1.Introduction
From the economic and intranational trade perspective, the Strait of Malacca is one of the most important and busiest maritime waterways in the world. The strait is a vital shipping channel linking European Union the with the major Asian countries of South China Sea and the Pacific Ocean such as China, Japan, South Korea and ASEAN (Wan et al., 2013). Through the strait, one-quarter of the world's traded goods including oil, the large assortment of valuable manufactured goods in container boxes by tankers and container vessels every year in vice versa direction to India, Middle East and Europe. In addition, for the coastal states (or countries) along the strait, it is a very important entrance into the internal waters (Chia, 1998).
There are several studies to investigate the strategic importance of the Strait of Malacca (Evers and Gerke, 2006; Xiaobo Qu et al., 2011). If the strait was blocked by any reason, vessels especially deep draft tankers and container carriers might use the other alternative routes such as Sunda and Lombok straits. In this case, there will be additional 1.5 to 3.5 days for each vessel (Xiaobo et al., 2011), which have adverse effects on economic development of Korea because Korea has been highly depended on the imported raw materials and international trade by ships.
There are also many other studies of enhancing maritime safety and security in the Strait of Malacca (Roach, 2005; Mohd Hazmi, 2012a; Sam et al., 2007). However, the most of these studies did not have detailed analysis on a marine traffic distribution in the strait, which has great impacts on the marine traffic safety.
Therefore, the purpose of the study is to fill this research gap to analyze marine safety of the Strait of Malacca more concretely. To do this, this study used the 30 days AIS data collected from 16 May to 15 June 2012 and found out parameters of traffic in the waterways such as routes, frequency, types, length, speed and draft of vessels using TOAIS program (Thanh et al., 2013). Based on these information, the lateral traffic distributions in gates were determined by using the EasyFit Professional program (MathWave, 2014). Thereafter these analysis results were used as input data for risk assessment of collision in the Strait of Malacca by using the IWRAP model(IWRAP Wiki, 2014).
2.Marine Traffic Analysis
2.1.Introduction about the Strait of Malacca
The strait of Malacca is the shortest shipping channel between the Indian Ocean and the Pacific Ocean, in term of navigation. And in term of economy, it is the traffic route which carries 80 % of the oil transported to the Northeast Asia as well as one third of the world's traded goods (Xiaobo et al., 2011).
This paper focus on the waterway of the Strait of Malacca within the STRAITREP (from sector 1 to sector 6) which covers the area between longitudes 100°40'E and 103°23.5'E and includes the routing system (Traffic Separation Scheme) in the Strait of Malacca as shown in figure 1. Comparing with the Western part between longitudes 95°26'E and 100°40'E, this area is narrower, shallower and has many ports locate along the two shores so that the marine traffic safety is also more important.
2.2.Marine traffic in the Strait of Malacca
Marine traffic in the Strait of Malacca consists of two main types of traffic: transit traffic and cross-strait traffic. The first one is a traffic of vessels pass through the strait with or without call one or some ports along the strait. The second one is a traffic of vessels operating on the routes between Malaysian ports of Port Dickson, Melaka, Muar and Kukup with various Indonesian ports on the opposite shore, including Pelabuhan Belawan, Tanjung Balai, Dumai and Bengkalis in Sumatra (Mohd Hazmi, 2012b).
2.2.1.Analysis of Transit traffic using North-South route
Because the Strait of Malacca is long and wide strait so that most of vessels passing through the strait are international voyage vessels and equipped with AIS. The figure 2 shows traffic density in the research area.
According to this figure, East bound and West bound vessels sail on the two routes which are almost separated. The East bound vessels which have deep draft sail on special route when passing sector 3 of the STRAITREP.
In the area, there are three crossing areas between vessels sailing on the main route and vessels arriving or departing from ports that located along the strait, including off Port Klang, off Port Dickson and off Malacca Dumai.
To analyse details of transit vessels’ traffic, nine gates were set as figure 3. By using the TOAIS program, frequency, types, speed, draft and length of vessels passing through each gates were investigated.
As the results of the TOAIS analysis, the number of vessels passing gate No.1 is 6,484 (about 216 vessels/day and 77,800 vessels/year). Type and size of passing vessels are displayed in figure 4 and 5. The two figures show that general cargo vessels such as container vessels, general cargo vessels, bulk carrier and tanker vessels (include oil tankers, chemical tankers and GAS carriers) accounted for about 95 % of all passages and most of vessels have length over 175 meters (average length of vessels is about 210 meters).
