The use of measurement uncertainty to assess the reliability of on-site field test kits for the investigation of contaminated land. Proceedings of ConSoil 2008

It can be seen from the case study that the generally poor quality of the lab data makes validation of the on-site kit measurements difficult, particularly for total PAH and TPH. Shortcomings in the sample preparation by the lab have probably also contributed to the high measurement uncertainties, however this highlights this issue of the quality of routine lab analysis that could affect any user of test kits aiming to validate results using lab measurements. It is possible that the PXRF could have been used at this site for on-site decision making e.g. where to sample or not to continue sampling, which samples should be sent for external analysis etc, as long as the measurement uncertainties estimated are quoted and taken into consideration. However for TPH and total PAH the quality of the lab data would make it unadvisable in this case to use the on-site test kits to make on-site decisions, as there isn't a strong relationship between the field and lab data. From this Case Study lessons have been learnt that will benefit further case studies in the project and other on-site tool users. These include: o the importance of a rigorously designed and implemented sample preparation in field and lab, o good communication with the lab prior to the experiment to ensure that all the data quality requirements and sample preparation are carried out as required, o knowing the definition of the analyte used in the on-site test kit and in the lab (e.g. TPH). If there is a discrepancy in the definition of an analyte between the field test kits and the lab that isn't known about, it could cause problems in the data comparison between the different techniques. However, with knowledge of it and if a relationship can be established between the field and the lab data then the field data could still be used effectively in the field to help manage sampling and remediation decisions. o the need for duplicates and matched CRMs in both the field and the lab for the estimation of measurement uncertainty. o the need for good data quality from the lab in order to validate field measurements. In general it can be seen that the estimation of measurement uncertainty is a simple and effective method for the validation of field test kits. Estimating and quoting the measurement uncertainty provides the user with a given confidence in the quality of the data and therefore gives confidence in the risk management decisions that are based upon the data.