INVESTIGATION INTO MARINE TRAFFIC AND A RISKY AREA IN THE TURKISH STRAITS SYSTEM: CANAKKALE STRAIT

!e Turkish Straits comprising the Strait of Canakkale, the Strait of Istanbul and the Sea of Marmara are unique in many respects. All dangers and obstacles characteristic of narrow waterways are present and acute in this critical sea lane. !is research reveals the simulation of Canakkale (Dardanelle) Strait under di#erent tra$c conditions and identi%es risky areas. !e results of this simulation show that an increase of 25% in the existing tra$c grows 43 times in the number of waiting ships (from 1.663 to 73.73), whereas waiting time increases 29 times (from 24.267 to 737.07). As a result of simulations and risk analysis, it is found that Nara turning point is the bottleneck point of the strait due to its topographic structure and the current system.


Introduction
e Turkish Straits System (TSS), consisting of the Marmara Sea, the Strait of Istanbul (Bosporus) and the Strait of Canakkale (Dardanelles), are very complicated and narrow waterways connecting the Black Sea to the Mediterranean Sea (see Fig. 1). It is an established fact that the Turkish Straits are one of the most hazardous, crowded, di cult and potentially dangerous waterways for marines in the world. e Turkish Straits located between the Black and Mediterranean Seas are 164 nautical miles (nm) in length and have unique physical, geographical, hydrological an oceanographic characteristics and complicated navigational conditions prevailing in the area. e Turkish Straits form a waterway of strategic and economic importance since being comprised of Istanbul and Canakkale Straits and the Sea of Marmara. As the only water route between the Black Sea and the Mediterranean Sea, the Turkish Straits both geographically and metaphorically connect Europe to Asia. e TSS are the most unique amongst other straits due to di erent physical, hydrological and oceanographic characteristics as well as because of complicated navigational conditions. e length of the Strait of Canakkale is about 37.8 nm with a general width ranging from 0.7 nm to 1.08 nm (see Fig. 2). A very sharp course alteration is needed at the narrowest point at Nara turning (approximately 90 degrees), the westernmost section of the waterway that divides Europe from Asia and connects the Mediterranean and Black Seas. e strait, through the Sea of Marmara, is a narrow, winding passage anked on the north by the Gallipoli peninsula. Due to the geography and an increasing volume of tra c of the straits, the existence of harbours is of real danger for the safety of passage and navigation, life, property and the environment. Big ships and especially tankers face serious di culties when navigating the sharp turns because of geographical structure, strong surface and undercurrents of the strait.
Despite unpredictable weather and swi surface currents, the Dardanelles has been a strategic water route and an object of conquest throughout history. e aim of this paper is to investigate marine tra c at Canakkale Strait and to identify the risky areas of the place. e paper has been divided into two main sections. First, a description of tra c at Canakkale Strait and second, simulations of marine tra c at the strait are provided.

e Current System of Canakkale Strait
e current system consists of two parts. e upper part runs from Marmara to the Aegean Sea, and below, Aegean salty waters run with the speed of 50 cm/s. e upper current sometimes reverses due to the shape of the coast and meteorological conditions. ese reverse currents are more visible in the middle and south part of the strait. Surface currents up to Nara are about 1.5-2 knots, whereas at Nara and Kilitbahir they reach about 4 knots (Black Sea Pilot 1990). Ors and Yılmaz (2004) modelled the current system at Canakkale Strait (see Fig. 3). e Fig. 3 shows that the current change at Nara turning point is obvious which would also negatively a ect navigation at this point.

Maritime Tra c in the Strait
Maritime tra c in the Turkish Straits is exceptionally dense due to merchant tra c, coasters, shing vessels and local tra c crossing the strait and causing di culties in the navigation of the transit passage. Such dense tra c includes the transport of noxious, dangerous and hazardous cargo (oil, LNG, LPG, chemicals and other explosive and environmentally hazardous substances). e volume of tra c is expected to increase by 40-50% with additional tra c coming from the Main-Danube, Volga-Baltic and Don waterways. Traffic congestion will further increase in oil supply and the volume of foreign trade from the Black Sea states and neighbouring countries.

