Hydrographic Doppler Sonar System
(HDSS)

Background

Doppler sonars with a capability to profile to depths approaching 1 km were developed at Scripps in the late 1970's and first used operationally from the Research Platform FLIP in 1980. Commercially available Doppler sonars (ADCPs) are now on most ships in the research fleet. It is natural to attempt to extend sonar velocity and shear profiling to greater depths. This is an uphill battle. The typical velocity of baroclinic ocean currents decreases with depth, a consequence of decreasing density gradient. Expected current speeds at kilometer depths are of order 1-2 cm s-1, less than the (short-term) error in the best ship navigation systems. With a ship moving at 12 kts, a bias in the velocity estimate as small as 1 part in 600 can obscure the natural signal. The primary value in deep Doppler systems is in exploring a-typical sites and situations, where large deep currents and/or shears might be expected. Deep-sea topography plays a key role here. Energetic boundary currents are often closely coupled to topography. Inertial and internal waves have a virtually unexplored geography, with topographic wave generation being a major unknown. In these situations, relatively small vertical scale motions are expected in close proximity to the sea-floor. Thus, precise depth resolution and careful control of sonar beam side lobes are even more critical for long-range sonars than for their short-range counterparts.

hdss under Rev ship

Figure 1. A view of the HDSS installation on the R/V Revelle,
showing the 50 kHz (brown) and 140 kHz (yellow) starboard forward transducers.
The 50 kHz units have since been replaced by improved transducers.

In the mid 1990’s, as part of a NSF Facilities and Infrastructure Grant, funds were obtained to develop a long-range Doppler sonar for permanent installation on the R/V Roger Revelle (Figure 1). A dual-frequency (50-140 kHz) system was designed and constructed at Scripps and installed in December 1999. Termed the Hydrographic Doppler Sonar System (HDSS), it has been operating continuously, with incremental improvements, since January 2000. The sonar has been used in the NSF experiments HOME, EPIC, IWAP and SWIR as well as the ONR Sea of Japan, ASIAEX, NLIWI, and AESOP experiments (Figures 3,4).

Both HDSS sonars are configured in the conventional 4-beam Janus geometry. In plan, the beams are oriented 45° relative to the fore/aft axis of the ship’s hull. The depression angle of each beam relative to horizontal is 60 degrees. The HDSS sonars transmit through a protective polyethelene "window" mounted flush with the ships hull.

PADS array Ver.1

Figure 2. One of four second-generation 50 kHz arrays presently installed in the Revelle. The transmitter (upper) consists of ten hexagonal transducers, each containing 19 individual tonpiltz elements. The receiver (lower) is a cluster of seven transducers. The center element is recorded independently, enabling a modest degree of shading. The orientation of the transmitter is rotated slightly relative to the receiver such that the transmit and receive grating lobes do not coincide.

Upgrading and maintenance of the “wet end” of the system occurs at infrequent intervals when the ship is dry-docked. By 2006, the original 50 kHz transducers were replaced with more-efficient units (Figure 2). In 2007, a major upgrade of the sonar electronics and processing software was undertaken. The present (2009) technical effort is directed toward improved real-time data displays and performance diagnostics.

shear plotting

Figure 3. Cross sections of cross-shelf (a) and along shelf (b) velocity and shear in the Kuroshio, south of Okinawa.
Note the extremely coherent shear layers that appear to emanate from the shelf-break front.
Data were obtained by the 140 kHz sonar on the Revelle. (From Rainville and Pinkel, JPO, July 2004)

hdss shear plotting

Figure 4. Sections of meridional velocity (top) and shear (bottom) across the California borderland as
the Revelle approaches San Diego from the southwest.
The abscissa corresponds to approximately a distance180 km.
Both horizontal and “slanting” shear layers are seen with horizontal coherence scales approaching 100 km.

Why such a “big” sonar?

