Technical Description - System Capabilities

This section describes the performance and capabilities of Bathyswath in various survey situations.


Grazing angle and spreading loss

Line spacing is the distance between adjacent survey lines. The spacing is determined by the sonar horizontal range expected at that depth, and the amount of overlap required. The horizontal range expected depends on the water depth under the sonar-head, as well as the seabed type and the sea state.

The term "Horizontal range" is used to describe the sonar coverage from one transducer. For a twin transducer configuration, the total swath width, from port edge to starboard edge, is therefore twice this range.

The horizontal range is limited by two factors: grazing-angle and spreading loss. The grazing angle limit is related to the angle that the sound ‘beam’ makes with the seabed.

Directly under the transducers, sound is reflected directly, and there is little loss when sound is scattered by the seabed.

Moving away from the transducers, much of the sound is reflected away from the transducers, but enough sound is scattered back for the seabed to be properly detected.

At the grazing-angle limit, the sound makes a very small angle with the seabed. Most of the sound is reflected away, and the signal scattered back from the seabed is too small to be detected. Actually, the configuration of the Bathyswath transducers is similar to sidescan sonars: the swath width ends at the point where the backscattered signal is not sufficiently above noise to enable detection of the angle of the backscattered wave.

Backscattered acoustic signals from a seabed generally follow “Lambert’s Rule”; that is, the backscattered signal falls with the square of the cosine of the angle of incidence of the acoustic “ray”.

Plotting this rule, we can see that the Lambert’s Law function is low at depth to swath width ratios above about 15:1.

In poor seabed conditions or turbid waters, this ratio can be reduced to about 10:1 or even less. Bottom types such as soft mud or peat can indeed reduce the expected range by as much as 30%. Sand, rock and shingle all give good sonar backscattering.

The spreading loss limit is simply caused by the sound spreading outwards, and being absorbed by seawater. The rate of absorption is related to the frequency of the sonar signal. The spreading loss limit is thus determined by the distance from the transducers to the farthest point on the seabed (the slant range).



(1) There is no real physical limit to the depth at which measurements can be made by an interferometric system. Close to the transducer, the sonar system is operating in the near-field domain, where sound rays are not yet fully formed. This means that the accuracy of the angle measurement is not as good as a fraction of the sonar range as it is in the far field zone, but the very small ranges (within the 15-times depth limit) means that the depth measurement accuracy remains easily within that specified. A practical limit is close to that of the size of the transducer itself.

(2) The operational slant range is the one you should get in most cases. The maximum slant range is the one you should get in the best environmental conditions.

Swath width and coverage angle

The following table shows the optimum water depth for which the maximum swath coverage angle and width are reached without being limited by the operational sonar range.



Elevation angle

Bathyswath is typically configured with two transducers, one facing port and the other starboard, and with both transducers pointing downwards at 30° from vertical (recommended). Each of the two transducers can, in principle, measure the angle of any sonar wavefront that approaches its front face, from -90° to +90° of its normal. If the two transducers were pointing horizontally, the angle of coverage would be a complete 360°.

The 30° elevation angle means that the top 30° is not covered on each of the port and starboard sides. In addition, the shape of the transducer beam in elevation means that the signal is weak for the last 20° of the -90° to +90° coverage. Therefore, we calculate the angular coverage range to be (360° - 2 x 30° - 2 x 20° =) 260°. This configuration also gives 20° overlap in the nadir region between coverage of the 2 transducers.




Consideration must also be given to the accuracy required from the survey. Bathyswath is essentially an angle-measuring instrument, so that depth accuracy reduces with horizontal range. The angular accuracy of both the Bathyswath sonar and commonly available MRUs is better than 0.05 degrees. The accuracy of the combined system is thus better than 0.1 degrees. The maximum range required for a given depth accuracy can easily be calculated. One accuracy specification is that of the International Hydrographic Organisation (IHO) S44 specification [Ref. 1]. Bathyswath has been used, and quality checked, in surveys at all IHO S44 accuracy orders, including Special Order.

