Archaeology is the study of the human past - the reconstruction of history, cultures and lifestyles through the collection and analysis of artifacts that survive the ravages of time and development. Marine archaeologists are faced with special challenges. The artifacts they seek are hidden from view, often by thousands of meters of water, and subject to decay and degradation from the harsh marine environment. Historical records describing clear landmarks and the continuity of human development often pinpoint the location for terrestrial archaeological investigations, but marine archaeologists are faced with searching vast expanses of featureless ocean with little initial indication of where their targets may be. Given the very poor light transmission properties of seawater, marine archaeology depends on acoustic and other remote sensing techniques, in particular sidescan sonar and marine magnetometers, to carry out the initial search for objects on the seafloor.
The sidescan sonar is the most commonly used tool for the physical search phase of marine archaeological projects. Sidescan sonar is typically deployed in a towed body and produces fan-shaped acoustic beams (broad [typically > 150°] in a swath orthogonal to the direction of travel and narrow [typically less than 1°] in the direction of travel). Pinging at a rapid rate (depending on the frequency and the range of the sonar), the energy returned from this insonification is displayed as a function of travel along track and range across track (converted from acoustic travel-time using a nominal speed of sound in the water column). The result is a plan view acoustic image of the seafloor that is sensitive to changes in topography (mostly through the generation of shadows) and to the composition or small-scale roughness of the seafloor through changes in the amount of energy backscattered to the sonar. Objects of archaeological interest (wrecks or other man-made objects) will sometimes sit proud above the seafloor and thus cast a recognizable shadow or be different enough in composition from their surroundings to present a change in acoustic backscatter.
As with all acoustic systems, the sidescan sonar experiences the typical trade-offs between range and resolution. Low-frequency sidescan sonar operates at frequencies of a few kHz to tens of kHz allowing for the insonification and search of swaths that are kilometers wide. While the ability to insonify many kilometers in a single pass is an efficient means of searching large areas of the seafloor, the resolution of these low frequency systems is such that only very large targets can be found. For example when the 11/12 kHz MR-1 long-range sidescan sonar from the University of Hawaii's Institute of Geophysics found the 245-m- long aircraft carrier USS Yorktown (CV-5) sunk in 5,200 m of water during the battle of Midway, the sonar target was nothing more than a few darkened pixels (Fig. 1).
On the other end of the spectrum, a high-frequency sidescan sonar can have extraordinary resolution, but very limited range. The Klein 5000 dynamically focused high-speed sidescan sonar operates at 455 kHz and can detect objects less than a meter in size, but only over ranges of tens to a few hundred meters. Systems are available at many frequencies and resolutions between these end-members; the appropriate system must be selected based on the particular circumstances of the search.
Even when the highest resolution sidescan sonar is used, however, the data collected cannot necessarily provide an easily interpretable and unambiguous result. It is the unusual case when a sidescan sonar image can be unambiguously identified as a man-made artifact and even rarer when specific details of the artifact can be gleaned from the sidescan imagery. The standard sidescan sonar does not provide information on the depth of the target being insonified (interferometric sonar can provide depth information but this technique is inappropriate for most surveys over wrecks) and thus the images provided by sidescan are created by assuming a linear, monotonic increase in travel-time away from the sonar transducer. This "flat seafloor assumption" leads to a range of distortions when a target that has much local relief (like a wreck) is insonified. Additionally, poor control on the precise position and motion of the sonar tow vehicle leads to other distortions along with the inherent danger of towing a vehicle near a wreck.
The result is a distorted, plan-view image of the seafloor, which contains backscattered reflections off targets on the seafloor and, most importantly, shadows cast by objects on the seafloor. In areas where a man-made object sits proud above a relatively flat seafloor, the backscattered reflection from the object and particularly the shadows cast by the object leave little doubt of the presence of a target (e.g., Fig. 2a). Even in the very high-resolution example shown in Figure 2, however, neither the nature or details of the object are clear.
In environments where the seafloor is rocky and rough, it is often impossible to separate man-made objects from natural features using the sidescan sonar record alone. To aid in the detection of man-made artifacts in these rough environments as well as in those areas where objects may be buried (and thus not detectable by sonar), marine magnetometers are used as an additional search tool in support of marine archaeological studies. A marine magnetometer measures anomalies in the earth's magnetic field caused by ferrous objects. The size of the anomaly will be a function of the size and composition of the object as well as the distance of the object from the magnetometer (typically towed behind a vessel, much like a sidescan sonar). As the magnetometer is towed through a search area, the anomalies can be contoured (assuming that the position of the tow-body is being tracked) providing a coarse picture of the general distribution of ferrous objects on or below the seafloor (Fig. 2b).
While sidescan sonar and magnetometers provide a means to locate potential archaeological targets, neither of these approaches resolves the detail needed for a complete archaeological study. In order to be able to identify marine artifacts, ascertain their state of preservation, make historical inferences, and plan recovery, the marine archaeologist must call upon optical techniques (underwater photographs, video, and when possible, direct observation by a diver or submersible). Optical techniques provide the ultimate level of resolution (Fig. 3a) but are severely limited by the attenuation of light in most marine environments. Cameras must be deployed, or direct observations made, within a few meters of the object. As a consequence, the field of view of most optical images is rather limited, often making it difficult to extract the context of the image (Fig. 3b) and thus get an overall feel for the nature of a large target. Mosaicing techniques applied to wreck imagery allow larger areal coverage (e.g., Singh, et al., 2000), but these techniques are complex and often very time-consuming.
In this paper we explore another approach to marine archaeological studies, the use of a new generation high-resolution multibeam sonar. This new type of sonar uses dynamically focused beams to collect extremely detailed bathymetric data over relatively long ranges (as compared to optical systems); when operated to hydrographic standards and combined with state-of-the-art visualization tools, a quantitative 3-D image of the targets and the surrounding seafloor can be generated at a resolution that addresses many of the key questions posed in an archaeological study. The feasibility of this approach was first demonstrated in a recent study of the scuttled WWI German High Seas Fleet off Scapa Flow, Scotland (http://www.ccom.unh.edu/project/scapamap). Here we use examples from the ongoing NHC survey of the D-Day beaches off Normandy, France to demonstrate the tremendous potential of this approach for marine archaeological investigations.