2 MDL DEFINITION Defined at 40 CFR 136 Appendix D The minimum concentration of a substance that can be measured and reported with 99-percent confidence that the analyte concentration is greater than zero and is determined from the analysis of a sample in a given matrix containing the sample
4 MDL CALCULATIONSeven (7) replicates at ~ [5 times] the background noise: should be at or near the low calibration standardMultiply by (z) distribution factor3.14 time the S.D. = MDLThe MDL is more dependent upon precision than accuracyLabs. must demonstrate that they can see a signal at the MDL
5 MDL ISSUES TO CONSIDER Many regulatory standards near detection limit Lab credibility often associated with detection limitsMaze of tests and matrices make it difficult to find a universally applicable techniqueCalculating detection limits can be expensive
6 WHY YOU CAN’T SET PERMIT LEVELS @ THE MDL MDLs vary by day, lab., analyst & methodPrecision at low levels are usually poorGood for false-positives; false negatives may be reported 50% of the timeInherent uncertainties make enforcing permit MDL values impossible
7 MDLs to PQLs PQL: Practical Quantitation Limit PQL = Average (Median) MDL x 5Sources of MDL dataInterlaboratory low-level studiesApplying statistics to low-level cal. dataMethod data where availableMost data is now 5-12 years old!!!
8 PQLs to NJQLsThe Department recognizes that the current PQL data is based upon out-dated technologyOQA already requires certified labs. to collect MDL data by method and mediaWhat we need is current, representative data and a way to manage it
9 NJQLs & NJDEPOQA is contracting for the development of a database to collect MDL data from NJ certified labs. and then generate NJQLs (avg. MDL x 5)NJQLs to be generated by analyte, matrix, methodDWWPSHWDEP programs will have a choice in selecting an NJQL for their individual needs