Deciphering spatial pattern of groundwater chemistry and nitrogen pollution in Poyang Lake Basin (eastern China) using self-organizing map and multivariate statistics

Abstract

Characterizing regional groundwater chemistry/quality is of great importance for sustainable water resource management, and is still a challenge due to spatial complexity caused by natural and anthropogenic influences at large scales. This study reported the spatial variability of groundwater chemistry and associated influencing factors in the Poyang Lake Basin (eastern China). A total of 68 groundwater samples collected from different aquifers were divided into hydrochemically distinct six clusters and prototypes representing each cluster were determined using the Self-Organizing Map-K-means (SOM-KM) clustering approach. Among them, Cluster 2, Cluster 4, Cluster 5 and Cluster 6 represented the contaminated groundwater with different type and degree of pollution. Multivariate analysis (e.g., principal components analysis (PCA)), together with classical hydrogeological methods, were applied to identify factors responsible for these clusters. Three principal components (PCs) were extracted in PCA with explaining more than 68% of the total variance, which reflected the major factors controlling the spatial characteristic of groundwater chemistry. Firstly, nitrate contamination significantly altered the groundwater chemistry, and the contamination level is heavier in the southern and western parts of the study area. Particularly, groundwater contamination could not only be attributed to local contamination sources (e.g., sewage) occurring at the sites, but also be transported there by groundwater flow. Secondly, groundwater chemistry was influenced by water-rock interaction, depending on the aquifer, lithology and matrix. Thirdly, the reducing environment was confined to localized areas and favors the formation of high NH4+ groundwater. Through this work, the spatial pattern of groundwater chemistry and associated anthropogenic contamination and hydrogeochemical processes are characterized. The outcomes of study are significant for groundwater management and pollution control in the Poyang Lake Basin and elsewhere.