This section considers the accuracy and resolution of Bathyswath; these two parameters are closely related, and can be selected for, against each other, using statistical methods in data processing. It is derived from a model of the accuracy of Bathyswath, related to resolution. This model has been validated using data collected and published for the 2008 Shallow Survey conference. Bathyswath can output pulses down to 2 cycles (A single-wavelength pulse will not have sufficient transfer efficiency into the water), so the physical limit of measurement resolution is as per the following table.

(1) Bathyswath is a single frequency (also known as Continuous Wave – CW) system; the sonar frequency is built into the electronics systems. The frequencies are derived by dividing a 60 MHz base frequency in powers of two (i.e. 60000 kHz / 29 = 117.1875 kHz)

(2) Any sonar system can detect objects that are larger than half of its wavelength. The limit to measurement resolution is determined by the sonar pulse:

(3) These are theoretical values, but practically there is no way any sonar could tell the difference between a seabed at 20.000 metres and another one at 20.003m.


The software allows the user to select between 28 to (215 + 214) samples.

For a 234 kHz system, at 50m with 4096 samples, we get 12 mm sample interval (so 6 mm at 8192, 3mm at 16384 and 2mm at 24576; halve these values for a 468 kHz system).

However getting a very low interval by collecting huge amounts of data is meaningless and would in some cases reach USB buffer limits.

Filtering, Accuracy and Resolution

Interferometric swath bathymetry systems give many data points per side, typically 2000 to 8000. However, the spread (standard deviation) of raw data points is usually greater than that of beam-forming multibeams, which typically produce 100 – 200 points per side.

Software filtering can be used to reduce this standard deviation to internationally acceptable survey limits. However, this filtering also reduces the resolution of the filtered data.

The International Hydrographic Organisation (IHO) Special Publication number 44 [Ref. 1], sets out standards for both accuracy (termed “uncertainty”) and resolution. These are defined in a set of “Orders”, of which “Special Order” is the most stringent, followed by Order 1.

Accuracy is specified in terms of the depth uncertainty to 95% confidence.

Resolution is specified in terms of the size of objects that must be detected, and the number of acceptable data points per square meter.

Beam Width, Azimuth

The beam width of the transmit and receive elements of Bathyswath, for the three sonar frequencies available, is given in “Beam Width, Azimuth” below. These figures are for both the transmit and receive elements. The effective width of the sonar “footprint” on the seabed is found by combining the “footprint” of both the transmit and receive staves, which is done by halving the beam width of the separate staves, to get a “two-way” azimuth angle.

Total Propagated Error

This analysis is concerned only with the depth error component of the sonar system. Other error components include position, attitude, heading and height relative to datum. The last of these is usually measured using GPS height or tide height, and is often the largest component. This discussion concerns itself only with the contribution of the interferometric sonar.

Estimating Depth Uncertainty

This analysis uses Bathyswath sonar depth profiles extracted from the data sets that were submitted to the 2008 Shallow Water data set. Profiles are analysed at two Bathyswath frequencies: 468 and 234 kHz.

Depth error is first estimated by comparing raw data points with an averaged profile.

Data Smoothing

The Bathyswath software includes a range of data filters. These include averaging filters, which smooth the data, thereby reducing the statistical spread (and thus the uncertainty) of the depth data, but at the expense of resolution.

This smoothing filtering is mathematically modelled using a Gaussian sliding window filter.

The width of this window is selected to satisfy the S44 requirement for a given number of accepted data points per square meter. Both across-track (along the profile) and along-track (profile separation, from vessel speed and ping repetition frequency) are used.

The smoothed points are converted to 95% uncertainty, using the statistical method recommended in IHO S44.

< 95% Uncertainty of smoothed data.


Error Modelling

The data smoothing procedure is repeated for a number of “pings” in the data set, and averaged by range.

