Risk evaluation and safety optimization on narrow waters

Marine accidents occur frequently in narrow waterways such as straits and channels. Daily increasing maritime traffic increases the risks that threaten the safety of navigation in these regions. Occurring marine accidents bring not only ship and cargo loss but also loss of life and environmental disasters. Despite the additional measures taken with the developing technology, the targeted reduction in accidents could not be achieved. For this reason, it is necessary to reveal the fixed and variable risks in narrow waterways and to determine the current risk control options to ensure safety and sustainability in maritime trade. In this study, contact, collision, grounding and sinking accidents that occurred between 2004-2018 in the Istanbul Strait, Dover Strait and Singapore Strait, which are important junction points of maritime trade, were analysed qualitatively and quantitatively. In the analyses, Geographical Information Systems, Human Factors Analysis and Classification System and Bayesian network approaches were used, respectively. As a result of the study, accident networks have been developed that can reveal the accident probability of a ship, that will pass through the examined narrow waterways, depending on the variable conditions. If accident networks are used by shipmasters and vessel traffic operators in the relevant regions, it will be possible to increase awareness and reduce accidents. In this way, possible loss of life and environmental pollution cases will be reduced, contributing to the maritime traffic safety of the region. In addition, as a result of the study, measures were presented to prevent accidents in narrow waterways. The effectiveness of the measures presented for the Istanbul Strait on the accidents was analysed with the Evidential Reasoning approach. Thus, a quantitative feasibility study has been conducted on how the effects will be if the measures are implemented. Key Words : Narrow Waterways, Human Factor Analysis, Bayesian Network, Geographic Information Systems, HFACS, Evidential Reasoning