Multi-scale near-long-range flow measurement and analysis of virtual water in China based on multi-regional input-output model and machine learning

Abstract

At present, China is facing a relatively serious water security problem, and virtual water trade is an important means to alleviate the shortage of water resources. It is of great practical significance for China to analysis of the multi-scale near-long-distance flow of virtual water between regions. In this paper, the multi-regional input-output model was used to measure the virtual water near-long-range transfer of 17 departments in eight regions in 2012 and 2017 in China. The measured results are further discussed in the context of K-means and structural decomposition analysis (SDA) methods. The results show that from 2012 to 2017, the overall trend of interregional virtual water flow and import and export volume of various regions and departments in China showed a downward trend, and there were significant differences in inflow, outflow, and distribution. In 17 sectors, the agricultural sector has a much higher virtual water flow than other sectors, particularly in the north-west and central regions. In addition, the food and tobacco industry and construction industries in most regions also have the characteristics of high outflows. This paper argues that China should Fully exploit the effect of industrial structure on the suppression of virtual water rise. At the same time, it is necessary to improve the virtual water trade pattern, especially the virtual water trade pattern of the agricultural sector, and then gradually optimize the distribution pattern of regional water resources in China.