Abstract

To ensure the risk level associated with continuously increasing maritime traffic through particularly sensitive sea areas remains at acceptable level, a periodic risk assessment needs to be carried out by the relevant authorities. As a part of such assessment, allowing for proactive countermeasures to mitigate risk, the frequency of accidents is estimated along with the assessment of geographical locations where the accidents are most likely to happen. To this end scientific literature offers a number of approaches, however only a few solutions are recognized by the maritime authorities and applied world-wide. One of such approach is a evidence-based, semi-dynamic, network-based model called IWRAP Mk2. Despite its advantages, the tool lacks the verification procedure of the model development process that governs the reliability of the results. This ultimately may undermine the reliability of the obtained results. This shortcoming seems to be quite common in the field of maritime risk assessment, as revealed by the recent analysis of the risk assessment method and tools. Therefore, this article attempts to close this knowledge gap by providing a novel framework for ship-ship collision probability estimation and identification of the collision-prone locations, encompassing novel verification procedure suitable for network-based maritime risk models such as IWRAP Mk2 tool. As a results this new, wider modeling framework offers more reliable, evidence-based estimates of the probability of ship-ship collision and identifies more accurately the collision-prone locations in a given sea area. To demonstrate the usability of the framework a case study is performed, with the use of 10 months of ship traffic data recorded in the heavily trafficked and enclosed sea area of the Gulf of Finland during ice-free season with the special attention paid to the oil tankers. The updated framework delivers the annual probability of ship-ship collision, where at least one ship is an oil tanker, which is higher by 16% compared to the results obtained from regular IWRAP Mk2 software, that lacks verification procedure. Also the framework identifies the most collision-prone locations in the Gulf of Finland, which are located in the eastern part of the Gulf, explaining over 60% of the total collisions in the whole GoF, for ice-free seasons.

Citations

Authors (6)