 2.

Acoustic scintillation instrument parameters for the Bosporus canyon bottom boundary study.

Table 2.
Save