The need to install high-resolution sonars on research vessels is motivated by the problem of deep-ocean mixing and its role in climate change. It is now felt that most of the intense mixing sites on the planet are associated with under-sea topographic features, where wind or tidally driven currents generate energetic internal waves. These waves propagate into the mid-water column and subsequently break. It appears that one can quantify mean ocean mixing rates by assessing the level of small-scale (~10 m-50m) shear (Gregg, 1989, Alford and Pinkel, 2001, Polzin et al, 2002). While the reasons for this remain the subject of debate, the shear-based parameterizations estimate mixing rates to within a factor of three (whereas the rates themselves can very by orders of magnitude.

Froude Spectra

Figure 5. Froude spectra of vertical shear from 20 to 400 m.
The thick gray lines are from the shipboard 150 kHz ADCP (RDI) during the Kyushu cruise: inshore of the Kuroshio, in the Kuroshio, and offshore.
The nominal ADCP depth resolution was “set” at 8m. Spectra from the HDSS (in the Kuroshio) are shown by the thin black lines
(solid: the 50 kHz sonar (20-400m) , dashed: the 140 kHz (15 to 210 m) .The dotted line is the GM spectrum.
The HDSS spectra appear as simple “low passed” versions of a more energetic GM spectrum. Interpretation of the ADCP spectra is more complex.

There is thus a premium on measuring the spectrum of shear (and perhaps strain) at vertical scales as small as possible, over broad areas of the ocean. For basic surveying near shallow topography, shipboard Doppler sonar is an ideal tool. Near deep topography lowered sonars have been used with great success (eg. Garabato, et al 2004).

The in-situ performance of these systems depends on a mix of environmental conditions, instrument design parameters, and processing algorithms (eg. Pinkel, 1981, Polzin et al, 2002). Existing commercial systems have a puzzling inability to resolve high wavenumber shear, relative both to their stated specifications and to SIO-built sonars (Figure 5). The apparently poor resolution has been ascribed to intense small-scale horizontal variability (Polzin et al, 2002). In comparison with the HDSS 140 kHz, it appears to be a property of the commercial measurement system.

When mounted on a moving ship, the depth resolution, bias and velocity precision of a sonar are all influenced by the beamwidth (Figure 6). For example, a 6° wide beam spreads one meter for every 10 m of increased range. At 1000 m range, the “depth spread” of such a beam is about 50 m, assuming a beam depression angle of 60 degrees below horizontal. Beyond a few hundred meters, depth resolution is nearly independent of transmit pulse length, being set by the width of the spreading acoustic beam. A significant investment in narrow acoustic beams is thus required if one hopes to see fine-scale motions.

Schematic

Figure 6. Schematic of a finite-width sonar beam intersecting an isolated scattering layer (shaded). Even though the acoustic pulse is relatively short (δR), the effective depth resolution of the sonar (δZ) is much broader, set by the finite width of the acoustic beam. As the pulse enters the scattering layer from above, a, the observed Doppler shift is proportional to cosθlower. On exit, b, the shift is proportional to cosθupper. At typical ship speeds, this difference can be as great as expected ocean signals, for broad (~ +/- 3o) acoustic beams.

The Revelle HDSS was designed as a “benchmark” system, able to quantify the small-scale variability of the sea on a global scale. It was envisioned that 3-10 such systems (with the large 50 kHz transducers, at least) might be useful to the research community, globally. This is probably too small a “market” to attract commercial interest. Thus, development in a university research environment was warranted. The HDSS is one of a very few sources of moving-ship raw-echo data available to the academic community. Raw data is essential in the development of new signal processing schemes for shipboard applications.

Velocity

Figure 7. HDSS data are presently collected whenever the Revelle is underway (in international or cleared waters), Figure 7. Data support the research of the seagoing PIs and are archived at by the SIO Ocean Physics Group for distribution to non-seagoing users. PIs’ who whis to use the HDSS as a primary sensor can retain proprietary use of the data up to two years following its collection, in accord with NSF policy.