Occurrence of natural and anthropogenic hexavalent chromium (Cr VI) in groundwater near a mapped plume, Hinkley, CA


Task 6

Evaluation of the occurrence of natural and anthropogenic chromium

A wide range of geologic, geochemical, and isotopic data will be collected from the different subareas within the study area. No single type of data is expected to provide the complete answers to questions concerning background Cr VI concentrations. However, the combination of these data is expected to provide understanding to address this issue. It is likely that background Cr VI concentrations may differ by subarea, or that anthropogenic Cr VI concentrations attributable to PG&E may be identified at concentrations below background in some areas.

Initial data analyses will be done using a pattern-recognition approach developed using a summative scale. Most people are familiar with a type of summative scale commonly used to assess public opinion known as a Likert scale (Likert, 1932). To analyze results from a Likert opinion survey, ordinal values are assigned to responses to questions within the survey (known as items) such as: strongly agree, agree, no opinion, disagree, strongly disagree (in a typical Likert survey scores could range from 1 through 5, respectively). The ordinal values are then summed to determine the overall opinion of the participant. Summative scales have application in the physical sciences as an initial data-analysis tool for problems where multiple processes may affect the measured data, and there is not a strong consensus or understanding of the importance of the individual processes. Their use in this study is intended to provide a transparent process that facilitates input from TWG members during data interpretation.

For the purposes of initial data analysis, binary-ordinate (nominal) values will be assigned to items developed on the basis of metrics measured during this study that compose the scale. The items that compose the scale, analogous to questions in a Likert opinion survey, will be relevant to the occurrence of natural (value = 0) or anthropogenic (value = 1) Cr VI. For example: 1) Is tritium absent (value = 0)? , Is tritium present (value = 1)? , or 2) Is chromium less abundant than average in local geologic material (value = 1)? , Is chromium more abundant than average in local geologic material (value = 0)? Summed values will be used to categorize (bin) data to assist with pattern-recognition on graphics developed as part of this study. These graphics will include maps for spatial analyses, and representations of geochemical processes, such fractionation or mixing trends on plots of δ53Cr as a function of Cr VI concentrations.

After the initial analysis, geologic, hydrologic, and geochemical knowledge coupled with traditional parametric and non-parametric statistical analyses will be used to refine results to: 1) identify items within the scale that contribute most to interpretable patterns; 2) identify items within the scale that covary and may be eliminated from the scale without loss of interpretative power (Principal Component Analysis may be especially suited for this); 3) refine items within the scale from data oriented responses to process oriented hydrologically based responses; 4) if appropriate, values other than binary-ordinate (nominal) values will be assigned to refined items within the scale; and 5) if appropriate, weights will be assigned to selected items within the refined scale. Graphics will be revised to improve pattern recognition and interpretation of the data as the scale is modified. It is likely that mixtures of natural and anthropogenic Cr VI may be present in some wells, and in some areas it may not be possible to identify with certainty the source of low concentrations of Cr VI.

Although the use of a summative scale is intended to reduce complex data to easily understandable visual graphics that facilitate input from TWG members during data interpretation, final interpretation of the data will be scientifically based and process oriented; and preparation of the final report remains the responsibility of the U.S. Geological Survey.

Cooperating Agency: Lahontan Regional Water Quality Control Board
Project Chief: John A. Izbicki
Phone: 619-225-6131
Email: jaizbick@usgs.gov

California Water Conditions

Real-Time California Streamflow Conditions