Uncertainty

Website by Jean E Us

All chemical analysis is ultimately an approximation to the ‘right’ value, this applies as much to the laboratory as it does to on-site testing. Each laboratory will probably have differing techniques of analysis and extraction. Consequently they will generate different results for similar samples, despite the differing techniques being certified.

In the vast majority of cases the greatest cause of discrepancy of results between samples is down to actually taking the sample in the field. The Mcerts committee is currently evaluating guidance to allow the sampling procedure to become Mcerts accredited, emphasising just how vital sampling is for the analytical process. QROS has in place a standard operating procedure for sampling which will adopt the relevant Mcerts recommendations when they become published. For on site analysis, many of the procedures recommended that will try to preserve the integrity of the sample are not needed because the sample is extracted or analysed within a few minutes of taking the sample. Extraction processes do have a substantial effect on the final results with contaminants being easier to extract in differing solutes and at different temperature and pressures. The QROS extraction techniques are designed to be as efficient as possible whilst using acceptably safe and environmentally benign reagents.

However many sites have an extremely heterogeneous profile and it will often be the case that two samples taken from the same area of ground will have widely disparate contamination levels. Our heavy metals analysis report can help highlight if a sample is likely to be hetero or homogeneous in nature. Hydrocarbons, due to their nature tend to cause a more homogeneous dispersion in soil, but hydrocarbon concentration at the plume boundary can have significant variations even over a few metres.

There are studies being conducted to give more insight into the acceptability of on-site measurement by estimation and optimisation of uncertainty one of which can be found opposite. In these studies, the best way found to reduce uncertainty was to have more samples analysed from a given sampling location. The use of on site analysis can help to achieve this aim without increasing the analytical budget to unacceptable amounts.

Estimating and Optimising Analytical and Sampling Uncertainty in Environmental Investigations: Application and Evaluation

10.1111/j.1751-908X.2007.00854.x

Authors: Boon, Katy A.; Taylor, Paul D.; Ramsey, Michael H.

Source: Geostandards and Geoanalytical Research, Volume 31, Number 3, September 2007 , pp. 237-249(13)

Publisher: Blackwell Publishing.

Evaluating uncertainty in routine analysis

doi:10.1016/S0165-9936(99)00151-X

Authors: Alicia Maroto, Ricard Boqué, Jordi Riu and F. Xavier Rius

Source: TrAC Trends in Analytical Chemistry,
Volume 18, Issue 9-10, September 1999 , pp. 577-584

Publisher: Reed Elsevier.

Website by Jean E Us

Environment and Energy Award