Measurement uncertainty in environmental analysis
- 1. Measurement Uncertainty in Environmental Analysis Todd R. Crawford, BA Center for Toxicology and Environmental Health, LLC
- 2. Standard Procedures ISO 1993 – ‘Guide to the Expression of Uncertainty in Measurement’ NIST 1994 – ‘Guidelines for Evaluating and Expressing the Uncertainty of NIST Measurement Results’
- 3. Uncertainty Statements Inform the reader of how sure the writer is that the answer lies within that range.
- 4. Uncertainty Statement? “…if by any chance our value is shown to be in error by more than 1 part in 1,000, we are prepared to eat the apparatus and drink the ammonia.” Attrib. Dr. C.H. Meyers
- 5. Error Error is a single value – an idealized concept, error cannot be known exactly. Error is the difference between the true value and the result of the measurement.
- 6. Uncertainty Uncertainty is the state of knowledge about the unknown error. Uncertainty is given as the range in which the error is to be expected.
- 7. Accuracy Accuracy is the closeness of a measurement to the true value. Accuracy is expressed as the percent recovery.
- 8. Precision Precision is the closeness of agreement between repeated measurements. Precision is expressed as the relative standard deviation (RSD).
- 9. Measurement Uncertainty in Environmental Analysis The true value cannot be known. The reported value is a ‘consensus’ value.
- 10. Reporting Uncertainty (ISO) Reporting is required when: The information is relevant to the application of the results, When the client requires it, When the Uncertainty affects compliance with a specification limit.
- 11. Evaluating Uncertainty (ISO) Evaluating is required for calibrations. Procedures for estimating uncertainty are needed and need to be applied.
- 12. Data Quality The Uncertainty of a result is a quantitative indication of its quality. Quantitative Uncertainty allows comparison of the results. Measurement Uncertainty needs to be taken into account when interpreting the data.
- 13. TPH in Soil Bioremediation of a former petroleum terminal to residential clean-up standards. Approximately 500,000 cy of soil treated. Approximately 15,000 samples analyzed on-site over five years. On-site data demonstrated compliance with the remedial criteria.
- 14. Interlaboratory Study Measure the Accuracy and Precision of Five Labs Analyzing TPH in Soil by GC-FID. Homogenous soil containing biodegraded diesel fuel with 10% moisture. Soil submitted in blind triplicates to each lab. Each lab is certified for this analysis.
- 15. Summary Results (n=15) Control Lab TPH = 921 ppm RSD = 2.4% Independent Labs TPH = 755 ppm RSD = 35%
- 16. Assessing Measurement Uncertainty Any set of numbers (n>3) has a mean and standard deviation. Complex statistical tests are beyond the expertise of most data users. Most data sets are too small to assess with any confidence.
- 17. Charting to Evaluate Uncertainty Rank the data in order of concentration. Normalize the data against the maximum value. Chart the value against its rank, where rank equals 1/n. Determine the slope and correlation coefficient.
- 21. Conclusion Forensic environmental investigations require evaluation of measurement uncertainty. Quantitative results are most useful when reported in the context of their measurement uncertainty.
- 22. Conclusion Measurement uncertainty is significantly improved with the familiarity of the laboratory to the analysis. On-site is better than Off-site! Measurement uncertainty can be evaluated (more effectively) by charting.
- 23. Acknowledgements John E. Ross, de maximis, Inc. (Charter International Oil Company) James S. Smith, Les Eng, Trillium, Inc. Willem Schre ü der, Principia Mathematica, Inc. Glenn Millner, Center for Toxicology and Environmental Health, LLC