PAPER Acoustic Sensing for Research

AUTHORS ABSTRACT Bruce M. Howe Ocean observatories have the potential to examine the physical, chemical, biological, Applied Physics Laboratory, University of and geological parameters and processes of the ocean at time and space scales previ- Washington ously unexplored. Acoustics provides an efficient and cost-effective means by which these James H. Miller parameters and processes can be measured and information can be communicated. Inte- Department of Ocean Engineering, grated acoustics systems providing navigation and communications and conducting acous- University of Rhode Island tic measurements in support of science applications are, in concept, analogous to the Global Positioning System, but rely on acoustics because the ocean is opaque to electro- magnetic waves and transparent to sound. A series of nested systems is envisioned, from small- to regional- to basin-scale. A small number of acoustic sources sending coded, low power signals can service unlimited numbers of inexpensive receivers. Drifting and fixed receivers can be tracked accurately while collecting ocean circulation and heat con- tent data (both point and integral data), as well as ambient sound data about wind, rain, marine mammals, seismic T-phases, and anthropogenic activity. The sources can also transmit control data from users to remote instruments, and if paired with receivers en- able two-way acoustic communications links. Acoustic instrumentation that shares the acoustic spectrum completes the concept of integrated acoustics systems. The presently in the planning and implementation stages will require these inte- grated acoustics systems.

INTRODUCTION as NEPTUNE (NEPTUNE Partners, ics, geodesy, hot vents, gas hydrates). Tech- ecent advances in have 2000), and enhanced coastal observatories nical questions about the navigation and prompted investigation of integrated acous- (Jahnke et al., 2004). A unifying theme for communications are discussed with thoughts Rtics systems for ocean observatories (IASOO, the OOI infrastructure investment is the about implementation, followed with con- 2002) that include navigation, communi- provision of seafloor junction boxes pro- cluding remarks. cations, and acoustical oceanography. Such viding power and communications. The an integrated system may be considered a expectation is that the OOI request for combination of Underwater GPS (UGPS), major research equipment and facilities 2. Science topics acoustic communications, and acoustic sen- centers will be included in the 2006 U.S. sors that support science. The purpose of federal budget. This investment (equivalent 2.1 Ocean Circulation this paper is to investigate the evolving con- to several large ships) will, with planned The use of neutrally buoyant and profil- cepts, primarily within the context of sci- increases, lead to a concomitant large sci- ing floats has been a cornerstone of ocean- ence scenarios, and to describe briefly some ence investment. ography for the last half-century. In the late technical aspects. We examine first some science scenarios 1950s Swallow floats misbehaved by mov- A major motivating factor is the recently (not meant to be exclusive) and the integra- ing at “high” speed (10 cm s-1) in seemingly created National Science Foundation tion of these ideas into the infrastructure of random directions rather than slowly in a (NSF) ORION (Ocean Research Interac- planned research oriented ocean observatory straight line (baffling the crew on the sail- tive Observatory Networks) program and and operationally oriented observing system boat trying to track them acoustically), thus the embedded Ocean Observatories Initia- efforts. We explore applications of acoustics awakening oceanography to the presence of tive (OOI) to further the sustained study in the areas of ocean circulation (currents, the mesoscale “weather.” A detailed descrip- of the ocean (Clark, 2001). The three in- heat content, turbulence and mixing, wind tion of the history of acoustically tracked frastructure elements of the latter consist and rain), the living ocean/ocean biosphere floats can be found at http:// of a coarse global array of buoys (DEOS, (fisheries, marine mammals, nekton), and www.po.gso.uri.edu/rafos/general/history/ 2001), a regional cabled observatory such seafloor processes (earthquakes, plate tecton- index.html.