Maritime Tra c Regulations in the Turkish Straits
Along with the introduction of the Regulations, the Turkish authorities have also established 'Tra c Separation Schemes' (TSSc) in the Straits, in accordance with the provisions of 'International Regulation for Prevention of Collusion at Sea' (COLREG). TSSc was approved by the International Maritime Organization (IMO) General Assembly in November 1995, in association with 'Rules and recommendations on navigation through the Strait of Istanbul, e Strait of Canakkale and the Marmara Sea' (IMO 1995;e Strait of Istanbul … 1995). e sectors at Vessel Tra c System (VTS) and tra c separation at Canakkale Strait are shown in Fig. 4 (IMO 1995) indicating that the strait consists of 3 sectors including Gelibolu, Nara, and Kumkale. Nara is the narrowest sector and has a sharp turning point.
Slowing Down (Turkish Straits Vessel … 2008): • An immediate notice should be given to the TSVTS Centre when a vessel is forced to slow down in either Istanbul or Canakkale Strait. e TSVTS Centre shall assess the tra c situation and provide information, recommendations and instructions regarding the situation. Overtaking (Turkish Straits Vessel … 2008): • Vessels shall not overtake another vessel unless there is an absolute necessity. In case of such necessity, a vessel intending to overtake another shall inform the TSVTS Centre prior to commencing the overtaking. e TSVTS Centre shall  Fig. 4. e tra c separation system at Canakkale Strait assess the tra c situation and provide information, recommendations and instructions regarding the situation. • E orts shall be made to overtake the vessel in a single manoeuvre. However, overtaking shall not take place between Nara and Kilitbahir in Canakkale Strait. Visibility (Maritime Tra c Regulations … 1994): • Whenever visibility is 2 NM or less in any part of the Strait, the vessels passing through the Strait will keep their radar turned on constantly to provide radar readings. On the vessels with two radars, one of those will be assigned to the pilot's usage. • When visibility is 1.5 NM or less in any part of the Strait, the vessels the radar of which does not provide complete display ability shall not enter the Strait. • When visibility in the Strait is 1 NM or less, the vessels carrying hazardous cargo and large vessels shall not enter into the Straits. • When visibility in any part of the Strait is 0.5 NM, maritime tra c shall be open in the appropriate direction and closed in the opposite. In such situations, only vessels less than 100 meters in length and carrying no hazardous cargo can navigate in the direction open to tra c. • When visibility in any part of the Strait is less than 0.5 NM, the tra c ow in the Strait shall be closed in both directions. • When visibility in the Strait is suitable for navigation, the arrangement and order of entering the Strait shall be determined and noti ed to the waiting vessels and persons concerned by the Tra c Control Centre. • When a large vessel with hazardous cargo enters the Strait, a similar vessel approaching from the opposite direction may not enter the Strait until the previous vessel has exited. ere shall be a distance of at least 20 NM between two such vessels proceeding in the same direction. • e competent authorities may temporarily suspend two-way tra c and regulate one-way trafc to maintain a safe distance between vessels. As can be seen from Fig. 5, a large vessel during manoeuvring can violate the tra c separation lane due to the sharp turning point at Nara and Kilitbahir.

Statistics on Canakkale Strait
e passage through Canakkale Strait is given in Table 1, (UMA 2006). In 2006, 48915 vessels in total passed the Strait of Canakkale. e total numbers of tanker passages are 9567 from which LPG and LNG carriers make

Material and Methods
A simulation model can be used for determining the effects of changes (scenarios). For example, Hayuth et al. (1994) used a simulation model to evaluate the future of the port and ensure optimum investment strategies. In this study, the simulation language AWESIM was used as primary modelling tools (AweSim! User's Guide 1997). iers and Janssens (1998) made detailed models of tra c on the rivers, including navigation logic, tides and lock planning. Köse et al. (2003) investigated marine tra c at Istanbul Strait. ey found the bottleneck points at the strait. Demirci (2003) investigated port and new investments using AWESIM. Somanathan et al. (2009) simulated passage tra c at the North of Canada.  (2006). ree sub-systems were used to simulate the system: • tra c ow in direction 1 (from Marmara to the Aegean Sea); • tra c ow in direction 2 (from Aegean to the Marmara Sea); • two information systems representing big ships and the simulation of bad weather conditions. Each of these processes is modelled considering the movement of the entity through a sub network. Big ships (ship length L > 200 m) are modelled by gates open representing no big ship through. To ensure that only one ship enters the strait from one side, a resource, where only one ship is allowed to pass, is employed in conjunction with the gate. ese resources are named Marmara and Aegean corresponding to m1 and a1 and represent the starting location before each direction (the Marmara and Aegean Sea entrance). e starting location is seized by each ship entity before passing through and then freed immediately a er it passes. FIFO rule is applied at the entrances. e network model is depicted in Fig. 6. Although, the decision logic of tra c ow at both directions is the same, arrival time and time in the system are different. erefore, two networks are used to model both tra c ows with the attributes employed to specify the Fig. 6 resources and the gates required. ATRIB (2) is used to maintain the resource number and le number associated with the rst location. If ATRIB (2) equals 1, the ship entity requires the resource m1. If ATRIB (2) equals 2, then resource a1 is required. e entities representing ships are created at two CREATE nodes, one for each direction. Time between ship arrivals is uniformly distributed. Following the creation of the entities, ATRIB is called ATRIB (2) =1 for direction 1 and ATRIB (2) =2 for direction 2. e ships waiting for entering will be put in les 3 or 4 and ATRIB (3) is used to indicate these numbers. us, entities are assigned and ATRIB (3) =3 for direction 1 and ATRIB (3) =4 for direction 2. Once the entity is allocated for the starting location, it proceeds to the next AWAIT node waiting for the gate is de ned by ATRIB (2) which is either a1 or m1. If an appropriate line is closed (which means a big ship goes through the line), then, the entity will wait in les 3 or 4 in accordance with the value given by ATRIB (3). e COLCT node is used to record the values of the waiting time of the ship at the entrance.

