Formation Mechanism And Forecasting Of High Arsenic Groundwater Using Multivariate Statistical Method

Chronic As poisoning has occurred in recent years. Accordingly, studies of high-arsenic groundwater become a hot topic. However, the research mainly focuses on shallow groundwater. Little research has been done on characteristics of spatial distribution of arsenic in deep groundwater in arsenic-affected areas. Multivariate statistical method was applied in this dissertation to investigate the quantitatively relationships among physiochemical parameters of high-arsenic groundwater, to extract the main influencing factors, to reveal formation mechanisms of high As groundwater, and to build prediction model of its spatial distribution.Ninety samples of deep groundwater were collected in the front of Langshan Mountains. Twenty-two representative physicochemical parameters were obtained for each sample, which were evaluated with the methods of principal component analysis and hierarchical cluster analysis. Results show that As were highly correlated with ORP, U, NO3, SO4, Fe, NH4-N and HCO3. Moreover, the relationships between As and ORP, U, NO3, SO4 were negative, indicating that groundwater arsenic should be related to reducing environment.Three main factors controlling groundwater arsenic concentration was extracted by principal component analysis, including geological, reducing and oxic factors. Their roles are different in different areas. Factor scores obtained in principal component analysis were used as variables for statistical grouping in hierarchical cluster analysis. It shows that all samples were classified into three statistically significant clusters, which corresponded to the alluvial fans, the distal zone and the flat plain. In the alluvial fans, groundwater was controlled by oxic factor, where release of As into the groundwater was suppressed, and low As concentrations were observed. In the distal zone, groundwater was under conditions of suboxic, which is dominated by reducing and geological factors, where the levels of groundwater As were elevated. In the flat plain, groundwater was characterized by reducing conditions and high As concentrations, which is dominated by the reducing factor.Statistical analysis of water chemical data in three clusters shows that the formation mechanism of high-arsenic groundwater includes reductive dissolution of Fe/Mn(oxyhydr)oxides and sulfate reduction, which lead to As release under the impact of reducing factors. The processes would be mediated by organic matter and microorganism, with the generation of NH4-N, Fe(II) and H2 S. Based on this mechanism, ORP, U, and NO3, and NH4-N and Fe(II) were selected in the principal composition regression analysis as the representative oxic and reducing parameters in regression model to forecast As concentration logically. The built model was used to forecast groundwater As concentration in the inclined Piedmont plain of the Hetao basin and the Songnen plain. The predicted data were generally consistent with the monitoring data, which shows that this model is practical in simple geological unit with less anthropogenic influence.