Hydrogeochemical interpretation of South Korean groundwater monitoring data using Self-Organizing Maps

摘要

The National Groundwater Monitoring Network (NGMN) of South Korea provides data since 1995 to monitor the water level and quality of groundwater on a national scale. Major ions such as Ca, Mg, Na, K, HCO3, Cl, SO4 and NO3 have been monitored since 2008 to assess groundwater quality. Hydrochemical data of bedrock groundwater samples collected from 299 monitoring stations in 2009 were examined using the Self-Organizing Map (SOM) approach. Based on hydrochemical characteristics, bedrock groundwater is clustered into two groups and six subgroups. Group I containing 70.2% of groundwater samples (and monitoring stations) is characterized by lower TDS values and NO3 concentrations than Group II, indicating that Group I waters are less affected by contamination. Subgroup 1-1 (39.1%) represents Ca-HCO3-type groundwater with relatively low pH, TDS and concentrations of most ions compared with groundwater of Subgroups 1-2-1 (26.1%) and 1-2-2 (5.0%). Subgroup 1-2-2 represents a moderately alkaline, F-rich, Na-HCO3-type groundwater. Group II records either anthropogenic or natural processes. Subgroup II-1 (16.1%) contains groundwater with low values of TDS, HCO3 and pH, and moderately high NO3 concentrations due to nitrification, while groundwater of Subgroups II-2-1 and II-2-2 is characteristically high in Ca and Mg. Subgroup II-2-1 is also very high in SO4 and HCO3 but very low in NO3, while Subgroup II-2-2 is substantially enriched in CI and NO3. The hydrochemistry of groundwater of Subgroup II-2-1 likely results from dissolution of carbonates and gypsum in clastic sedimentary rocks and is affected by dissolution of pyrite and/or S-bearing fertilizers in crystalline rocks. The enrichment of NO3, CI, Ca and Mg in groundwater of Subgroup II-2-2 is the result of substantial contamination from agrochemicals and manure. Thus, about 20.5% (Subgroups II-1 and II-2-2) of bedrock groundwater in South Korea records anthropogenic contamination. This study shows that the SOM approach can be successfully used to classify and characterize the groundwater in terms of hydrochemistry and quality on a regional scale.