Many U.S. Geological Survey (USGS) studies focus on interpretations of low concentrations of chemical constituents. The National Water Quality Laboratory (NWQL) of the USGS has developed a new low-concentration reporting convention that produces an important benefit to these studies by providing estimates of analytes detected at low concentrations. These values are reported rather than being censored at a higher minimum reporting level, as has been done historically in many NWQL methods. Providing estimates of low concentrations gives the user the flexibility to decide when it is appropriate to censor the reporting level and when it is appropriate to use estimated values below the reporting level for interpretations that are based on particular objectives of a study. Some studies require a high degree of certainty for an individual analysis, such as those that address regulatory issues. Others do not require the same high degree of certainty for an individual analysis but rather base interpretation on results from large numbers of samples, such as studies that assess the occurrence of a constituent in the environment.
A key advantage of the new reporting convention is that the NWQL will use two concentration markers, the long-term method detection level (LTMDL) and the laboratory reporting level (LRL), for data reporting to minimize the risk of critical measurement errors. The first measurement error, which is called a false positive, is relevant to analyte detections and occurs when an analyte is incorrectly reported as present in the sample when it is not. As detailed below, the false positive risk is minimized to no more than 1-percent probability at the LTMDL concentration. At concentrations less than the LTMDL, this risk can increase dramatically. The second measurement error, which is called a false negative, occurs when an analyte is reported as less than a specific concentration (the reporting level) when it is, in fact, present at a concentration equivalent to or greater than the reporting level. The false negative error risk, therefore, is relevant to the non-detection condition. The NWQL is setting the default less than reporting level at a concentration called the LRL where the false negative error rate is minimized to no more than 1 percent. If the analyte is present in the sample at a true concentration equal to or greater than the LRL, it should be detected 99 percent of the time, although measurement error may cause the reported concentration to be slightly different (higher or lower) than the true concentration.
This document describes the statistically based approach that the National Water Quality Laboratory is implementing to determine the LTMDL and to set the LRL. The new low-concentration reporting conventions are detailed, and advantages of its use are highlighted. Historical use and limitations of the minimum reporting level (MRL) in NWQL methods are noted. Preliminary guidance on interpretation of data reported by the NWQL using the new reporting conventions is provided.