Integration of self-organizing maps and entropy-weighted quality index methods to study hydrochemical processes and quality assessment in Qilian coal base of China

A holistic understanding of groundwater hydrogeochemical characteristics and water quality is an essential task for managing and protecting groundwater resources. However, it is a challenge for the complex hydrogeological conditions of the Tibetan Plateau coal base, where many factors can cause regional hydrochemical changes. In this study, the groundwater hydrochemical processes in the Jiangcang basin of Qilian coal base are analyzed by the combination of self-organizing map (SOM) and K-means clustering. Traditional hydrogeochemical methods such as Durov diagram and Gibbs diagram are used to further demonstrate the feasibility of the SOM clustering method systematically. Further, the entropy-weighted theory is used to describe groundwater quality and assess the suitability of groundwater for drinking purposes. The results show that the groundwater samples (Cluster I and III) are mainly distributed around the mine area and coal transport channel, with high Na + and high SO 4 2− type. Their groundwater hydrochemical formation mechanisms are influenced by human factors, the main ion sources are related to coal production. The groundwater samples (Cluster II) are distributed in river valley talik and bedrock mountains far away from the mine area, mainly HCO 3 − type water. The rock weathering is the dominant mechanism controlling the groundwater chemistry of Cluster II. Among the 42 groundwater samples, 80.9% of the above-medium groundwater samples were suitable as drinking water. The results of this study demonstrate that the SOM can be used as a data-driven approach to understanding the hydrochemical processes in alpine regions, which can facilitate better water resource management and ecological environmental protection of Tibetan Plateau coal base.