144 Marine Technology Society Journal Since then it has become abundantly clear temperature and salinity data and to obtain chored sound sources provide an acoustic that the ocean circulation needs to be mea- a satellite navigation fix. During the time navigation system whereby precisely timed sured on all scales, from global and basin submerged the float is untracked. Accord- acoustic signals (an 80-second long 3-Hz scales that are climatically relevant down to ing to Davis and Zenk (2001): wide FM sweep at 260 Hz) spread out radi- the small scales that are important to deter- ally from each of the sources. Given the ar- mining elusive mixing processes. There are Some studies place a high premium rival times at a receiver equipped with a syn- complementary observing technologies that on using floats to represent fluid-parcel chronous clock, the receiver’s position can are relevant—satellites and radars for the sur- trajectories requiring the uninterrupted be determined to an accuracy of a few kilo- face (not considered here), and Lagrangian current following that can be achieved meters. Since the mid-1980s floats have been platforms and fixed Eulerian sensors for in only with acoustic tracking. The pen- deployed in many different projects world- situ point measurements, both of which can alty for autonomous [profiling] opera- wide to study ocean currents. be spanned by acoustic tomography. tion is a long time interval between The floats, also known as RAFOS floats known positions, which precludes re- (SOFAR spelled backwards), can be deployed Fixed small scale measurements The solving eddies unless cycling is rapid, to drift at any depth although greater acous- acoustic Doppler current profiler (ADCP) and periodic surfacing that interrupts tic ranges are possible if both the sound is now a ubiquitous oceanographic instru- the quasi-Lagrangian trajectory. For sources and the floats are not too far off the ment. It can measure profiles of velocity observations of subsurface velocity, it is sound channel axis. The floats record the ar- (more precisely the velocity of scatterers fol- likely that both the continuously rival times in their microprocessor memory. lowing the water) at ranges up to about 1 tracked neutrally buoyant RAFOS At the end of each float’s mission underwa- km. Because there are no moving parts, the floats and autonomous floats will be ter, it surfaces and telemeters all data back to instrumentation is robust and reliable. More needed. For acoustic floats a major shore. Typical missions last from months to recently the acoustic travel time current limitation to economical sampling can several years. Special float designs have been meter (ACM) has been used on scales of 0.1 be overcome by widespread deployment developed whereby a float surfaces briefly to to 1 m. The acoustic scintillation current of high-energy sound sources. It is, for telemeter its data before returning to depth meter (ASCM) has been developed to pro- example, entirely feasible today to in- to continue its underwater mission. To make duce average measurements of cross-beam stall enough sound sources that a float the floats mimic fluid motion accurately, they velocity averaged along the path, typically could be continuously tracked anywhere can be configured to drift with the waters about 1 km. These various acoustic current in the tropical or North Atlantic. Since both horizontally and vertically; these are measurement devices have largely replaced sound sources, like radio stations, can known as isopycnal floats. mechanical current meters. serve different users, the presently rather In regions of high shear, profiling floats Inverted echosounders (IESs) measure simply structured network of moored can give erroneous estimates of velocity if the round-trip travel time from the seafloor sound sources will require greater in- they are based solely on surface fixes. Simi- to the ocean surface. This is a measure of ternational coordination in the future. larly, if trying to measure abyssal currents, the integrated, averaged sound speed (~tem- The benefits of an organized RAFOS estimates can be severely compromised if the perature) along this path. When IESs are network will be linked with the respon- float comes to the surface. In these cases it is combined with bottom pressure and a ref- sibility of contributing parties to main- advantageous to remain submerged. The erence absolute velocity (such as a point- tain such arrays of sound sources over same is true for other mobile instrument ACM or an electrometer that gives an extended period of time on a basin- platforms such as gliders, autonomous un- barotropic velocity), much can be learned wide scale. Miniaturization of receiver dersea vehicles (AUVs), and bottom rovers. about the low mode ocean velocity and den- electronics and production in great With UGPS (using low amplitude, long sity structure (Meinen et al., 2002). These numbers could result in significant de- duration, coded, high bandwidth signals, sensors provide fundamental information cline in float prices. This, and the de- and new low power clocks), the precision about the gravest vertical modes of ocean velopment of new sensors, could open and accuracy of the tracking can be much circulation, and because they are so simple other fields of research… better than with RAFOS systems (cf. LO- and robust (and may be used also for navi- RAN and GPS) resulting in correspondingly gation, communication, and short-range This statement is one motivation for the better velocity estimates. With appropriate tomography), they belong in any ocean ob- present work. transmission schedules, high temporal reso- servatory “basic sensor suite.” A widely used and accurate method for lution is possible, and at the limit (depend- studying the ocean in motion employs the ing on the spatial array size), Lagrangian platforms Autonomous pro- deep sound or SOFAR channel to track neu- time and space scales, and perhaps even filing floats, such as , typically come to trally buoyant drifters over great horizontal mixing scales, may be approached. the surface once every ten days to telemeter distances, O(103) km. Three or more an-

