Multivariate statistical analysis for the assessment of groundwater quality under different hydrogeological regimes |
| |
| Authors: | N Kazakis C Mattas A Pavlou O Patrikaki K Voudouris |
| |
| Institution: | 1.Laboratory of Engineering Geology and Hydrogeology, School of Geology,Aristotle University of Thessaloniki,Thessaloniki,Greece;2.Decentralized Administration of Macedonia-Thrace, Water Directorate,Thessaloniki,Greece |
| |
| Abstract: | Multivariate statistical analysis has been widely used for hydrogeochemical characterization of groundwater quality. In this study, hydrochemical data from three hydrological basins were used and two methods (factor and cluster analyses) were applied. The first area is the coastal area of Eastern Thermaikos Gulf where groundwater is influenced by seawater intrusion and geothermal fluids. The other two areas are the inland basins of Gallikos and Perdikas in which agricultural and industrial activities constitute the main anthropogenic pollution sources of groundwater. Initially, the aforementioned methods were applied for each area separately and resulted in a different number of significant factors and clusters, while the natural and anthropogenic influences were spatially determined in each area. Additionally, factor and cluster analyses were applied coupling data from all areas. Therefore, five clusters and three major factors were determined distinguishing the hydrochemical processes and impacts from anthropogenic activities in more detail. It is worth mentioning that the application of cluster analysis in the coupled groundwater samples of all studied areas resulted beneficially in the most hydrochemically complex area. Salinization dominates in the coastal area, while in Gallikos and Perdikas basins high concentrations of NO3 occur mainly due to agricultural activities and small livestock units. The numerous hydrochemical samples are identified as the main issue for the higher discretization and reliability of the second approach. Nevertheless, this study is associated with a number of limitations of multivariable statistical analysis regarding extreme concentrations of Cl and Na. This issue stimulates further research in overcoming and understanding these drawbacks. |
| |
| Keywords: | |
| 本文献已被 SpringerLink 等数据库收录！ |
|
