Commercial fishing, one of the most dangerous sectors globally and, particularly, in South Korea, faces fatal accident rates significantly exceeding other industries, and an urgent improvement in safety measures is necessary to reduce these numbers effectively. The main objective of this study is to analyse fishing vessel accidents in South Korea and present the direction of safety policy for fishing vessels based on the lessons learnt from previous accidents, and, ultimately, to contribute to reduce fishing vessel accidents. For this purpose, we have analysed the marine accident data from the last five years provided by the Korea Maritime Safety Tribunal (KMST), the statistical yearbook of oceans and fisheries provided by the Ministry of Oceans and Fisheries (MOF), and the statistical survey of marine fisheries provided by the Korean Statistical Information Service (KOSIS), using both quantitative and qualitative approaches. After thorough review and analysis of the accident data, we have proposed seven suggestions for the direction of future safety policy for fishing vessels in South Korea: adopt international safety benchmarks, focus on occupational and collision accidents, tailor safety measures to vessel size and age, enhance lookout practices to prevent collisions, implement comprehensive strategies to mitigate human errors, and address communication issues among Korean and international crews.

In this paper, we introduce a model designed to predict human error probability (HEP) in the context of fishing boat operations utilizing the cognitive reliability and error analysis method (CREAM). We conducted an analysis of potential accidents on fishing boats and calculated the cognitive failure probability (CFP) for each identified accident. The common performance conditions (CPCs) from the original CREAM were adapted to better reflect the conditions on fishing boats, with the adapted CPCs’ validity confirmed through expert consultations. To apply CREAM, data were gathered via a survey of fishermen, with the uncertainty in the collected data addressed through the application of fuzzy set theory (FST). We then established a Bayesian network (BN) model to elucidate the relationship between the fuzzy data and HEP, utilizing a weighted sum algorithm to determine conditional probabilities within the BN. Both basic and extended versions of CREAM were applied to analyze the most common accidents among fishermen, calculating the CFP for each type of accident. According to our analysis, the poorer the dynamic CPC, the higher the probability that a fall accident will occur inside the boat due to human error, necessitating a countermeasure. The paper proposes safety enhancements for small fishing boats and illustrates the increased precision of human reliability analysis (HRA) models in forecasting human error by incorporating quantitative methods. It calls for further data collection and refinement of the model for more accurate operational risk assessments.

Fisheries and Aquatic Sciences