Evolution trends and influencing factors analysis for the severity and pollution of maritime accidents in Arctic waters from multi-source data
利用2004-2023年北极水域海事事故与水文气象数据，分析事故严重性和污染的演化趋势，并用双变量Probit模型识别影响因素，发现船旗国特征、机械损坏、强风等对严重性和污染有不同影响。
As Arctic maritime activity has increased, Arctic shipping risk management has attracted extensive attention. Given the limited availability of rescue equipment in Arctic waters, maritime accidents in ice-covered regions tend to result in more severe consequences and a greater environmental impact. This paper proposes a data-driven framework for analyzing evolution trends and identifying the influencing factors of maritime accidents in Arctic waters, in terms of both their severity (casualties and property loss) and pollution (environmental impact). First, a data preparation approach is proposed to integrate maritime accident data and associated hydrometeorological data from 2004 to 2023 for Arctic waters covered under the Arctic Search and Rescue Agreement. Second, the evolutionary trends of maritime accidents in Arctic waters are analyzed in terms of their severity and environmental pollution. Third, a bivariate probit model is proposed to explore the factors affecting the severity and environmental pollution of maritime accidents in Arctic waters. Finally, a marginal effect analysis is conducted to quantify the impacts of these factors on the severity and environmental pollution. The results indicate that flag state characteristics significantly influence both the severity and the pollution of accidents. Additionally, factors such as machinery damage, wrecked, allision, hull damage, fire/explosion, strong winds, and sea ice contribute positively to the severity of accidents, while these factors negatively influence pollution accidents to varying degrees.