Summer 2004 Volume 38, Number 2 145 One illustrative application of use of respiration or the Chl-a record indicates high ing in space, and can sample at the speed of navigated floats is in the Atlantic Climate rates of production. When the float outcrops, sound. It can very efficiently measure depth- Change Experiment (ACCE) where the float the oxygen saturation level remains at 104% average temperature and velocity (if recipro- provided in situ observations of dissolved ± 2%. However, at the end of the first win- cal transmissions are used), with the amount oxygen. All University of Rhode Island ter the oxygen shows momentary drops as of depth dependent information a function (URI) ACCE floats measured dissolved oxy- large as 30% (2 mL/L). These occur at a of geographic location and the sound speed gen using the “pulsed” technique pioneered time when the Chl-a indicates the onset of profiles. A possible array showing the scale by Dr. Chris Langdon (Lamont Doherty primary production and suggests that sink- and the nesting aspects builds upon the Earth Observatory). Here the sensors are ing of detrital matter may be concentrated Acoustic Thermometry of Ocean Climate only briefly activated once a day, and thus to these newly-stratified density surfaces. (ATOC)/North Pacific Acoustic Laboratory can operate over the entire lifetime of the Around day 340, late June, the oxygen record (NPAL) array, using the future capability of floats. From a Lagrangian viewpoint, inter- approaches a local minimum in dissolved seafloor junction boxes provided by NEP- preting the variability of dissolved oxygen oxygen. This takes place after the peak of TUNE and other programs (Fig. 3). Note presents the difficulty of separating the bio- springtime production. This may be another, that when one source is added, data along logical from the physical processes. How- although less intense, indication of removal the paths to all the receivers are obtained. ever, there are times when the nature of the from detritus. On day 360 the summer These ideas can be extended to under- variability clearly indicates that one process bloom peaks simultaneously with a slight, ice operations in polar regions. One project dominates the other. but noticeable, 2% (0.1 mL/L) momentary will deploy gliders under the Labrador Figure 1 shows the track of a float (550), drop in dissolved oxygen. However, this winter ice with real-time RAFOS tracking; which outcrops in the Irminger Sea with the bloom is accompanied by a O(40 m) in- another will deploy profiling RAFOS floats corresponding measurements of pressure, crease in depth, which could have also ac- under the Antarctic ice (C. Lee and S. Riser, temperature, and dissolved oxygen, plus the counted for the oxygen drop. personal communications, 2003). In some satellite-derived values for significant wave cases, it may be advantageous to implant a height (SWH) and SeaWiFS Chl-a. The Acoustic tomography Acoustic tomogra- satellite/GPS-acoustic/UGPS transponder track and property plots are color-coded to phy can complement point float (Fig. 2) and in sea ice to provide two-way communica- indicate two distinct periods: red when the moored measurements by providing spatial tion to mobile platforms (J. Morison, per- float outcrops into the wintertime mixed integrals between the navigation sources and sonal communication, 2003). layer, and green when either the oxygen the receivers, whether fixed or moving. Tomography over a 1000-km scale with record indicates significant removal due to Acoustic tomography is inherently averag- a drifting receiver has been demonstrated (AMODE Group, 1992). There are numerous simulation papers de- FIGURE 1 scribing tomography using drifting receiv- The track (left) of an isopycnal float (550), which outcrops in the Irminger Sea with the corresponding measure- ments of pressure, temperature, and dissolved oxygen (right) (from Lazarevich et al., 2004). ers, where the position is simultaneously determined (Cornuelle, 1985; Gaillard, 1985; Duda et al., 1995) as seismologists do to determine earthquake location. This scenario of simultaneously determining float/receiver position and the sound speed field will require three or more sources to be “in view.” If mobile platforms are used as receivers it is likely that a data assimilation system will be used that takes in surface fixes, travel times, Doppler velocities, depths, es- timates of C(x, t), etc., to simultaneously estimate the float state (position, velocity, and acceleration as a function of time), along with the ocean state (e.g., temperature, sa- linity, velocity).

Turbulence Small-scale turbulence in the ocean has significant consequences for both carbon fluxes and marine biodiversity

146 Marine Technology Society Journal FIGURE 2 through direct and indirect effects on the The ATOC 0–1000-m depth-averaged temperature along a path from Kauai to the northwest (red, very small error marine ecosystem, as detailed in the SCOTS bars), compared against Argo float data in a box surrounding the path (B. Dushaw, personal communication, 2003). Report (Glenn and Dickey, 2003) and sum- marized below. In addition, turbulence has profound effects on air-sea exchanges of greenhouse gases and on the magnitude of the . Ignorance of the processes and locations of turbulent mix- ing in the ocean has such serious implica- tions for global climate modeling and cli- mate change projections that the NSF Mil- lennium Report (Brewer and Moore, 2001) identifies turbulence as one of the remain- ing ‘big questions’ in ocean science. While the focus of this section in the SCOTS re- port was the combination of turbulent mix- ing and biophysical interactions, the authors felt it important to emphasize that our basic knowledge of small-scale ocean turbulence will be significantly advanced by the ability FIGURE 3 to make long time series measurements in A possible acoustic array on regional and basin scale for collecting ocean circulation and heat content data while an observatory setting, and to cover scales meeting navigation and communications requirement. up to the mesoscale. Cabled systems are essential for tackling problems in small-scale turbulence and its interactions with all levels of the marine eco- system. Both the ample power and, espe- cially, the broad bandwidth that will be pro- vided by cabled systems are required by present methods for measuring turbulence quantities (dissipation scales with microscale sensors or large- scales with acoustics), and for those biological techniques (such as multi-frequency acoustics, multi-wavelength and sheet light optics, holography, and video/ still photography) that can provide detail sufficient to adequately delineate ecosystem responses. A regional-scale cabled system with an adequate footprint is necessary for the investigation and comparison of turbu- lent mixing regimes with very different char- acters—deep-sea mixing driven by tidal flow over steep topography, upper mixed layers subject to very different combinations of wind and buoyancy forcing, and shelf re- gimes strongly influenced by tidal bottom boundary mixing. The character of embed- ded ecosystems also varies dramatically: the challenge is to discover how much of the bio-geographical variability arises through the direct and indirect effects of small-scale turbulence.

