Multivariate statistical analyses were performed on data from 156 mineral-spring sites in nine Western States to analyze relationships among the various parameters measured in the spring waters. Correlation analysis and R-mode factor analysis indicate that three major factors affect water composition in the spring systems studied: (1) duration of water circulation, (2) depth of water circulation, and (3) partial pressure of carbon dioxide. An examination of factor scores indicates that several types of hydrogeologic systems were sampled. Most of the samples are (1) older water from deeper circulating systems having relatively high salinity, high temperature, and low Eh or (2) younger water from shallower circulating systems having relatively low salinity, low temperature, and high Eh. The rest of the samples are from more complex systems. Any of the systems can have a relatively high or low content of dissolved carbonate species, resulting in a low or high pH, respectively. Uranium concentrations are commonly higher in waters of relatively low temperature and high Eh, and radium concentrations are commonly higher in waters having a relatively high carbonate content (low pH) and, secondarily, relatively high salinity.
Water samples were collected and (or) measurements were taken at 156 of the 171 mineral-spring sites visited. Various samples were analyzed for radium, uranium, radon, helium, and radium-228 as well as major ions and numerous trace elements. On-site measurements for physical properties including temperature, specific conductance, pH, Eh, and dissolved oxygen were made. All constituents and properties show a wide range of values. Radium concentrations range from less than 0.01 to 300 picocuries per liter; they average 1.48 picocuries per liter and have an anomaly threshold value of 171 picocuries per liter for the samples studied. Uranium concentrations range from less than 0.01 to 120 micrograms per liter and average 0.26 micrograms per liter; they have an anomaly threshold value of 48.1 micrograms per liter. Radon content ranges from less than 10 to 110,000 picocuries per liter, averages 549 picocuries per liter and has an anomaly threshold of 20,400 picocuries per liter. Helium content ranges from -1,300 to +13,000 parts per billion relative to atmospheric helium; it averages +725 parts per billion and has an anomaly threshold of 10,000 parts per billion. Radium-228 concentrations range from less than 2.0 to 33 picocuries per liter; no anomaly threshold was determined owing to the small number of samples. All of the anomaly thresholds may be somewhat high because the sampling was biased toward springs likely to be radioactive.
The statistical variance in radium and uranium concentrations unaccounted for by the identified factors testifies to the complexity of some hydrogeologic systems. Unidentified factors related to geologic setting and the presence of uranium-rich rocks in the systems also affect the observed concentrations of the radioactive elements in the water. The association of anomalous radioactivity in several springs with nearby known uranium occurrences indicates that other springs having anomalous radioactivity may also be associated with uranium occurrences as yet undiscovered.
Additional publication details
USGS Numbered Series
Radioactivity and geochemistry of selected mineral-spring waters in the Western United States; basic data and multivariate statistical analysis