Dynamic risk assessment of natural environment based on Dynamic Bayesian Network for key nodes of the arctic Northwest Passage

Recent global warming has made it possible to exploit and utilize resources in the Arctic Northwest Passage. However, the harsh natural environment in this sea area poses a major threat to safety during navigation. Although this passage is extensive, the natural environmental state of few key nodes affect the navigability of the entire passage. In this paper, we describe dynamic assessment of natural environmental risks of key nodes in the Arctic Northwest Passage using Dynamic Bayesian Network (DBN). Specifically, index selection and data processing, determination of key navigation nodes, calculation of evidence-based reasoning and verification of DBN-model are discussed. Results show that the DBN-model effectively handles uncertainty of information, and generates highly accurate inference results. In addition, it integrates historical information in the reasoning process, enables accumulation of information, reduces the influence of data errors on the final result, and makes the result closer to the real value. Overall, this model provides an important reference for judging the comprehensive risks of natural environment at key nodes.