Multivariate analysis of water quality of Sacramento-San Joaquin Delta

Understanding ambient water quality characteristics is critical for identifying the key issues related to water quality and for deriving potential solutions for protecting public and environmental health. Water quality monitoring is used to characterize physical and chemical characteristics, and data driven approaches are often employed for evaluating water quality. However, observation-based data driven approaches pose challenges in decision making because of temporal and spatial variability of water quality, which lead to challenges in decision-making. Combining the monitoring approach with multivariate statistical analysis may provide an improved understanding of water quality of ambient water bodies. In this study, we used rapid and portable sensors to determine water quality of samples collected from four different locations in Sacramento-San Joaquin Delta. Then, we used a multivariate statistical technique to evaluate spatial water quality characteristics. Four sampling sites with seven water quality parameters produced clusters reflecting different pollution levels. Factor analysis extracted two varifactors explaining a majority of the total variance and representing the dominant water quality parameters. In this study, we used handheld sensors, which are relatively in-expensive, and easy to use with accuracy varying between 5 and 10%, which is comparable to conventional sensors. These traditional sensors are relatively expensive and challenging to use compare to portable the rapid test sensors. The results showed that the average pH of delta water was within permissible limit for drinking water (6.5–8.5 pH). However, the average turbidity value was often greater than the permissible limit for drinking water, which is set to less than 1 NTU. The range of turbidity varied depending on the sampling locations (0.1–7.5). The permissible conductivity value for drinking water is set to 1000 μS/cm, and the delta water conductivity was found to vary between 305 and 6,600 μS/cm. The study presented here conducted a water quality sampling to evaluate the water quality, and demonstrates the capacity and application of multivariable statistical analysis for water quality assessment and identifying pollution factors in ambient waterbodies.