Spatial clustering based meteorological fields construction for regional vulnerability assessment

Chemical accidents have affected the social-environmental system. For the regional vulnerability assessment, which is the baseline work to assess the impact on the environment, a meteorological field is needed to determine how chemicals from multiple adjacent companies are propagated. In this study, we present the method of meteorological field based on the spatial cluster which is the main component of vulnerability assessment on regional chemical accident scenario. To integrate spatially dense chemical companies into a cluster, we adopt spatial clustering algorithms. Experiment result shows that DBSCAN-based approach reduces 80.5% total area of the meteorological field against brute-force algorithm, and shows good performance on the average of the overlap ratio, and utility ratio for clustering results.