Simulation of the System
Two other segments of the model control the entrance of a big ship or dangerous cargo vessels in any direction and consist of a series of OPEN and CLOSE nodes. In this segment of the model, resources and gates are referred to the label given them in the RESOURCE and GATE blocks. GATE 1 refers to the line from the Black Sea to the Marmara Sea and GATE 2 is vice versa. If a gate is open, the ship can proceed, otherwise, they have to wait until the gate is open. e system closes the gate according to uniform distribution between 190 and 230 minutes and duration of 200-245 minutes which corresponds to approximately 1325 hours per year which means that one line is closed about 1325 hours. e last segment of the model is simulating dangerous passages and bad weather conditions to close two lines.
To simulate the system, the following assumptions are made: • vessels do not overtake each other; • vessels enter the strait one at a time from each entrance; • all vessels are trans-passing (not stopping for loading and unloading within the strait); • local crossing tra c does not interfere with transit tra c. Ship arrivals were simulated with 5 di erent uniform distributions and the intervals of those are given in Table 2. a1 direction was assumed to be closed with the uniform distribution between 190 and 230 minutes which corresponds to closure once every 1300 hours. m1 direction was closed once every 1350 hours and close time varied between 200-245 minutes. e closure of two directions was modelled also with uniform distribution between 2300-2500 minutes which corresponds to about 219 hours of closure a year.
The model was running for 43200 minutes (1 month) for six di erent arrival times. e results of six scenarios are given in Table 2.

Results
e third scenario simulates the existing situation modelled by uniform distribution between UNFRM (14, 16) for Marmara entrance and UNFRM (14, 17) for the Aegean Sea entrance. e average arrival times for Marmara and Aegean entrances are 15 and 15.5 minutes and the average waiting times make 24 and 18 minutes respectively. When arrival time reduces to the average, 10 minutes waiting time and the number of waiting ships increase rapidly (see Figs 7, 8 and Table 3). Fig. 7 shows that an increase of 25% in ship number grows 43 times taking into account the number of waiting ships at Marmara entrance (from 1.663 to 73.73) and 76 times at the Aegean Sea entrance (from 1.094 to 84.81). Fig. 8 shows that only 25% increment in ship arrivals caused ship waiting time to increase from 24.267 to 737.73 minutes for Aegean entrance. Similar results were also found for Marmara entrance (from 18.386 to 883.73 minutes).

Conclusions
1. Everyone recognizes that given the nature of Canakkale Strait and the existing grave situation created dense tra c congestion. In this case, the strait cannot bear additional oil shipments without putting into danger the safety of Canakkale, the lives of its population and its unique historical and precarious environment. 2. e third simulation is the current level of maritime tra c. e obtained results of this simulation show that waiting time at Marmara entrance is about 24 minutes. 3. As seen from Table 3, if maritime tra c at Marmara entrance increases in 25%, the average waiting time changes from 24 to 737. 4. Table 3 also indicates that an increase of 25% in tra c at Aegean entrance would result an increase in waiting time from 18 to 883. 5. As simulations showed, there would be congestion at the entrances of the strait in case of the increased number of ships which would put pressure on either vessels to increase their speed or the owners to carry goods using large ships. 6. e current systems and approximately 90 degrees turning make ship manoeuvring di cult at Nara Turning point. Large ships and speed increase would regularly make manoeuvring di cult.
us, Nara point can be considered as a risky area. erefore, necessary actions like keeping emergency sta and equipment should be taken.