Summer 2004 Volume 38, Number 2 147 Holbrook (2003) recently showed that speed and rainfall rate. The signal from 2.2 The Living Ocean/ one can image density (sound wind has a characteristic shape that rises The Ocean Biosphere speed) gradients on vertical scales of ~10– and falls in amplitude with wind speed. The The fundamental difference between 100 m using seismic airgun data. In the past signal from rain is much louder, contain- conventional acoustic sampling platforms the water column data was ignored; by ad- ing relatively more high frequency sound. and ocean observing systems is the poten- justing the gain one can look for “layers” in In particular, there is a unique bubble en- tial for broad-scale spatial and temporal sam- the ocean just as one does beneath the seaf- trapment mechanism associated with small pling from a suite of sensors in real time. loor. Images of eddy interactions and inter- raindrops that produces sound at 13–25 This unprecedented combination provides nal waves breaking on the continental slope kHz. This is a surprisingly loud sound that both opportunities and logistic challenges are possible. Equivalent cabled systems will allows the acoustic detection of light drizzle to the fisheries and marine mammal science be able to observe the time dependence of at sea. While biological and anthropogenic communities. A combination of Eulerian these structures. noises can also occur in this frequency band, (i.e., fixed grid) and Lagrangian (i.e., mo- these noises are generally local or intermit- bile) sampling strategies are required to ad- Rainfall and wind In the frequency tent and usually do not interfere with acous- equately sample fish and marine mammals range from 500 Hz to 50 kHz, the domi- tic wind speed and rainfall rate measure- through the range of ecologic scales encom- nant sources of underwater ambient sound ments. The measurement of wind and rain passed by an ocean observatory. Detection in the ocean are bubbles generated by at sea is notoriously difficult, especially ranges of most active and passive sensors lim- breaking waves and raindrop splashes under storm conditions (when air-sea fluxes its the feasibility of relying solely on a grid (Nystuen, 2001; Medwin, 1992). Figure are highest), and the acoustic method pro- sampling design. Strategic choice and place- 4 shows the mean spectral shapes of rain- vides one of the few ways, if not the only, ment of sensors should enable a suite of or- and wind-generated underwater sound to do so reliably with exceptionally robust ganisms to be continuously monitored over gathered from over 100 months of ambi- instrumentation—a simple hydrophone. It a broad range of spatial and temporal scales. ent sound measurements on deep-sea can be a cornerstone of widespread verifi- Commercial fisheries (i.e., fishes and inver- moorings. The spectral shapes associated cation of satellite-derived wind and rain- tebrates) research opportunities include: (1) with wind and rainfall are distinctive, al- fall estimates. examining spatial and temporal fluxes of lowing quantitative measurements of wind biomass, (2) resource assessment with ex- tended temporal sampling, (3) quantifying FIGURE 4 distribution variability in space and time, and Wind- and rain-generated underwater sound has unique characteristics that allow acoustic measurements of (4) species-specific habitat use. Marine mam- wind speed and rainfall rate (courtesy of J. Nystuen). mal research opportunities include: (1) quantifying spatiotemporal distributions of animals, (2) investigating biological re- sponses to oceanographic variability, and (3) monitoring behavioral responses to anthro- pogenic noise sources. A common element to both groups is the opportunity to moni- tor changes in densities and distributions of aquatic organisms in response to variability in the marine ecosystem. Many life history characteristics (e.g., timing and path of an- nual migrations) are not known for fish and mammal species. A continuous suite of sen- sors provides a networked platform to quan- tify and understand aquatic life cycles. Underwater sound is produced by ma- rine mammals over a very wide frequency range, from a few Hz to more than 100 kHz. These sounds are used for communication, navigation, and hunting. The signals are generally unique to animal species and can therefore be used to monitor animal behav- iors, including migration patterns, feeding

148 Marine Technology Society Journal behaviors, and communication between tracked in an appropriately instrumented single one requires an integrated knowledge animals; an example of blue and fin whale ocean basin. When the fish are caught, data of the others. For example, rock alteration, spectra is shown in Figure 5. Because sound in the tag can be downloaded. In the active mineral precipitation, and the formation of travels through water more easily than light, mode several groups are working with small gas hydrate (methane “ice”) caused by fluid acoustic rather than visual investigations of salmon tags which allow the fish to be flow change the mechanical and hydrologic marine animal populations is usually more counted as they go by a bottom-mounted properties of the structure. Marine gas hy- effective. By listening to the animals’ signals hydrophone array. The range of the system drates represent an untapped hydrocarbon on multiple , they can be is several hundred meters at best (Fig. 6). resource of great magnitude, and at the same tracked. A possible scenario for studying Modern imaging can reveal criti- time, a possible source of greenhouse gas that large numbers of marine mammals is to use cal information about the behavior and hab- could lead to global climate change if the arrays of mobile receivers (serving multiple its of marine animals. Figure 7 shows a so- gas were released in large quantity. A pri- users, such as Argo floats) positioned using nar image of two fin whales foraging on a mary goal over the next few decades will be a combination of UGPS and/or GPS to school of herring. Concepts for autonomous to understand the factors in the hydrologi- track the animal vocalizations. This assumes vehicles that can image krill with 3D arrays cal system that affect hydrate stability. that future in situ recording/processing and are under development for large volume bio- Sites with acoustic instrumentation can communications capabilities increase, as is mass assessment (P. Chandler, personal com- provide long-term monitoring of seafloor expected with time. munication, 2003). hydrocarbon vents and the associated struc- Both marine mammals and fish can be tural changes with hydrate formation. These tracked using acoustic tags, either passive (in 2.3 Seafloor Processes surface sites should be linked to borehole the RAFOS mode) or active (in the SOFAR Integrated acoustics systems can support experiment sites where the presence and/or mode) depending on the situation. There a broad range of scientific objectives in ma- absence of structural controls for fluid flow have been significant developments in tag- rine geology and geophysics including crustal (e.g., faults, high permeability layers, cracked ging technology for marine mammals and hydrology/geochemistry/biology, clathrates rock) have been adequately mapped by geo- fish. For example, G. Fisher, C. Recksiek, and zone fluids, continental physical surveys and/or drilling. and T. Rossby at University of Rhode Island margin sedimentation, ridge crest processes, A generic suite of acoustic sensors should are developing a “fish chip” RAFOS receiver and seismology/geodynamics. include the following: broadband seismom- that can operate 2–3 years with a 100-mA- Hydrologic systems in the ocean crust eter/sea floor hydrophone; broadband seaf- hr battery. The entire tag is 2.5 cm long and and sediments at ridge crests, ridge flanks, loor moored hydrophone arrays, both verti- 1 cm in diameter. In addition to position, and subduction zones play key roles in in- cal and horizontal; associated hydrophone the tag provides temperature and pressure fluencing rock alteration, mineral formation, and geophone arrays in boreholes drilled that is be stored in non-volatile memory. The and hydrocarbon migration. Hydrologic near the hydrate sites. A suitably instru- small size and long lifetime would allow in- processes and their consequences are strongly mented borehole array in a gas hydrate prov- dividual fish to be tagged and potentially linked, and therefore to understand any ince would provide a tomographic image to monitor the hydrological system over time. FIGURE 5 Observatories will be used to study the Multi-year ambient sound spectra from Point Sur, California. Daily average spectra are plotted so signals from temporal variability of ridge crest processes earthquakes, ships, and other sources are not shown. Marine mammals at ~17 Hz are the loudest signals here. (tectonic/physical/chemical/biological). It is often stated that we know the topography of Mars and Venus better than we do Earth, largely because of the masking nature of the . Mapping is one of the natural first tasks of exploring a new environment, and its importance in understanding and devel- oping the oceans over the next century should not be underestimated. While near-autonomous mapping is just beginning it is expected that long duration missions will be possible in the future, mak- ing use of seafloor charging and communi- cations stations/docks. This is one possible way to cover relatively large areas at a rea- sonable cost. AUVs can also be used in a

