Modeling and characterization of shallow aquifer water based on ion concentrations to salinity variations using multivariate statistical approach

Abstract

Abstract Water chemistry data (1965–2015) from the Chicot and Evangeline aquifers in Nueces and Kleberg counties of Texas were analyzed to assess the hydrogeochemical processes based on variations in ion concentrations affecting groundwater salinity levels in those coastal aquifers. A further aim of this study was to identify a proxy using major-ion concentrations to predict groundwater salinity. Correlation analysis suggests that the hydrogeochemical processes operating in these aquifers differ significantly. In principal component analyses, the first three principal components explain 91% and 94.8% of the total variabilities of the variation of groundwater chemistry in the Evangeline and Chicot aquifers, respectively. The rotated factors (factors 1, 2, and 3) using the varimax rotation method implied that depth control and salinity variations predominantly cause the highest variabilities in water chemistry in the Chicot and Evangeline aquifers in the study area. Results showed depth control on water quality parameters was more pronounced in the Evangeline aquifer compared to the Chicot aquifer. Stepwise elimination of the least significant predictors to identify the proxy for groundwater salinity revealed chloride (Cl) could be the most significant predictor to estimate groundwater salinity variations in both aquifers. However, regression models generated from 75% of the training datasets predicted total dissolved solid (TDS) variations with 78% and 43% accuracies in Chicot and Evangeline aquifers, indicating that Cl can be considered the proxy for the Chicot aquifer only but not suitable for the Evangeline aquifer in the study area.

Keywords

• Aquifer
• Groundwater chemistry
• Regression
• Salinity intrusion