Evaluation model and management strategy for reducing pollution caused by ship collision in coastal waters

Abstract

This study addresses the Spatio-temporal associations of internal and external factors (e.g., ship type, flag registry, port inspections and season, etc.) with ship collisions in coastal waters. To prevent collision and subsequent pollution, effective identification of various potential risk factors related to collisions is important. Based on 10-year collision reports, a Bayesian Spatio-temporal (BS) model is developed to assess collision risk by taking into account space and time mutual influences. In this research, the research area including parts of North China, Korean Peninsula, and Japan, is categorized into three different risk levels of potential collision. An association rule mining algorithm is then developed to analyze the spatial-temporal mutual relationships of collision elements in neighbor regions. Corresponding association rules among different collision risks are revealed for different regions. The results show that small ship size, general cargo type, and spring or summer season have a strong association with the collisions occurred in high-risk regions, while large ship size in accordance with summer or winter season is remarkable in moderate-risk regions. It is also found that large ship size and tanker ship type are significant factors in low-risk areas. Finally, management strategies are proposed to help preventing ship collisions and reducing potential pollutions.