A mathematical model of phase error in Bathyswath is created, and validated against the observed uncertainty; see the blue line on the left.

The process is repeated for all three Bathyswath sonar frequencies.


Error modelling, validated with data from the Shallow Survey 2008 data set, shows that Bathyswath is capable of providing survey depth accuracy within the requirements of IHO S44, Special Order, at good ranges. Longer ranges are achieved with the lower-accuracy Orders.


The time in which a given seabed area can be surveyed depends on the distance between the survey lines and the forward speed of the survey vessel.

Line Spacing

The spacing between survey lines is determined by a combination of range limit and accuracy required. There must also be some overlap allowed to account for variations in the survey line followed. Otherwise, any small helmsman’s errors will cause gaps in coverage of the seabed.

Pulse Repetition Frequency

The time taken for a ping cycle is that for a round trip from the transducers, to the farthest range, and back again. The speed of sound in water is about 1500 meters per second. For example, a 150 meter ping takes 0.2 seconds. This gives a pulse (or ping) repetition frequency (PRF) of 1 / 0.2 = 5 per second (or Hertz).

The software allows the nominal sonar range to be set in meters. The corresponding PRF is calculated in software and used in data acquisition.

Platform Speed and Along-Track Coverage

Bathyswath is capable of providing total ensonification of the bottom at practical and efficient survey speeds.

The distance between pings along the track of the vessel is determined by the pulse repetition frequency (PRF) and platform speed. In order to minimize cross talk between the two sides, the system can be used with alternating sonar transmissions, port and starboard. Thus, in the alternating mode, this distance is doubled:

Bathyswath also provides the option of firing both transducers simultaneously. This doubles the coverage rate, so that the along-track ping spacing reduces to (V.PRF). However, this mode should be used with caution in surveys with a requirement for high bathymetry accuracy, because some cross-talk between the channels is likely. That is, the signals from one side can affect the other side.

Coverage is also determined by the width of the sonar beam. A narrower beam gives better resolution, but carries a greater risk of missing targets between beams. See section 8.3.4. At 50 metres, the 234kHz and 468kHz beams cover 0.43m, and the 117kHz beam covers 0.74m.

Increasing the speed over the ground will reduce survey time, but will also reduce the along-track coverage. Five or six knots is generally a good compromise. At 5 knots, with a 100m range, giving 6.7 pings per second, each side is covered every 75cm along-track. At 10 knots, this spacing doubles. In the 5 knot example, ground is covered at 500 square metres per second, or 1.8 square kilometres per hour.

When using the system in simultaneous pinging mode, on a typically flat seabed, the directionality of the two transducers is sufficient to prevent the signals from one side appearing on the other. However, if one side contains a very strong reflector (e.g. a harbour wall), or is very weak (e.g. contains acoustic shadows), then there can be “cross-talk” between the sides. The operator needs to be aware of the risks and priorities. Typically, it may be safest to use alternating mode where bathymetric accuracy is paramount, and simultaneous mode when using the system to detect small objects on the seabed. When high coverage is required in a limited area, and channel cross-talk is a problem, it may be beneficial to ping on one side only, thus doubling the along-track coverage on that side.



Some users have a requirement to survey at relatively high speeds, 8 knots or more. At excessive speeds, along-track coverage is reduced, and data density could fall below specification or the sidescan imagery might not fully cover the seabed. This section shows the options for meeting this specification with Bathyswath.

Data Density

Survey data specifications such as IHO S44 [Ref 1] define a minimum detection resolution, which is often interpreted by hydrographic organisation in terms of sounding density. For example S44 Special Order surveys require around 9 soundings per square metre, and Order 1a requires round 9 soundings in a 2x2 metre patch, so 2.25 soundings per square metre. Special Order is reserved for limited areas such as docksides, where high vessel speeds are not usually allowed. Therefore, we use the Order 1a requirement for this analysis. Bathyswath can provide S44 accuracy at around 3 soundings per metre across-track (see section 8.3), so it is necessary to achieve no fewer than (2.25 / 3 =) 0.75 pings per metre along-track.

