Shipwreck Discoveries and their Analyses
During a month-long survey of the coastline along the south-eastern Bozburun peninsula, Turkey, nine shipwreck sites were discovered. Of these, five have historical significance and represent a chronological range from the Roman Imperial to Renaissance periods. This article provides a description of the sites and associated artefacts, and attempts a provisional analysis for each wreck’s operational date as well as the nature of the finds in their historical context.
The picturesque Turkish coast features a profusion of finger-like projections where sparsely-vegetated cliffs are battered by waves. On the south-easternmost portion of one particular isthmus in the Aegean Sea, the Bozburun peninsula, the ancient settlements of Physkos, Lorima and Tios were founded. Over the centuries, this segment of coast was part of an active trade route between the ancient cities of Rhodes to the south and Knidos to the west, and part of the greater Aegean and eastern Mediterranean mercantile network. During the summer of 2005, a coastal survey of this shoreline was undertaken to locate and document submerged cultural remains. This co-operative project was carried out by the Turkish Ministry of Tourism and Culture, represented by Commissioner Gulnaz Savran, the Institute of Nautical Archaeology (INA), and RPM Nautical Foundation (RPM), a non-profit institute dedicated to nautical archaeology research. Founded by INA director George Robb junior in 2000, one of RPM’s aims is to support INA projects all over the world.
INA has been carrying out expeditions along this section of coast over the past 40 years. During the summers of 1965, 1967 and 1968, Dr George Bass led survey expeditions in response to reports of archaeological finds by sponge fishermen during the 1950s, noted by Peter Throckmorton. In 1968, 26 of 145 side-scan sonar images obtained in previous seasons were investigated and a scattered, unidentified wreck was located at 100 m (Bass, 1976: 29–30). Bass led subsequent side-scan surveys along the south-west Turkish coast, including the Bozburun peninsula, in 1973, 1974, and 1980, and documented an apparently-looted wreck-site near Marmaris in 30 m, as well as several near-shore dump sites, including one near ancient Loryma (Rosloff, 1981: 277–81; Bass, 1982: 45–7). Also located along this section of coast is the small Bay of Serçe Limani where Drs Bass and van Doorninck excavated the 11th-century Byzantine glass wreck in 1977–79, and where Cemal Pulak excavated a Hellenistic wreck in 1978–80 (Pulak and Townsend, 1987; Bass and van Doorninck, 2004). More recently, in 2004, Jeremy Green of the Western Australian Maritime Museum and Faith Hentschel of INA returned to the area from where a bronze statue was reported to have been raised in 1953. A limited side-scan survey produced numerous anomalies; unfortunately, equipment problems and poor sea-conditions hampered the verification process (pers. comm. Jeremy Green, 2005).
2005 objectives and survey area
Considering the advanced technology available, the nature of the survey area according to Green, and the information on reported wreck-sites in a wider area provided by George Bass and Tufan Turanli of INA, the scope of the 2005 campaign was significantly expanded from previous expeditions. Our objectives were to conduct a systematic multibeam survey of the entire southeastern coast from near-shoreline out to a depth of 100 m, and subsequently to locate and document all significant cultural deposits. The survey area extended approximately 37 km from Kadirga Burun at the north-east, just outside the Marmaris approaches, to Bozuk Bükü near the peninsula’s south-western end (Fig. 1). Over 120 km2 were completed, from close inshore to the 80-m contour. The vast majority of this designated area had not been surveyed previously. Considering the greatly increased scope of the 2005 survey relative to previous seasons, the objectives no longer centred on the search for a single hypothesized vessel. Rather, the goals were to locate, document, identify, and assess all submerged archaeological sites within the designated survey area.
This portion of coast is dominated by cliffs which plummet into the sea to depths of 30–50 m. Thereafter, a sandy sea-floor with a relatively gentler gradient is typically encountered until reaching the Rhodes channel. Exposed rocks and small islands dot the coastline, forming natural hazards for maritime traffic. Two particularly interesting small bays within the survey area are Bozuk Bükü, at the end of which the ancient city of Loryma was situated, and the bay of Serçe Limani, which has produced several noteworthy shipwreck sites.
Survey was conducted by RPM Nautical Foundation’s two research vessels: the R/V Hercules and R/V Juno. Both are equipped with multibeam echosounders among other remotesensing, verification, and analysis equipment. Based on field experience and the nature of sea- floor in the survey area, the multibeam systems were deemed to have the best potential for locating cultural resources. A dual-head system for depths up to 100–120 m is fixed to the Hercules, and a single-head system for depths up to 45 m on the Juno. Accordingly, the Juno surveyed the areas from the coastline to the 45-m contour, while the Hercules surveyed the deeper area. Multibeam survey provides three-dimensional data that can provide highly-detailed topographical maps of the sea-floor, making it possible to exclude many of the geological anomalies which often plague two-dimensional side-scan images, as well as providing a better overall context for all anomalies. This is important as a pile of amphoras or ballast-stones appears very similar to geological formations, which results in significantly more spurious anomalies in side-scan images than in those from multibeam. Moreover, low-profile mounds formed by shipwreck sites can more easily be missed in side-scan survey, and positioning information for multibeam data is much more precise. The 2005 multibeam survey area, therefore, included the entire coastline regardless of previous side-scan survey work.
Multibeam data was processed onboard the R/V Hercules and subsequently reviewed for potential shipwreck sites, which were plotted for investigation with the remote operated vehicle (ROV). Outfitted with still and video cameras, lights, and sonar, ROV deployments always recorded video for documentation; where wrecksites were located, both still and video photography were used. An experimental laser device attached to the ROV provided a photographic scale that assisted in the identification of individual objects and the construction of preliminary site-plans.
To date, a total of 77 anomalies have been identified in the multibeam data. Each of these was assessed on their resemblance to geological formations or potential cultural remains, in order to prioritise verification efforts. Two anomalies located by the Juno, designated as sites TK05-AA and TK05-AG, were clearly discernable as shipwrecks in the multibeam imagery. Both vessels appeared modern, intact, and had little sign of burial, so were designated for diver verification at a later time. Of the remaining 75 anomalies, 29 (39%) were investigated by ROV during the 2005 field season, and 7 of these 29 (24%) were shipwreck sites. The following is a preliminary assessment for these nine shipwrecks.