Three-dimensional investigation of buried structures with multi-transducer parametric sub-bottom profiler as part Jens LOWAG, Germany, Dr. Jens WUNDERLICH, Germany, Peter HUEMBS, Germany Key words: parametric, transducer, sub-bottom profiling, acoustics, marine archaeology SUMMARY Coastal and harbour projects like waterway extension, port development or marina construction require extensive hydrographical surveys prior to any dredging activities. Often the typical application is limited to seabed surface measurements with single or multi-beam echo sounders and side scan sonar systems. However the sub-seabed sediments might contain buried archaeological remains which need to be investigated. The multi-transducer parametric sub-bottom profiler SES-2000 mtx provides the possibility to survey the sub-seabed in three dimensions with excellent vertical and lateral resolution. Parametric (non-linear) sound generation is the only available technique to design acoustical systems with small transducer sizes and narrow sound beams at low frequencies. Project and data examples from sub-bottom surveys in German coastal waters will be presented which revealed buried archaeological structures. It has been demonstrated that this new system is a valuable extension to standard hydrographical surveys, if marine archaeological remains are of concern. 1. MULTI-TRANSDUCER PARAMETRIC SUB-BOTTOM PROFILER The presented SES-2000 mtx system is a product variant based on the SES-2000 compact system of Innomar Technologie GmbH. For the parametric sound generation two high frequencies around 100 khz are transmitted with a high source level. These frequencies interact in the water column and new frequency components arise. Especially the difference frequencies which are adjustable between 5 khz and 15 khz are able to penetrate into the sediment and can reveal information about sub-seabed layering and buried objects. The use of parametric acoustics has some significant advantages compared to linear sub-bottom profiler systems: - narrow pencil-like sound beams for low frequencies at small transducer sizes (this system has an aperture angle of +/- 1.8 degrees @ -3dB with an active area of 20 cm by 20 cm only) - virtually no presence of side lobes in the directivity of the low frequency sound beam (Figure 1), which reduces ambiguities during the interpretation of individual reflectors - reduced reverberation noise due to highly focused sound beam - transmission of short signals without any ringing effect at the beginning and at the end of the transmit signal (it is possible to transmit a single sinus cycle for each of the selectable frequencies) - based on the above parameters a very high lateral (small acoustical foot print) and very high vertical (short transmit signals) resolution can be achieved 1/5
- it is possible to work in extreme shallow waters of less than one metre - very high pulse repetition rates are achieved (i.e. more than 50 pings per second) depending on the water depth Figure 1: Directivity diagram for high and low frequencies generated by the parametric transducer of the SES-2000 mtx system, the aperture angle of the parametric frequencies is equal to the primary frequency of 100 khz with +/- 1.8 (-3 db) The property of narrow sound beams allows the configuration of line arrays of individual parametric transducers without overlapping acoustical foot prints (Figure 2). Such a configuration provides a much higher data density compared to a single transducer system. Furthermore the consistency of data is increased due to fixed transducer and sounding point distances. The geometry of the transducer array can be configured depending on the water depth and application (i.e. required data density and resolution). Figure 2: Array beam pattern for SES-2000 mtx transducer line array of three transducers with a separation of 50 cm and a distance of 5 m below transducer surface 2/5
The line array of transducers is rigid and fix mounted to the survey platform. Therefore no complicate offset and attitude corrections are necessary, like for towed systems. However emphasis must be put into the positioning accuracy in order to combine multiple survey lines into a common data set of high spatial resolution. Since transmitter and receiver are on identical positions, no extensive migration processing is necessary for the acoustical signals. Collected data sets are immediately ready for 3D visualization with volume rendering methods after data acquisition. 2. SITE INVESTIGATION AND RESULTS Several sites were surveyed with the SES-2000 mtx system already. To operate this system an accurate positioning system, like an RTK GPS receiver, a true heading sensor and a motion sensor are required. The first data set collected is from an archaeological site in Northern Germany (Schleswig). Wooden pieces have been recovered during dredging works for a small ship channel some years ago. Further investigation by side scan sonar, diving and dating of material has revealed the presence of a submerged wooden barrier within the mud of the Schlei, built during the Viking period. In 2008 this site was surveyed again during a summer camp of the HafenCity University Hamburg for students of hydrography and related subjects. The SES-2000 mtx array configuration had a transducer spacing of 33 cm and an area of 120 m by 40 m was covered with a dense line spacing during about two to three hours. Individual echo plots of the sub-bottom strata show plenty of reflectors below the sediments, but it is very difficult if not impossible to correlate these small targets over dozens of cross sections and to obtain a clear picture of any structure present (Figure 3). Figure 3: Typical echo plot of SES-2000 mtx system from an individual transducer of the line array, frequency of 10 khz and pulse length of 100!s, height of echo plot is 5 m, length of echo plot is 140 m, water depth is about 3 m Data were processed on site, which included quality check, offset correction and interpolation of coordinates for individual transducers and finally the conversion from the non-uniform into a uniform grid of data points. This grid with a lateral cell size of 33 cm and a vertical cell size of 2 cm was visualized with a volume rendering software and has clearly shown that the wooden barrier extends much further below the mud than visible in data sets collected with 3/5
side scan sonar and multi beam systems (Figure 4). The rendered volume can be manipulated in real time, cut planes can be applied and features of interest can be highlighted by the removal of unwanted parts of the data, for example the seabed layer. Figure 4: Time slice and 3D volume of the archaeological site with possible location of debris, total size of the area is 130 m x 40 m Another data example is from a site survey prior to the development of a coastal marina in the Baltic Sea. The planned site contained a historical ship wreck at the bottom surface and it was necessary to investigate the sub-seabed for any further archaeological remains. During a few days of survey the whole area of about 500 m by 500 m was covered and a data set of a few hundred individual cross sections was acquired. Figure 5: Time slices of increasing depth from the marina development area with shipwreck and possible buried artificial structures, total size of the area is 50 m x 50 m Apart from several single reflectors scattered over the entire area an elongated feature with a length of about 80m was found close to the wreck location. Furthermore another few distinct 4/5
and regular features (i.e. triangular and rectangular) were detected below the sandy sediments (Figure 5). The origin and archaeological relevance of these features need to be investigated by diving and excavation work in more detail in future. The third example was acquired during a harbour channel extension project for a port in Northern Germany. Both sides of the existing channel were surveyed in 2010 with the SES2000 mtx system. The aim was to find any buried structures in the historically active harbour area, already used during the time of the Hanse. Some unknown structures were identified and will be investigated by divers in near future (Figure 6). Figure 6: Time slice and 3D volume of the unknown circular structures, total size of the area is 50 m x 50 m 3. CONCLUSION It has been demonstrated that a multi transducer system based on parametric acoustics can be successfully applied to the identification and mapping of buried structures and archaeological objects, even made from wooden material. Narrow sound beams with a small footprint and the high vertical resolution due to short transmit signals allow the generation of 3D subbottom data sets, which can be immediately visualized after data acquisition by volume rendering methods without the need of extensive migration processing. CONTACTS Jens Lowag Innomar Technologie GmbH Schutower Ringstr. 4 D-18069 Rostock GERMANY Tel. +49 381 44079201 Fax +49 381 44079299 Email: jlowag@innomar.com Web site: www.innomar.com 5/5