The Strait of Istanbul is one of the most dangerous and busiest sea passages, according to its narrowness, sharp turns, currents, heavy traffic and many other important factors. Despite the latest precautions, marine incidents still occur, especially in narrow seaways. This study shows that a management tool that is specifically applied for reducing marine casualties and consequently enhancement of safety and ship management performance can be used to prevent accidents and casualties in maritime transportation. In this study, factors which affect marine casualties examined and determined with SWOT (strengths, weaknesses, opportunities and threats) analysis method and weighting of the factors determined by using the AHP (analytic hierarchy process) method. With this approach, strategic action plans were developed for minimizing shipping casualties at the Strait of Istanbul, taking into account the weighting factors and previously happened accidents.

Oil/chemical tanker ships carry a large volume of liquefied chemical cargoes in bulk condition. Chemical cargoes can be hazardous and most of them being flammable and/or toxic. It may pose acute dangers for human life and maritime environment. Risk assessment is of paramount importance to enhance operational safety and to protect environment for oil/chemical tanker ships. This paper aims at conducting an extensive risk analysis for oil/chemical tanker pollution due to leakage since the consequences can be fatal for human and marine environment. To accomplish this, fault tree analysis (FTA) is used. The failure occurrence probabilities of basic event (BE) of FT are calculated under fuzzy logic. The extended CREAM (Cognitive reliability and error analysis method) approach is used to determine human-oriented errors in FT since human error contributes the most to various types of maritime accidents. The outcomes of the paper can be helpful to minimize chemical leakage pollution risks and to improve operational safety in cargo loading/discharging process in oil/chemical tankers.