WORK PLANS - Laserfiche WebLink

Home Browse Search

Upper Limits to Estimate Background Level Threshold Values (BTVs) or Not-to-Exceed Values ProUCL 4.0 can be used to compute several parametric and nonparametric upper limits that are used to estimate the BTVs or not-to-exceed values for data sets with NDs and without NDs. These upper limits include: upper prediction limits(UPLs),upper tolerance limits (UTLs), and upper percentiles. Some of the nonparametric methods such as the Kaplan-Meier(Meier, 1958) method and ROS methods are applicable on left-censored data sets having multiple detection limits. The background statistics as incorporated in ProUCL 4.0 are particularly useful when individual site observations from some impacted site areas(perhaps after some remediation activities) are to be compared with BTVs to determine if adequate amount of remediation and cleanup has been performed yielding remediated site concentrations comparable to background level concentrations; that is if the site concentrations can be considered as coming from(or approaching to)the population of background concentrations. The process of comparing individual site observations with BTVs or some other not-to-exceed values is also used for screening purposes (e.g.,before performing any cleanup and assessment) to identify the COPCs, and to determine if site areas under study need further sampling and remediation actions. Specifically,the process of comparing onsite data with the BTVs may help the working crew, project team, or the decision makers to take immediate decisions if more remediation and more onsite sample collection need to be performed at the site areas under investigation. The first step in establishing site specific background level contaminant concentrations for site related hazardous pollutants is to perform background sampling to collect appropriate number of samples from the designated site specific background areas or some agreed upon site reference areas. An appropriate DQO process(EPA, 2006) may be followed to collect an adequate number of background samples. It is desirable to collect at least 10-15 background samples to compute reliable estimates of BTVs. Furthermore, it is suggested not to use estimated BTVs and not-to- exceed values based upon background data sets of sizes smaller than 8-10. Once, an adequate amount of background data have been collected,the next step is to determine the data distribution. This can be achieved by using exploratory graphical tools (quantile-quantile (Q-Q) plots and histograms) as well as formal GOF tests as incorporated in ProUCL 4.0. Once the data distribution of a background data set has been determined, one can use parametric or nonparametric statistical methods to compute background statistics. A review of the environmental literature reveals that one or more of the following statistical limits are used to compute the background statistics; that is to determine and estimate background level contaminant concentrations. Collectively,these statistics represent estimates of the background threshold values (BTVs). The BTVs are estimated by statistics representing values in the upper tail(e.g., 95%upper percentile, 95%UPL) of the background data distribution. Typically, a site observation(preferably based upon a composite sample) in exceedance of a BTV (e.g.,UPL, upper percentile) can be considered as coming from a site area(location),which might have beeri impacted by the site-related activities. In other words, such a site observation may be considered as exhibiting some evidence of contamination at that site area(location) due to site related activities. For data sets with NDs,the BTVs can be estimated using upper limits based upon KM