A Knowledge-Graph-Driven Framework for Intelligent River Pollution Traceability

Abstract

Frequent river pollution incidents threaten drinking water safety, ecological balance, and social stability. Rapid and precise source identification is critical for emergency response, yet conventional methods suffer from high computational costs, reliance on prior knowledge, and outcome uncertainties. This study proposes an intelligent traceability framework using knowledge graph technology. Topological relationships were established among multiple factors-mainstream-branch connections, source-sink dynamics, and upstream-downstream monitoring dependencies-along with spatial linkages between units, river reaches, and outfalls. Distinctive discharge patterns specific to different industrial sectors were also characterized. An intelligent river pollution traceability framework was developed based on these topological relationships and characteristic discharge patterns, implementing a three-stage hierarchical localization structure comprising: target river reach, risk spatial unit, and potential pollution source. Field implementation in the River Nanfei Watershed achieved significant enhancements in river pollution traceability, demonstrating improved efficiency (less than 20 seconds) and enhanced positional accuracy for the precise localization of risk units or pollutant discharge entities. This methodological advancement holds significant scientific value, as it enhances emergency response to river pollution and enables precise aquatic resource management─critical needs for addressing the escalating pollution challenges facing the world’s major rivers.