Summer 2004 Volume 38, Number 2 149 FIGURE 6 earthquake processes at active margins such Proposed hydrophone array locations of the Pacific Ocean Salmon Tracking Project (POST). Additional infor- as ridges, transform faults, and subduction mation can be found at www.coml.org/descrip/post.htm. zones and to study intraplate earthquake activity. Any and all acoustic receivers can play a role in the detection of earthquakes because they will observe seismic T-phases (water borne signals). When signals from even a sparse array of receivers are combined, even small earthquakes can be observed and earthquake locations can be determined rela- tively precisely (Fig. 8). This has been ap- plied with success in monitoring the Juan de Fuca plate, East Pacific Rise, and mid- Atlantic Ridge for volcanism and seismic- ity. Recently, these water borne signals have been used to infer information on the state of icebergs and ice sheets in Antarctica. Deep “shadow zone” arrivals (Dushaw, 1999; But- ler, 2002) have been detected below the sound channel. While the theoretical expla- nation for these is lacking (it is thought some scattering process is responsible), this means FIGURE 7 that navigation in parts of the deep ocean Two fin whales foraging in a school of herring off the coast of Norway (courtesy of O. Misund, IMR, Norway) that were previously thought to be excluded should be possible at some level. These T- phase data can be used for CTBT monitor- ing purposes, and could contribute to tsu- nami early warning systems. The combination of IASOO and AUVs will provide a unique opportunity to obtain complementary long-term measurements of crustal deformation for the entire life cycle of a plate from creation to subduction. Geo- detic instruments will be concentrated near the plate boundaries, but it is also impor- tant to incorporate techniques to measure plate scale deformation. Compared to land- based observations, seafloor geodesy is in its infancy. However, several techniques are emerging to measure both the vertical and survey mode to locate hydrothermal vents during an eruption, to concentrate activity horizontal deformation (Chadwell et al., in remote locations. In many situations pre- around lava flows. In the past temporary 1999; Chadwick et al., 1999, Spiess, et al., cision repeat surveys over large areas (100s local transponder nets have been set up at 1998). It should be feasible to use AUVs as of kilometers) are desired as in, for example, significant expense for a particular operation; the intermediate vehicle for acoustic geod- studies of seafloor deformation near subduc- future work will require permanent naviga- esy, and acoustic “beacons” will be required tion zones (with centimeter accuracy) and tion and communications capability. to navigate the AUVs. Just as the GPS satel- deep sea ecology. Sensor networks for ocean Most of the earthquakes in the world lites are used for both navigation and geod- observatories should include navigation and occur either under the oceans or involve at esy on land, a network of acoustic transpon- communication with AUVs and bottom least one oceanic plate. Seafloor seismic ob- ders could serve a dual purpose in the oceans. rovers. The latter might be controlled in real servatories will be a valuable tool to extend Deep ocean hydrothermal systems are time using two-way acoustic communica- the global distribution of seismic stations to notoriously difficult to study. Our ability to tions to make routine surveys, and then the seafloor (Stephen et al., 2003), to study study these systems and their relation to tec-

