Award Date


Degree Type


Degree Name

Doctor of Philosophy (PhD)



First Committee Member

David Kreamer

Second Committee Member

Michael Young

Third Committee Member

Matthew Lachniet

Fourth Committee Member

Ganqing Jiang

Fifth Committee Member

Ding Zhonghai

Number of Pages



Because of the complex geologic setting of the Basin and Range province, groundwater flow systems of the Intermountain basins of the southwestern United States are complex and remain poorly understood. Understanding these flow systems is important for water budgeting on a regional and local scale, and development of robust numerical groundwater models for sustainable water use and protection of water-dependent ecosystems. Although for decades hydrochemistry and isotopes have been used to characterize and trace subsurface water and surface water, effectively interpreting these data are still challenging, which can be attributed to existing subjective grouping of these data and the lack of methodological framework for analyzing and interpreting the data.

In this dissertation, new analytic approaches to analyze hydrochemical and stable isotopic data are described and used to trace the sources and movement of groundwater, and better quantify Virgin River interactions with groundwater in the lower Virgin River Basin in Nevada, Arizona, and Utah. In the new approach, data analytical techniques and data interpretations are combined in a sequential and mutually supportive way to test the hypotheses of potential interbasin groundwater flow, and Virgin River interaction with groundwater in the lower Virgin River Basin. To achieve these goals of the dissertation, three studies were conducted.

The objective of the first study was to optimize the multivariate statistical grouping of hydrochemical data of groundwater. The results indicated that repeated stable cluster solutions (robust hydrochemical facies) are obtained when cluster analysis is combined with Discriminant Function Analysis and agreement measures to group hydrochemical data. Also, this integrative approach allows for a quantification of the effect of analytical errors, outliers, and data transformation on the clustering of hydrochemical data. Using this approach, an optimal number of six robust hydrochemical facies were delineated for groundwater in the lower Virgin Valley. The results indicate that inappropriate data transformation can significantly impact the delineation of robust hydrochemical facies (Cramer's V < 0.8). In addition, the results indicate that analytical errors0.8) in clustering.

The objective of the second study was to precisely define the sources of groundwater and Virgin River interactions with groundwater in the lower Virgin River Basin through a new approach for analyzing hydrochemical data of groundwater and surface water that allows for the precise definition of sources and discharge end-members, and overlay of interpretations. The methodology developed in the first objective above was used to delineate 6 robust hydrochemical facies for all waters in the basin. Subsequently, hydrochemical end-members were identified using the Schoeller diagram and discriminant functions plot. Flowpaths were tested, based on geographic coherence and patterns of the characteristic facies. Additionally, the PHREEQC inverse modeling code was used, beyond the contemporary application for identifying processes of hydrochemical evolution, to diagnose the significance of potential recharge sources to the lower Virgin River Basin. Mineral phases and constraints selected were based on compositional plots, saturation indices calculations, and the available geologic information. PHREEQC inverse modeling indicates that at least 80% of the groundwater in the lower Virgin River Basin is derived from interbasin flow. In addition, based on the hydrochemical data, the floodplain aquifer interacts highly with the Virgin River, but both are unlikely to be hydraulically connected to the underlying Muddy Creek Formation aquifer.

Finally in the third study, linear regression, Spearman correlation tests, scatterplot, box-and-whiskers plot, and Wilcoxon Rank Sum test are unconventionally applied to glean information from δD and δ18O, and Na, K, SO4, and Cl data of the hydrochemical facies delineated in Chapter 3. The δD values of the recharge end-members from adjacent basins and within the lower Virgin River Basin are high and significantly different from the low δD values of the discharge end members. Box-and-whiskers plot of δ18O values, comparing to box-and-whiskers plot of δD values, indicate possible oxygen isotopic exchange between the discharge groundwater and the aquifer minerals. The isotopic exchange implies a long residence time of the groundwater which discharges in the basin. Correlation tests of δD and δ18O versus Cl, and box-and-whiskers plots of δD, δ18O, and solute data indicate the Virgin River and floodplain aquifer are more homogeneous and evaporated than samples of the source/recharge and discharge end-members. Regionally transported deep carbonate water is invoked as possible explanation of the low δD values of the discharge waters in the lower Virgin River Basin. Mixing calculations for stable isotopes indicate that at least 50% of the groundwater discharging in the lower Virgin River Basin derives from interbasin groundwater sources. Although this estimate is different from the estimates presented in Chapter 3, both results reinforce the significance of interbasin groundwater to the lower River Virgin Basin.


Groundwater flow; Groundwater flow systems; Hydrochemical facies; Hydrochemistry; Interbasin flow; Multivariate statistical; Stable isotopes; United States – Virgin River Basin; Water chemistry


Geochemistry | Geology | Hydrology