A Vine Copula-Based Modeling for Identification of Multivariate Water Pollution Risk in an Interconnected River System Network

The Interconnected River System Network (IRSN) has become a popular and useful measure to realize the long-term health and stability of water bodies. However, there are lots of uncertain consequences derived from natural and anthropogenic pressures on the IRSN, especially the water pollution risk. In our study, a Vine Copula-based model was developed to assess the water pollution risk in the IRSN. Taking the ponds around Nanyang station as research objects, we selected five proxy indicators from water quality indexes and eutrophication indexes, which included dissolved oxygen (DO), total nitrogen (TN), total phosphorus (TP), chlorophyll-a (Chla), and ammonia nitrogen (NH3-N). Models based on three classes of vine copulas (C-, D-, and R-vine) were utilized respectively to identify the water quality indicators before and after the operation of the connection project. Our results showed that TN, Chla, and NH3-N should be considered as key risk factors. Moreover, we compared the advantages and prediction accuracy of C-, D-, and R-vine to discuss their applications. The results reveal that the Vine Copula-based modeling could provide eutrophication management reference and technical assistance in IRSN projects.