Table 1 shows the number of vessels passed gates from No. 6 to 9 arriving and departing ports in Dumai, Melaka, Port Dickson and Klang. Table 2 and 3 show the type and size of vessels passing gates No. 6 to 9.
Table 1 shows that there are about 35,000 vessels called at ports in Dumai (in Indonesia), Melaka, Port Dickson and Port Klang (in Malaysia) among 77,800 vessels passing the Strait of Malacca. The remaining about 42,800 vessels passed through the strait without any call at the ports along the strait.5
Through tables 1, 2 and 3, we could recognize that the route to Klang ports has highest number of passages and most of vessels are container and general cargo vessels with average length about 180 meters. In the routes to Dumai, Melaka and Port Dickson, most of vessels are tankers which have average length about 160m, 150m and 155m respectively.
Figures 6 and 7 show type and size of vessels passing gate No. 5. Same as the gate No. 1, general cargo vessels and tankers account for about 98 % of number of passing vessels and average length of vessels is about 179 meters.
2.2.2.Analysis of Cross-strait traffic using North-South route
Figure 8 shows four main routes of cross-strait traffic in the Strait of Malacca.
Cross-strait traffic includes barter trade vessels, fishing boats and passenger ferries as illustrated at figures 9 to 1011. A tightly-knit network of trade relations, both formal and informal, spans the waterway. Barter trade activities in the Strait refer to the trade activities between the people who are living on opposite shores of the Strait of Malacca. Most of the cross-strait vessels call at the Malaysian ports of Port Dickson, Malacca, Muar and Kukup, all located at the south-western end of the Malay Peninsula. These ports have connections with various Indonesian ports on the opposite shore, including Pelabuhan Belawan, Tanjung Balai, Dumai and Bengkalis in Sumatra (Mohd Hazmi, 2012b).
3.Marine Traffic Distribution
3.1.Traffic Distribution Analysis based on AIS data
According to chapter V, regulation 19 of the SOLAS convention, “All ships of 300 gross tonnage and upwards engaged on international voyages and cargo ships of 500 gross tonnage and upwards not engaged on international voyages and passenger ships irrespective of size shall be fitted with an automatic identification system” (SOLAS, 2010).
Therefore, in waterways where the majority of the participating vessels are equipped with AIS equipment, as the Strait of Malacca, the use of AIS data for traffic survey brings enormous benefits.
The traffic distribution is determined through the following two steps. The first step will determine number of vessels passing each gate. In the AIS data, speed and position of vessels are provided at a interval of time (from 2 seconds to 3 minutes) so that position and time of vessels when passing a gate can be calculated exactly. Figure 12 shows a histogram of number of passing vessels at a gate (No. 4-East bound, as in the figure 3) for one day.
In addition, the AIS data also provides another information of vessel such as type, size and draft so that the number of vessels passing a gate could be classified by type, size and draft quickly In the second step, statistic data on number of vessels passing a gate are generalized by distribution function by using EasyFit Professional program. With this useful tool, parameters of the distribution function which is fit with the histogram of number of passing vessels at a gate will be found. And Figure 12 shows an example of the fitted distribution of the EasyFit Professional program.
3.2.Traffic Distribution Analysis based on statistic
The Automatic Identification System (AIS) was developed in the 1990s, however until 2000, vessels started to equip AIS transceivers onboard as a new requirement of IMO. Before this time, the works on traffic survey mostly were done by manual observation so that it is difficult to find the traffic distribution. Thus, at that time, the traffic distribution mainly was calculated based on the formula which was created by using statistical analysis.13
For example, traffic survey studies in the Japanese waterways showed that traffic distribution in a fairway(w) can be determined as follow (Park et al., 2010):
- Traffic is distributed under the Normal distribution
- In the case of fairway without center buoy, center of distribution is fixed at one-tenth of fairway width from the right side of the center of fairway
- In the case of fairway with TSS(Traffic Separation Scheme), center of distribution is fixed at one half of fairway width from the right side of the center of fairway
- Standard deviation is calculated by formula (1)
where:
σ : Standard deviation (m)
ω : Width of fairway (m)
Q: Traffic volume (vessel/hour)
Until now, this old method is still commonly used in the Marine traffic engineering field.14
3.3.A comparison on the traffic distribution between the old and the new methods in the Strait of Malacca
Along the strait, there are some areas where marine traffic is very complex because of a crossing traffic. So that these areas were not selected for comparison on the traffic distribution between the two methods. Instead of this, routes that have straight traffic such as routes at gates 3, 4 and 5 (as shown in figure 3), were selected.
