WORK PLANS - Laserfiche WebLink

Home Browse Search

estimates. Some of the statistical limits used to estimate the BTVs for data sets with and without NDs as incorporated in ProUCL 4.0 are listed as follows. 1. Upper Percentiles (e.g., 95%, 99%) for data sets without and with NDs (e.g.,based upon KM estimates) 2. Upper Prediction Limit(UPL) for a future (site observation) observation(using KM or other estimates for data sets with NDs) 3. UPL for future k(e.g., next k or k site observations) observations 4. Upper Tolerance limits (UTLs) - Upper Confidence Limits for Upper Percentiles 5. Upper percentiles, UPLs,UTLs based upon data obtained using ROS methods—data with NDs 6. IQR Upper Limit(upper end of the upper whisker in a Box Plot) 7. UPL and UTL based upon resampling bootstrap 8. UPL based upon Chebyshev inequality 9. UTL based upon bootstrap methods for data sets with NDs 10. BTVs using nonparametric methods based upon higher order statistics (Conover, 1999) Note: The behavior of the exploratory IQR based upper limit as an estimate of a BTV is not well studied. This limit should be used with caution to estimate the BTVs or not-to-exceed values. It should be noted that background versus site comparisons based upon the BTVs are performed when not enough site data(e.g., <4-6 observations) are available to perform traditional two sample comparisons using hypotheses testing approaches such as t-test, Wilcoxon Rank Sum test, and Gehan test. When enough site data are available (e.g., at least 8-10, more are preferable), it is preferable to use hypotheses testing approaches to compare site data with BTVs or not-to-exceed values. Thus, in the absence of adequate amount of site data, individual point- by-point site observations are compared with some BTVs to determine the presence or absence of contamination due to site related activities. This method of comparing site versus background', level contamination is particularly helpful to use after some sort of remediation activities have taken place at the site; and the objective is to determine if the remediated site areas have been remediated enough to the background level contaminant concentrations. Typically, a site observation(possibly based upon composite samples)in exceedance of a background threshold value can be considered as coming from a contaminated site area that may have been impacted by the site-related activities. In other words, such a site observation may be considered as exhibiting some evidence of contamination at the site due to site related activities. In case of an exceedance of the BTV by a site location, some practitioners like to verify the possibility of contaminated site location by re-sampling (collecting 2-3 additional samples) that location, and comparing the sampled value(s)with the BTV. Hypothesis Testing Approaches Both single sample and two sample parametric and nonparametric hypotheses testing approaches are available in ProUCL 4.0. The hypotheses testing approaches as incorporated in ProUCL 4.0 can be used on full data sets without any ND observations, and on left-censored with nondetect