Sidescan Sonar Data


The sidescan sonar data sets for the 2km x 2km area were divided into eight data sets. This was to show the effect that frequency and range setting have on the ability to detect small or ephemeral sites.
The sidescan sonar used for this project was capable of collecting both high and low frequency data simultaneously. The high frequency setting is able to detect small object and provided detailed images of objects on the seafloor but the energy is not normally capable of getting data at ranges greater than 100m. The low frequency setting is able to provide data out at much larger ranges than the high frequency but at a lower resolution.
The high frequency data set produced the best quality images for the interpretation but only at short range settings (50m or 75m). This is because the high frequency acoustic energy is attenuated before it reaches the end of the longer range settings.
The low frequency data could be transmitted to the end of the 150m range. Although this was suitable for geological interpretation, it was not suitable for detecting the small objects required for archaeological interpretation.
While it was still possible to detect the presence of small objects on both frequencies and at most ranges, only the data sets with high frequency and short range settings allowed the interpretation of whether an anomaly was natural or artificial in origin.
The position of ephemeral sites is also affects whether the sites will be detected, as is shown by the Thomas Lawrence.
Sites like this wreck, with low relief and lying in areas of complex bathymetry, can be difficult to identify as they will not always reflect enough acoustic energy to produce an anomaly. Such sites highlight the need for every anomaly in the data to be recorded and not just those which are identified from more than one survey line.
This means that it is not possible to determine whether an anomaly is a real acoustic signal or noise by simply determining if the anomaly is observed from two adjacent survey lines.
While it has not been possible to develop a system for the automatic discrimination of anomalies in sidescan data as part of this project, it was possible to get an impression of the scale of work that would be involved.
Any such system would have to be able to learn the difference between geological and archaeological anomalies and also discriminate against acoustic noise.
On this basis, it is likely that archaeological assessments of sidescan sonar data are going to continue to be best conducted by experienced geophysicists rather than by software or by people inexperienced with the problems and practicalities of sidescan sonar data.
A final factor highlighted by this survey was that in areas where the sea floor shoals rapidly along a survey line, it is not always possible to keep the towfish at the optimum height above the sea floor.
Where the towfish gets too high, the range is effectively shortened and so the expected data coverage is not achieved. It is advisable that the line spacing is less than the expected minimum range so as to ensure that all areas of the sea floor are surveyed at least twice.
The greatest number of anomalies found in any of the eight sidescan sonar data sets was in the high frequency, 50m range data. All the other data sets had less resolution and so fewer anomalies were identified. This implies that surveys specifically designed for archaeological purposes should use high frequency sidescan systems at low range settings. This project has shown that by using high frequency sidescan sonar systems at either 75m or 100m ranges up to two-thirds of the anomalies within an area may be missed.
At short range setting the low frequency setting did detect slightly less anomalies than the high frequency data set. The main difference between the resolutions of the two data sets was that certain anomalies classified as wrecks in the high frequency data set were described as seafloor disturbances in the low frequency data sets.
The interpretations of different sidescan sonar data sets underline the need for sidescan sonar surveys to be carefully designed so that the correct range and frequency are selected to detect the smallest anomalies expected to be of interest in the survey area. Also the combination of range setting and line spacing needs to be such that at least 220% coverage of the area is achieved so that all areas of the seafloor are ensonifed at least twice, from different orientations.