For typical 234 kHz Bathyswath swath widths (12 times water depth up to a maximum of 300 m), the speed at which Order 1a data density is just met can be calculated; this is 13 knots at depths below 25 metres.

For Special Order data density, the maximum speed falls to 3.2 knots with 12 times water depth swath widths for depths below 25 m.

To achieve Special Order data density at high speed in depths below 10 metres, the swath width must be reduced.

Complete Bottom Coverage


If it is necessary to find all sonar targets, no matter how small, then full bottom coverage is needed. This is determined by the area covered by the sonar beam on the seabed, called the “sonar footprint”. A sonar beam is generally taken to extend to the angle at which the power at the centre of the beam falls to half its power, or 3 decibels (dB). For resolution purposes, a “two-way” angular resolution is quoted; see section 8.3.4 below. In this analysis we assume that detection of seabed targets of interest is achieved within the “one-way” azimuth angle.

The footprint is narrowest under the transducers, and widest at far range. Under the transducers it is wider in deeper water.

< Beam Footprint vs. Horizontal Range, at 10m and 50m water depth.


Maximum Speed for Full Coverage

The maximum speed for full coverage can be calculated for a given water depth using the ping rate for that depth, and the footprint of the beam below the vessel. The maximum speed is that at which the vessel moves forwards by the width of the footprint within the ping period. Bathyswath is usually operated in simultaneous ping mode, in which both transducers fire at the same time. This doubles the along-track coverage at a given forward speed compared to systems that work in alternating mode (port-starboard-port-starboard etc.).

This graph shows the maximum speed at which full coverage is achieved immediately below the boat, operating in simultaneous ping mode.

< Max. speed for full coverage below boat.


It can be seen from these graphs that full coverage under the boat can be achieved at ten knots water depth if:

  • The swath widths are kept to five to six times water depth.
  • The sonar is operated in simultaneous mode Sidescan.

Search Mode

Most sidescan systems are run with every other line spaced for 100% overlap, so that the nadir of each swath is covered by the far range of at least one adjacent line. Bathyswath can be run in this way to give full bottom coverage with both bathymetry and sidescan.

The maximum speed is given by the footprint at far range and the ping period:


These calculations show that the Bathyswath sonars can easily provide both IHO S44 data density and 100% bottom coverage at relatively high survey speeds. Bottom coverage depends on the along-track and across-track beam patterns. Bathyswath provides very high data density across-track, so that coverage is complete in this dimension. The along-track coverage depends on ping rate, beam width, and speed of sound.

There are several options for obtaining 100% coverage:

Operate the system with minimal overlap, limit swath widths to approximately seven times water depth, or

  • Operate the system with 100% overlap between adjacent survey lines. In this case, 100% coverage is obtained in 10 metres of water at 15 knots.
  • Add a third, forward-facing transducer to boost data coverage in the nadir region. Click for Page.

There is a range of options lying between these solutions, so that the system can easily fit in with the operational practices of the user.


The Nadir Region The region of the seabed directly below the sonar transducers is called the “nadir region”. This coverage of an interferometer is greatest at medium ranges from the transducers, and is thus less in the area close to the centre than at far range. There are several reasons for this reduced data density in the centre.

Measurement Geometry

An interferometer samples the angle to the seabed at regular intervals of time. Each angle and time pair is converted to a depth and horizontal range pair. Time is converted to range from the transducer (slant range) using the speed of sound, so regular time steps translate to regular range steps. However, regular steps in slant range do not produce regular steps in horizontal range. Horizontal range is a distance measured along the seabed, starting from a point immediately beneath the transducers.