150 Marine Technology Society Journal FIGURE 8 tonic, magmatic, oceanographic, and biologi- A relatively few single hydrophones in the sound channel can be used to detect and locate many earthquakes in cal processes is limited by the lack of tools for an ocean basin that would be undetectable from land arrays. In this example six hydrophones (white stars) can resolving critical spatial and temporal scales be used to locate earthquakes (red dots) throughout the North Atlantic (courtesy of D. K. Smith). of flow that are characteristic of ridge tectonic processes. Observing the temporal/spatial variability and partitioning of energy between different types of flow is critical for under- standing fluid circulation in the crust, its in- teraction with crustal alterations, and its in- teraction with biological habitat. The ability to model hydrothermal fluid circulation af- ter it has left the seafloor and the entrainment of surrounding water into hydrothermally induced flow is limited by our inability to characterize flow at the scale of the entire plume within sub-tidal time scales. Point measurements made from a moving ROV or AUV, for example, are often subject to vari- ability associated with tidal cycles. Interpre- tation of such aliased measurements requires well defined forward models of tidal flow and its interaction with complicated topography. Acoustic systems are a primary technol- ogy for probing and monitoring hydrother- FIGURE 9 mal flow in the context of observatory sci- An acoustic volume backscatter image of a buoyant plume from which estimates of plume geometry as a function of time, particle mass concentration, flow velocity, and mass flux can be derived (courtesy of C. Jones). ence. In coordination with point measure- ments from both stationary and mobile sea- floor systems (AUVs), acoustic remote sens- ing can offer access to new scales of hydro- thermal processes and their interactions with the wider seafloor and ocean. Figure 9 illustrates an acoustic volume backscat- ter image of a plume from which estimates of plume geometry as a function of time, particle mass concentration, flow velocity, and mass flux can be derived.

3. Integrated acoustics systems

3.1. Navigation Acoustic navigation has taken two forms: fixed and floating ranges. Typical fixed ranges, e.g., Navy tracking ranges, use simple pings or relatively simple coded signals and are deployed over limited areas. Floating ranges emulate fixed ranges since the posi- tioning of their fiduciaries are accurately known with GPS (Howe et al., 1989), and there are several commercial products of- fered. There is at least one commercial prod- uct that has a bottom mounted system di-

Summer 2004 Volume 38, Number 2 151 rectly analogous to GPS, with continuous ments of an integrated system will be chal- mounted acoustic transducer that could transmissions of broadband pseudo-random lenging. Acoustic communications have serve multiple purposes: an inverted noise (PRN) signals. Float tracking for proven to be difficult. While basin-scale echosounder (depth-averaged temperature), oceanography first used the SOFAR mode communications have been shown to be fea- tomography, geodesy, ambient sound (wind, (drifting sources, fixed receivers) and later sible (Freitag and Stojanic, 2001), most ef- rain, marine mammals, T-phases, etc.), and the RAFOS mode (fixed sources and drift- fort has been devoted to short-range (kilo- navigation and communications. Together, ing receivers). Tomography instrumentation meters) applications either in shallow water all these applications will require manage- immediately used PRN signals. Because pre- (with relay stations, for instance) or with ment of the acoustic spectrum. requisites for tomography are accurate tim- vertical paths in deep water (bottom to a Following are some of the questions the ing and accurate navigation of the instru- surface ). Signal processing has in- community will need to address accompa- ments, tomography arrays can be used im- cluded both coherent (Fig. 10) and inco- nied by brief comments. mediately for navigation purposes and vice herent methods. Clearly more research and ■ What are the optimal frequencies and versa (cf. atmospheric and ionospheric to- development is necessary, and the commu- bandwidth to use? ATOC at 75 Hz has mography). nity needs to converge to a common set of shown that adequate signal-to-noise Because of and the nature standards that will accommodate the limi- (SNR) ratio can be obtained at ranges of the sound speed field throughout the tations of the medium and instrumentation of 5 Mm. Are there advantages to trans- oceans, we do not have an ideal geometry, while spanning the wide range of desired mitting two frequencies? Recent work by and cannot have the coverage that GPS space scales, frequencies, noise conditions, the NPAL Group (Worcester et al., does. There will be areas and depths not data and error rates. 1999) has shown that signals at lower covered. It will be necessary to treat spe- frequencies appear to suffer fewer detri- cial cases (e.g., shallow , deep trenches, 3.3. Implementation mental effects from internal wave double ducts, polar seas, stripping of Some uses of acoustics have been de- induced fluctuations (28 vs. 84 Hz). steeper angles by bathymetry) with cre- scribed briefly here. There are clearly many ■ What are achievable integration times? ativity, understanding there may not be more including volume imaging, side-scan Experience shows that at 75 Hz coherent just one solution. , fish and bio-acoustic backscatter, averaging times are about 14 minutes. turbulence and internal wave sensing, and In the Arctic coherence at 20 Hz is 3.2. Communications sediment transport. In an observatory set- sufficient that phase can be tracked over Simultaneously satisfying all the science, ting, a basic science/infrastructure element long time periods. navigation, and communications require- could be a single (in principle) bottom- ■ What data must be broadcast on top of the navigation signal? If a source is on a mooring should the source position be FIGURE 10 sent (or model parameters describing the Coherent signal processing for communications using spatial diversity (Rouseff et al., 2001) motion), much as ephemeris data is transmitted as part of the GPS signal? ■ What are the optimal types of acoustic sources to use under various circumstances? Should they incorporate directionality? ■ What type(s) of signals should be used? PRN signals are optimal in one sense (and are “orthogonal” to marine mammal vocalizations), but their use might exclude some types of sources. Code, time, and frequency division multiplexing are all possible. ■ What should the duty cycle be? If the source is deep, with a horizontally- oriented beam pattern, and relatively low level (all measures to mitigate possible effects on marine mammals), is it practical to consider continuous transmission?

