Analysis of temporal variability in groundwater quality provides useful information on processes occurring within a groundwater system. In this study, a high resolution (26 sites sampled 20 times over 24 months) geochemical dataset was analysed to inform conceptual model development and to understand what processes were driving observed variations in groundwater quality.
Multivariate statistical analysis (MVA), comprising of hierarchical cluster analysis (HCA) and principal components analysis (PCA) was utilised to compare monthly water quality subsets to the median observed groundwater quality (selected as the baseline, representing average rainfall/recharge conditions during the study). Variability in identified groundwater-types at the site scale, and temporal changes in the composition of groundwater-types, were analysed to determine what processes were causing changes in water quality at the different scales. At the site scale, the results of the conventional approach using the entire available dataset, were compared to those derived from analysing each monthly subset individually. Analysis of individual monthly subsets provided greater sensitivity in that more subtle variations which were not identified by the conventional approach were detected.
The analysis identified seven groundwater-types within the studied system. Salinity, weathering products and evapotranspiration (ET) effects were determined to be the main factors affecting groundwater-type classification. The distribution of the identified groundwater-types could be correlated with landscape features. Temporal changes in groundwater quality were driven by factors including upward leakage, the proportion of fresh recharge, mixing with surface water-derived recharge, and the degree of ET. Variability in groundwater-type composition was limited and short-lived, indicating relative stability in groundwater composition during the study period.
A classification scheme was developed based on the results of the MVA analysis with the aim of allowing preliminary analysis of groundwater quality based on a reduced number of parameters. This may assist managers in identifying potentially important local processes affecting groundwater quality.