海事事故严重性影响因素及影响程度识别

以上海、浙江、江苏海事局发布的海事事故调查报告为基础,通过建立有序概率模型(包括有序Logit模型和有序Probit模型),对海事事故严重性影响因素及影响程度进行识别.结果表明:有序概率模型适于研究该问题,且有序Logit模型优于有序Probit模型;季节、时段、天气、通航水域、船型、船长、船舶所有权等对较大及以上事故具有正向影响,能见度对较大及以上事故具有负向影响.其中,船长200 m影响最大(边际效应为0.378 3),夏季影响次之(0.282 2);能见度不良影响最小(-0.108 0),个体所有船舶影响次之(0.109 5).研究结果可为海事安全部门开展安全监管工作提供决策参考.

Identifying the Factors and Their Impact Levels on Severity of Maritime Traffic Accidents

Based on the investigation reports of maritime traffic accident, which were published by three maritime safety administrations, Shanghai, Zhejiang and Jiangsu, two kinds of ordered probability model (including the ordered Logit and the ordered Probit model) are established to identify the influencing factors and their impact levels on the severity of maritime traffic accidents (SMTA). The results show that the ordered probabilistic models are suitable for the study, but the ordered Logit model outperforms the ordered Probit model. In particular, the factors of season, time interval, weather, navigable waterways, vessel type, length and ownership have positive influences on larger and above accidents, while the influence of visibility is negative. Among them, the impact of the vessel length above 200 meters is the largest (marginal effect: 0.378 3), which followed by summer season (0.282 2). Meanwhile, the impact of poor visibility is smallest (- 0.108 0), which followed by individual vessels (0.109 5). Finally, the results can provide a decision basis for maritime safety administrations to carry out safety supervision work.