152 Marine Technology Society Journal ■ While ambient sound variability is not ray with receivers on floats will further ex- Acknowledgments well characterized, estimates are still tend the spatial coverage. The Acoustical Oceanography Subcom- required to determine SNR budgets. mittee on Integrated Acoustics Systems for ■ Should receivers be arrays with vertical 4. Concluding Remarks Ocean Observatories (IASOO), Acoustical or horizontal directionality? It would A navigation and communications in- Society of America, http://www.oce.uri.edu/ seem reasonable to use two hydrophones frastructure is a prerequisite to sustained ao/AOWEBPAGE, was formed in Novem- on a float (top and bottom) to get an human endeavors in the ocean. Many ap- ber 2002 at the Cancun meeting of the Acous- extra 3 dB of gain, for example. plications require or are enabled by this in- tical Society of America. The authors of this ■ For a given geometry of sources, a noise frastructure. Autonomous undersea vehicles paper drew upon the expertise of the com- scenario, and the sound speed field, it (powered and gliding) can navigate them- mittee members and other contributors: will be necessary to estimate the signal selves over the ocean bottom and through Bruce Howe (co-chair), James Miller (co- level, SNR, and position error estimates the water column without coming to the chair), Whitlow Au, Paul Chandler, Ross as a function of position. The estimated surface. They can navigate and communi- Chapman, Kerry Commander, Orest signal levels and SNR for a single receiver cate their data and status to users via acous- Diachok, Percy Donaghay, Lee Freitag, Dale will be necessary to evaluate the ocean tic modems to cabled or surface satellite te- Green, John Horne, Christopher Jones, Paul volumes in which marine mammals might lemetry systems without breaking away from Jubinski, James Mercer, Susan Moore, Jeffrey hear the raw signal above the noise. their underwater missions. Profiling and Nystuen, Gopu Potty, Thomas Rossby, ■ How do we resolve conflicts between drifting floats can more accurately measure Hanne Sagen, Henrik Schmidt, Yang Shen, active and passive users? the ocean’s velocity structure, tagged fish and Ralph Stephen, and D. Randolph Watts. The As even this sample of questions indi- marine animals can be tracked with high comments of the reviewers are appreciated. cates, there is much work required to ma- precision, and bottom-fixed instruments can ture this concept, let alone implement it. The measure seafloor motion. development effort will include: establish- The acoustic sources and receivers of References ing essential standards to unify the field (such such a system can serve multiple functions: AMODE Group. 1994. Moving ship as signal protocols), working on more effi- sources as navigation and communications tomography in the northwest Atlantic Ocean, cient broadband and directional sources and components as well as multi-static active EOS Trans. AGU., 75, 17, 21, and 23. miniature acoustic receivers, investigating transmitters, and receivers as communica- coherence times and lengths as functions of tions components as well as passive listen- Brewer, P. and T. Moore. 2001. Ocean frequency and range, understanding ambi- ing devices. With signal standards and pro- Sciences at the New Millennium, University ent sound variability, to name but a few. tocols for managing the acoustic spectrum, Corporation for Atmospheric Research. 152 pp. Over the last decade, much research has the system will be an extensive, multipur- (www.geo.nsf.gov/oce/ocepubs.htm) been done on the effects of anthropogenic pose acoustics infrastructure, capable of sup- Butler, R. 2002. Coupled seismoacoustic sound on marine mammals. Concerns about porting applications even beyond our modes on the seafloor. Geophys Res Lett. 29, possible effects will have to be taken into present vision. Analogous to the GPS and 10.1029/2002GL014722. account in designing an integrated system. its use for tomography of the atmosphere Some mitigations measures have already and ionosphere, the various acoustic sources Chadwell, C.D., J.A. Hildebrand., F.N. Spiess, been suggested, e.g., deep, directional, low- and multitude of receivers can function simi- J.L. Morton, W.R. Normark and C.A. Reiss. level sources (less than 250 W, 195 db re 1 larly in the ocean. This has significant im- 1999. No spreading across the southern Juan mPa at 1 m) using PRN codes. Much work plications for observing the ocean’s interior de Fuca ridge axial cleft during 1994-1996. is on-going in this area and it can benefit in real time and measuring long-term cli- Geophys Res Lett. 26:2525-2528. from an integrated system. mate variability. Receivers on globally dis- Chadwick, W.W., Jr., R.W. Embley, H.B. Actual fielding and operation of systems tributed floats can also listen to ambient Milburn, C. Meinig and M. Stapp. 1999. on various space scales might fall within the sound: wind and rainfall, seismic T-phases, Evidence for deformation associated with the ocean observatories and ocean observing marine mammals, and ships. Some of these 1998 eruption of Axial Volcano, Juan de Fuca systems that are being planned. On a re- natural sources of sound can in turn be used Ridge, from acoustic extensometer measurements. gional scale NEPTUNE will provide a rela- as sources of opportunity for other purposes. Geophys Res Lett. 26:3441-3443. tively dense array of seafloor nodes that can support the acoustics sensor networks. On Clark, H.L. 2001. New sea floor observatory a global scale significant acoustic coverage networks in support of ocean science research. can be obtained with fixed instruments us- Proceedings, MTS/IEEE Oceans 2001 ing ocean observatory and ocean observing Conference, Honolulu, Hawaii, November system assets; and augmenting this fixed ar- 5?–8. See also www.geo.nsf.gov/ooi.