A simple example illustrates this situation. Consider the system operating in a water depth D. The first depth measurement is taken immediately under the transducers. The next measurement is taken at a range step of dR. Simple trigonometry dictates that the horizontal range step, dH1, is much larger than dR. Now consider the situation further out along the profile, at some horizontal range H. Here, the horizontal range step dH2 is much closer to the range step dR.

Footprint of Transmit Beam

At the start of each ping, the sonar transmits a short pulse of sound. This pulse moves outwards at the speed of sound. Where the pulse hits the seabed, it returns an echo. The echo is picked up by the transducers and the angle of the returning signal is measured.

At any one instant, the pulse of sound will be “illuminating” a patch of seabed. The size of this patch is determined by the length of the pulse and the geometry. Immediately below the transducers, this patch is at its greatest, and thus the resolution is lower than it is further out.

Amplifier Response

When the sound signal first hits the seabed, the size of the returned echo signal goes from a very low level to a very high one extremely quickly. This fast change in signal level presents a challenge to the amplifier designer. One important part of a successful interferometer like Bathyswath is thus the way in which the amplifier responds to these fast signal changes.

Filter Response

As explained above, an interferometer measures angles to the seabed as a set of samples separated in time, and thus in range. Before the sonar signal reaches the seabed, the angles measured will be discarded due to low signal levels, or random due to noise pick-up or returns from objects in the water. These random signals from the water-column must be discarded before the seabed depths can be recorded. Bathyswath uses a collection of user-settable filters to separate the seabed from objects in the water column.

Deep Water Response

The shape of the sonar transducer beam in elevation has been chosen to maximize the performance for most survey situations. In shallow to medium water depths, the direct reflection from the seabed directly under the transducers is very strong, and can cause the electronics to ‘saturate’.

To reduce this, the sonar beam is shaped so that returns from this region are reduced in amplitude. However, at the limits of the depth capabilities of the system, this reduction can cause data-loss in the near-range area. Changing the transducer beam-angle so that the transducer normal makes a greater angle with the horizontal reduces this effect.

Coverage at Nadir

Most users find that the data coverage achieved in the nadir region is sufficient for their purposes. However, if very high data coverage is required at all parts of the survey, and if the sidescan component is also important, then a “sidescan survey” line pattern can be used. Survey lines are run alternately at the sonar range and twice the sonar range, so that the nadir is “filled” from an adjacent swath in every case.

Forward-Looking Transducer

Bathyswath can be supplied with a third, forward-facing transducer to boost the data density at nadir. See section 10 for more information.

Use with MBES Systems

Some users operate Bathyswath at the same time as an MBES system. The MBES provides more data in deeper water and helps to boost the nadir data density, and Bathyswath gives much greater swath width in shallow water and gives true, high-resolution sidescan imagery. Acoustic cross-talk interference between the two systems is usually acceptable, but Bathyswath can provide or accept sonar transmit synchronisation signals if these are required.


The standard Bathyswath systems use sonar frequencies above the limit of hearing of marine mammals. Only Bathyswath-XL uses a sonar frequency and transmitted power levels that are in the range that could disturb or cause injury to marine mammals. Bathyswath is fitted with mitigation features, including:

  • Programmable sonar soft-start: this slowly builds up the transmit power levels, allowing any marine mammals in the vicinity to move away before the sonar can cause them injury
  • Sonar transmit power linked to pressure depth: for use on deep-operated vehicles, this limits the transmit power until it has reached sufficient depth to be out of the diving range of marine mammals. This also helps to avoid cavitation issues on the transducers.


Bathyswath samples a high density of data. It therefore stores data to disk at a high rate. When recording a full set of bathymetry and amplitude points, the recording medium is filled up at a rate of 150 Kb per second (0.54 Gb per hour and 13 Gb per day, so a 1Tb external disk drive can store 11 weeks of continuous recording). If the data recording rate is a problem, the user may select a lower data sampling rate. This will save recording media but at the expense of resolution. Conversely, very high resolution surveys use more media.