Identifying critical energy-water paths and clusters within the urban agglomeration using machine learning algorithm

Energy and water shortages are two major problems in the process of urban development, and meeting the demands for energy and fresh water has become the key to global sustainable development. In this study, we developed a structure-based singular value decomposition (SSVD) method through incorporating techniques of multi-regional input-output (MRIO), structural path analysis (SPA), and singular value decomposition (SVD) within a general framework. The SSVD method is used to explore and track the system properties and flow paths of energy-water nexus network in the Pearl River Delta urban agglomeration (PUA) from 2012 to 2015. Our main findings are: (i) the largest final demand of inducing energy-related water (E-water) and water-related energy (W-energy) is the exports; (ii) Shenzhen mainly depends on other cities for E-water and W-energy, and Huizhou is the provider of E-water and W-energy; (iii) we identified over 10,000 energy-water clusters and found that Guangzhou's electricity and equipment manufacture drive the largest energy-water clusters, respectively. Our findings suggest that monitoring key paths and clusters of major energy-water consumption in the supply chains of urban agglomerations can provide new insights into energy and water policies.