Figures 15 shows normal distribution curves of traffic at gates 3 to 5 that were determined by analysing the AIS data. Figure 15 shows that the distributions of traffic at gates from 3 to 5 are very close to the normal distribution. However it is not the best fitness distribution.
Table 4 shows the mean and standard deviation of traffic distribution in each gate which are calculated by the two methods.
Table 4 shows that parameters of the normal distribution at gates 3 to 5 which are determined by analysing the AIS data have some features as follows:
-
the center of distribution is not fixed at one tenth of fairway width from the right side of the center of fairway;
-
and standard deviation of traffic distribution differs from the formula (1) (about 20 percent).
The figure 16 shows the normal and Weibull distribution curve over the histogram of traffic at gate No. 5-West bound. It shows that the Weibull distribution curve is more close to the histogram of traffic than the normal distribution curve.
4.Quantitative Assessment Traffic Risk by using the IWRAP model
4.1.IWRAP model and the IWRAP Mk2 program
The IWRAP Mk2 program was developed by the Canadian Coast Guard, Technical University of Denmark and the Marine Simulator Center Warnemende. It was based on the IWRAP model (IALA, 2012). In the IWRAP model, the frequency of accidents in a waterway can be calculated by a basic formula (2) as follow:
in here:
λ = frequency of collision or grounding accidents
NG = geometric number of collision or grounding candidates
P = causation factor
For calculating the NG value, the IWRAP model separates collisions and grounding accident by types as shown in Table 5. The NG value of a sub-type can be calculated based on specific scenarios for each one.
For calculating the causation factor, a causation possibility model was used. The IWRAP Mk2 program proposed to apply a Bayesian Networks model. It uses the default causation factor value as Table 6 (IWRAP Wiki, 2014).
4.2.Modelling traffic in the Strait by the IWRAP Mk2 program
Based on the analysis results on the AIS data, traffic in the Strait of Malacca is modeled in the IWRAP Mk2 program as shown in the figure 17. All the data related to traffic distribution and number of passing vessels, average speed, average draft of different types and size of vessels are calculated based on the AIS data which were collected in the Strait of Malacca.
4.3.Assessment result
Assessment results by the IWRAP Mk2 program are shown in the figure 18 and table 7.
As shown in the Table 7, it showed that the highest level of risk were overtaking and crossing situations in the Malacca Strait.
And Figure 18 shows that, collision risk in the area A (fairway connect the Strait of Malacca with Port Kelang area) is the highest in the Malacca waterway area. It caused by the high possibility of head on condition in this fairway, as shown in figure 19, together with the highest number of arrive/depart vessels compare with other ports that located along the Strait of Malacca. The second area of high risk traffic is the B area in figure 18, the Strait of Malacca from P.Udan light to Port Dickson. The collision risk mostly causes by over take condition. The third areas are area C and D in figure 18, in these areas the common traffic is crossing traffic.
5.Conclusion
As one of the most important fairway on the world, the traffic safety in the Strait of Malacca plays an very important role in the economic development of countries in the East Asia such as Japan, Korea and China.
So it is important to assess and improve the marine traffic in this area. This paper surveyed and analysed the AIS data for 1 month in the Strait of Malacca. By analysing the AIS data were collected in the Strait of Malacca, it shows as follow:
-
traffic in the Strait of Malacca consists of the two main routes: East-West route and North-South route. Most of vessels in the East-West route are tanker, container vessels and general cargo vessels which is about 210 meter in length. These vessels can pass through or call at one or some ports along the fairway. The East-West route is the route of ferries, small ships or fishing boats operates between ports along the two shores of the strait.
-
Generally, traffic distribution in the Strait of Malacca follows the normal distribution. However, the parameters of distribution is differ from the previous study in the Japan waterway. In addition, in some gates the normal distribution is not the best fitness distribution.
-
Quantitative assessment risk of collision in the Strait of Malacca by using the IWRAP model shows that the fairway which connects the Strait of Malacca and Port Kelang has the highest risk of collision. Following are the areas in the Strait of Malacca from P.Udan light to Port Dickson.
In the future, it needs to deduce the best expression method of marine traffic distribution including deviation and mean for passing vessels. And it needs to study the best countermeasure for improving the marine traffic in the Strait of Malacca.