Summer 2004 Volume 38, Number 2 153 Cornuelle, B.D. 1985. Simulations of acoustic Integrated Acoustics Systems for Ocean Spiess, F., D. Chadwell, J. Hildebrand, L. tomography array performance with untracked Observatories (IASOO), Sub-Committee of Young, G. Purcell and H. Dragert. 1998. or drifting sources and receivers. J Geophys the Acoustical Oceanography Technical Precise GPS/acoustic positioning of sea floor Res. 90:9079-9088. Committee of the Acoustics Society of reference points for tectonic studies. Phys America. 2002. See the Web page: http:// Earth Planet Inter. 108:101-112. Davis R.E. and W. Zenk. 2001. Subsurface www.oce.uri.edu/ao/AOWEBPAGE Lagrangian observations during the 1990s. Stephen R.A., F.N. Spiess, J.A. Collins, J.A. In: Ocean Circulation and Climate, Observing Jahnke, R., J. Bane, A. Barnard, J. Barth, Hildebrand, J.A. Orcutt, K.R. Peal, F.L. and Modeling the Global Ocean, ed. G. F. Chavez, H. Dam, E. Dever, P. DiGiacomo, Vernon and F.B. Wooding. 2003. Ocean Siedler, et al., 123-139. Academic Press. J. Edson, R. Geyer, S. Glenn, K. Johnson, seismic network pilot experiment. Geochem M. Moline, J. O’Donnell, J. Oltman-Shay, Geophys Geosyst. 4(10):1092, doi: 10.1029/ DEOS. See http://www.coreocean.org/DEOS O. Persson, O. Schofield, H. Sosik and 2002GC000485. Duda T.F., R.A. Pawlowicz, J.F. Lynch and E. Terrill. 2004. Coastal Observatory Research Worcester, P., et al. 1999. A comparison of B.D. Cornuelle. 1995. Simulated tomographic Arrays: A framework for implementation long-range acoustic propagation at ultra-low reconstruction of ocean features using drifting planning, Skidaway Institute of Oceanography (28 Hz) and very-low (84 Hz) frequencies. acoustic receivers and a navigated source. Technical Report TR-03-01. Proceedings, U.S.-Russia Workshop on J Acoust Soc Am. 98:2270–2279. Lazarevich, P., T. Rossby and C. McNeil, 2004. Experimental , 18–20 Dushaw B.D., et al. 1999. Multimegameter Oxygen variability in the near-surface waters of November 1999, ed. V. Talanov, 93-104. range acoustic data obtained by bottom- the northern North Atlantic: observations and Nizhny Novgorod: Institute of Applied mounted hydrophone arrays for measurement a model. Journal of Marine Research. In press Physics, Russian Academy of Sciences. of ocean temperature. IEEE J Ocean Eng. (September). 24:202–214. Medwin H., et al. 1992. The anatomy of Freitag, L. and M. Stojanovic. 2001. Basin-scale underwater rain noise, J Acoust Soc Am. acoustic communication: a feasibility study 92:1613–1623. using tomography m-sequences. Proceeedings Meinen, C.S., D.S. Luther, D.R. Watts, K.L. of MTS/IEEE Oceans 2001, Vol. 4, pp. 2256- Tracey, A.D. Chave and J. Richman. 2002. 2261. Honolulu, HI, Nove. 5-8, 2001. Combining inverted echo sounder and Gaillard F. 1985. Ocean acoustic tomography horizontal electric field recorder measurements with moving sources or receivers. J Geophys to obtain absolute velocity profiles. J Atmos Res. 90:11,891-11,899. Ocean Technol. 19:1653–1664.

GEO/Reference Timeseries Moorings. NEPTUNE Phase 1 Partners (University of See www.oceansites.org. Washington, Woods Hole Oceanographic Institution, Jet Propulsion Laboratory, Pacific Glenn S.M. and T.D. Dickey, eds. 2003. Marine Environmental Laboratory). 2000. SCOTS: Scientific Cabled Observatories for Real-time, Long-term Ocean and Earth Time Series, NSF Ocean Observatories Studies at the Scale of a Tectonic Plate. Initiative Workshop Report, Portsmouth, VA, NEPTUNE Feasibility Study (prepared for the 80 pp. www.geo prose.com/projects/ National Oceanographic Partnership scots_rpt.html. Program), University of Washington, Seattle. Holbrook, W.S., P. Paramo, S. Pearse and See also http://www.neptune.washington.edu. R.W. Schmitt. 2003. Thermohaline fine Nystuen, J.A. 2001. Listening to raindrops structure in an oceanographic front from seismic from underwater: An acoustic disdrometer. reflection profiling. Science, 301:821-824. J Atmos Ocean Technol. 18:1640–1657. Howe, B.M., et al. 1989. A floating acoustic- Rouseff, D., D.R. Jackson, W.L.J. Fox, C.D. satellite (FAST) range, Proceedings, Conference Jones, J.A. Ritcey and D.R. Dowling D.R. on Marine Data Systems, 225–230 (Marine 2001. Underwater acoustic communications Technology Society, New Orleans). by passive-phase conjugation: theory and experimental results. IEEE J Ocean Eng. 26.

154 Marine Technology Society Journal