Multi-beam sonars are now the standard technology used to survey the ocean bottom.   There are several manufacturers of multibeam sonar systems and Applied Microsystems Ltd instruments and sensors are used in support of all multibeam systems.

Unlike a traditional echosounder, with each acoustic ping a multi-beam sonar system can plot the bottom depths for dozens to hundreds of points in a line perpendicular to the the heading of the ship.  This gives the multibeam sonar two huge advantages over a traditional single beam echosounder: 1) The resolution is far superior to an echo sounder; and 2) The wide swath coverage dramatically reduces the time and cost of surveying an area.

The figure above shows the multibeam sonar footprint below the ship.  The transmitted acoustic beam is narrow along track and wide across track (area in yellow).   There are many received beams but each one is long along track and narrow across track.  One receive beam is highlighted in green. The intersection of the transmit beam area and the the individual receive beam area provides the footprint for that beam, shown in blue.  The depth is the average depth within the intersection area.

With this technique hundreds of tightly spaced depths can be determined from a single acoustic ping.

Multibeam sonar systems can provide highly accurate charts of the bottom bathymetery.  The accuracy is dependent on several factors. 

• Knowing the position of the transducer head relative to the earth.  Vertically and horizontally.  This data  a starting point for the acoustic pulse.
• Knowing the pitch, roll and yaw angles of the transducers head.  This data gives a start and finish orientation for the transducer head.
• Knowing the sound speed at the transducer head.  This allows the determination of the arrival angles of the various beams relative to the transducer head.
• Knowing the sound speed profile from the transducer head to the sea floor.  This allows the sonar processor to calculate range and to correct for refraction (bending) of the acoustic beams.

Applied Microsystems Ltd instruments and sensors are used to correct the latter two sources of multibeam error.

Beam angle errors due to inaccurate sound speed estimates at the multibeam transducer head will produce significant horizontal and vertical errors for the outer beam footprints.  Since the ship is underway during the survey it is possible to see rapid changes in sound speed at the surface.  This is especially possible at ocean fronts or in littoral waters near estuaries.  For this reason a sound speed sensor should be mounted at the multibeam transducer head and sampled  with every ping of the multibeam sonar.  Due to the possiblity for rapid changes of sound speed it is better to use a direct measuring sound speed sensor than to calculate sound speed from CTD data.  Sound speed data from CTDs will have 'Salinity Spikes' in the data which will cause the angle calculations of some multibeam pings to be in error whenever the ship transits an area of rapidly changing temperature.  Direct measuring sound speed sensors do not exhibit this problem due to the rapid resonse time of the sensor.

Surface sound speed (sound speed at the sonar transducer array) should be updated as often as possible for beam steering. Present day multibeam software programs allow a direct serial port connection of a sound velocity sensor.  With a direct connection the multibeam software can poll for a sound velocity update with every sonar ping providing the best possible accuracy.

For any acoustic array, including seismic arrays and multibeam systems, tuning the array to listen in one specific direction requires knowledge of sound speed at the hydrophones.  The listening angle is a function of the distance between the hydrophones and the speed of sound. 

For illustration purposes, we can simplify the calculation to two hydrophones - instead of many more - and work in two dimensions - instead of three.  In the diagram below, an acoustic plane wave approaching the array at the listening angle desired will encounter hydrophone A first and then it must travel the additional distance e before encountering hydrohone B.  There will therefore be a time delay t between the detection of the signal between hydrophones A and B. 

To maximize the response in a specific direction  we want the signal at both hydrophones to constructively interfere when the hydrophone outputs are added.  This can be accomplished by adding a time delay of t to hydrophone A and then summing the two hydrophone outputs.

As shown in the figure the sine of the listening angle is angle directly proportional to c, the sound speed.  This is the reason that sound speed measurements at the array are critical to an array's directional accuracy.

Since the seismic array is moving, the sound speed can change as the array is towed through the water.  In addition, since the array is large the sound speed at one end can be different from the other.  For this reason sound speed sensors should be distributed throughout the array.

Regardless of which deployment option is selected - sea chest installation or transducer deployment - a ship board sensor will have the following critical requirements:

The path the acoustic wave takes from the transducer to the seafloor is described by Snell's law.  The acoustic path will bend towards the slower sound speed.  In the diagram below the sound speed is lower at the surface than it is at depth.  Therefore the acoustic path will bend upwards.  If the sound speed is faster at the surface than at depth the path would bend down.

The diagram below also illustrates the errors generated by having an incorrect sound speed profile.  Sound speed profile errors are easily seen in multibeam sonar data.  When the multibeam data is cleaned (dot killing) a flat bottom will exhibit a depth error in the outer beams ('smiling' or 'frowning' lines).  The figure below shows a 'frowning' line.  That is, the bottom appears to be deeper away from the ship. 

Also note the apparent shift of the round target on the bottom.  Not only is there a depth error but there is also a horizontal position error caused by the sound speed profile error.

The sound speed profiles can be quite variable in the ocean due to factors such as fresh water run-off, daily heating and cooling cycles, upwelling and downwelling, etc.  Accordingly, an accurate and timely sound speed profile is required to process the multibeam sonar data.

Normally the ship must stop to take a sound speed profile.    Since time is money, when performing surveys it is important to obtain the profile as quickly as possible.  This usually means profiles are taken at speeds of 1 m/s or more. Even so, a deep 2000 m profile can take over an hour.  To make matters worse the profile should be refreshed at least every 6 hours.  In coastal waters the profile should be refreshed more frequently due to fast changing conditions.  The result is a requirement to profile as fast as possible.

Due to the 'Salinity Spiking' problems associated with profiling CTDs at high speed it is always better to use a direct measuring sound speed sensor rather than a CTD.

When conducting a vertical profile, ocean professionals normally want to map ocean characteristics by depth.  Examples might include sound velocity by depth, salinity by depth, density by depth, or perhaps even concentrations of dissolved oxygen, turbidity, or chlorophyll A by depth. 

Due to the expense of cabling, instruments used for vertical profiling are most often - but not always - self-powered, autonomous loggers. Well designed profiling instruments help the user to collect profile data as quickly as possible, without sacrificing data accuracy.   This means that a good profiler should incorporate the following characteristics: 

• rapid sensor response times
• matched sensor response times
• high sampling rate
• sufficient weight to minimize scoping
• adequate flow to the sensors
• uncontaminated flow to the sensors
• rapid wetting time

For accurate sound speed data, measurements should be made close to the acoustic system both, spatially and temporally. Spatial and temporal requirements vary by application, location and contract.  Examples of the variability are:

• Local conditions at a river mouth may dictate sound speed profiles for a multibeam survey be collected every 300 m or 10 minutes.
• A multibeam sonar operating in open ocean may have 25 km or 6 hour sound speed profile requirements.
• Both estuary and open ocean systems may have a requirement for a surface sound speed measurement at the sonar transducer for every sonar ping.

The frequency of sound speed profiles is a tradeoff between survey accuracy and the cost of stopping the survey ship to take the profiles. Areas with a highly variable water conditions require frequent profiles.  Therefore profiles may be required as often as every 10 min when operating at a river mouth.

In the open ocean the surface water conditions vary with solar heating during the day, cooling at night, evaporation, precipitation, waves, upwelling and downwelling. Conditions are less variable and profile frequencies of 2 to 4 hours in the open ocean are more typical.  Many surveyors are now taking advantage of more continuous SV profiles using moving vessel profilers (MVP) or towed undulating profilers.  These systems can provide SV profile updates every few minutes without stopping the survey vessel.