![[logo] US EPA](http://www.epa.gov/epafiles/images/logo_epaseal.gif){kind=logo}
Technical Support Center for Monitoring and Site Characterization
Statistical Software ProUCL 4.0 for Environmental Applications For Data Sets with and without Nondetect Observations
The statistical software package, ProUCL 3.0, has been upgraded to ProUCL 4.0 and includes statistical methods that can be used to estimate exposure point concentration (EPC) terms, not-to-exceed, and background threshold values (BTVs) for data sets with nondetect (ND) and without ND observations. ProUCL 4.0 retains all of the capabilities of ProUCL 3.0. ProUCL 4.0 addresses various statistical issues arising in: exposure and risk assessment studies, in background evaluations, and in background versus site comparison applications. Specifically, most of the statistical methods described and recommended in the Background Guidance Document for CERCLA Sites (EPA, 2002a), and in the Guidance Document to compute 95% Upper Confidence Limits (EPA, 2002b) have been incorporated in ProUCL 4.0. ProUCL 4.0 has statistical methods that can be used to verify the attainment of cleanup standards (EPA, 1989), and to estimate screening levels (EPA, 1996) for data sets with and without NDs. Some of the statistical methods (e.g., two sample hypotheses, upper prediction and tolerance limits) as incorporated in ProUCL 4.0 can be used in groundwater (GW) monitoring applications (EPA, 1992).
ProUCL 4.0 has goodness-of-fit (GOF) tests (e.g., normal, lognormal, and gamma) for data sets with and without NDs. For data sets with NDs, ProUCL 4.0 can create additional columns to store extrapolated (estimated) values for NDs obtained using regression on order statistics (ROS) methods including normal ROS, gamma ROS, and lognormal ROS (robust ROS) methods. ProUCL 4.0 can process multiple contaminants (variables) simultaneously in batch mode. ProUCL 4.0 also has the capability of processing data by groups. A valid group column should be included in the data file. ProUCL 4.0 has a couple of simple outlier test procedures, such as the Dixon test and the Rosner test. ProUCL 4.0 offers useful graphical displays for data sets with and without NDs including: histograms, multiple quantile-quantile (Q-Q) plots, and side-by-side box plots. The use of graphical displays provides additional insight contained in data sets that cannot be revealed by the use of estimates (e.g., 95% upper limits) and test statistics, such as GOF test statistics, the Rosner test, and various other statistics. In addition to providing information about the data distribution (e.g., normal, gamma), Q-Q plots are very useful in identifying potential outliers and the presence of mixture samples in a data set. Side-by-side box plots and multiple Q-Q plots are useful to visually compare two or more data sets (groups) such as: site versus background contaminant concentrations, surface versus subsurface concentrations, contaminant concentrations of groundwater monitoring wells (MWs), and areas of concern (AOCs) of a potentially polluted site.
ProUCL 4.0 has several parametric and nonparametric single sample and two sample hypotheses testing approaches. Single sample hypotheses (e.g., Student’s t-test, the sign test, the Wilcoxon signed rank test, and the proportion test) can be used to compare site mean concentrations (or some site threshold value such as an upper percentile) with some average cleanup standards, Cs (or a not-to-exceed compliance limit, A0), to verify the attainment of cleanup levels (EPA, 1989, EPA, 2006) after some remediation activities have been performed at impacted site areas. Two sample hypotheses tests (e.g., Student’s t-test, the Wilcoxon-Mann-Whitney test, the quantile test, and Gehan’s test) can be used for site versus background comparisons, comparing concentrations of site AOCs, and comparisons of contaminant concentrations of two or more MWs. The hypothesis-testing approaches can be used on both uncensored (without NDs) and left-censored (with NDs) data sets. ProUCL 4.0 also has parametric (e.g., maximum likelihood estimate (MLE), t-statistic, gamma distribution), nonparametric (e.g., Skewness adjusted CLT, Kaplan-Meier), and computer intensive bootstrap (e.g., percentile, BCA) methods to compute upper confidence limits (UCLs), upper prediction limits (UPLs), and upper tolerance limits (UTLs) for uncensored data sets and also for data sets with ND observations. Some of the methods, such as the Kaplan-Meier (KM) method and the robust ROS methods, are applicable on left-censored data sets having multiple detection limits.
References
USEPA (1989). Methods for Evaluating the Attainment of Cleanup Standards, Vol. 1, Soils and Solid Media, Publication EPA 230/2-89/042.
USEPA (1992). Statistical Analysis of Ground-water Monitoring Data at RCRA Facilities. Addendum to Interim Final Guidance. Washington DC: Office of Solid Waste. July 1992.
USEPA (1996). Soil Screening Guidance: User’s Guide. Office of Solid Waste and Emergency Response, Washington D.C. EPA/540/R-96/018, April 1996.
USEPA. 2002a. Guidance for Comparing Background and Chemical Concentrations in Soil for CERCLA Sites. EPA 540-R-01-003-OSWER 9285.7-41. September 2002.
USEPA. 2002b. Calculating Upper Confidence Limits for Exposure Point concentrations at Hazardous Waste Sites. OSWER 9285.6-10. December 2002.
ProUCL 3.0. (2004). A Statistical Software. National Exposure Research Lab, EPA, Las Vegas Nevada, October 2004.
USEPA. 2006. Data Quality Assessment: Statistical Methods for Practitioners, EPA QA/G-9S. EPA/240/B-06/003. Office of Environmental Information, Washington, D.C. Download from: http://www.epa.gov/quality/qs-docs/g9s-final.pdf [198 pp, 2.4MB, About PDF]
Changes from ProUCL 4.0 (version 4.00.00) to ProUCL 4.00.02
Although extensive changes were made in the code for ProUCL 4.0 (version 4.00.00) to produce ProUCL 4.00.02, those changes are transparent to the users. Most of those changes were made so that ProUCL 4.00.02 is compatible with our developing statistical software, Scout (e.g., both programs share the same statistical libraries). ProUCL will also reside as a separate module in Scout as a research tool.
There is a very minor correction of a displayed value in one of the hypothesis tests, the two sample t-test. The p-value associated with the t-test was computed in two different ways: one way is correct and the other way, although it produced subtle differences, is incorrect. The incorrect method has been removed from ProUCL 4.00.02.
Several extra warning messages have been added to ProUCL 4.00.02, mainly in regard to attempting tests when a data set is very small (n < 5), when the number of detected values is small (e.g., only zero, one, or two), or when all of the values are non-detected values. For an example, some screens depicting those warning messages are included in the newly added Section 2.11 (page 40) of this ProUCL 4.00.02 User Guide.
The only software files that were changed from ProUCL version 4.0 (4.00.00) to version 4.00.02 were updates in the ProUCL.exe file, and updates to the StatsLib.dll file to produce a more advanced ScoutLib.dll file. Very minor changes were made to this ProUCL 4.00.02 User Guide, including: changes to avoid inappropriate user inputs (warnings), changes to the title page, the inclusion of an acknowledgement page, and the inclusion of a contact information page.
No changes were made to the ProUCL 4.0 Technical Guide; that is, the ProUCL 4.0 Technical Guide is still applicable to the ProUCL 4.00.02 software and User Guide.
Contact Information for ProUCL 4.00.02
The ProUCL software is developed under the direction of the Technical Support Center (TSC). As of November 2007, the direction of the TSC is transferred from Brian Schumacher to Felicia Barnett. Therefore, any comments or questions concerning ProUCL should be addressed to:
Felicia Barnett, (HSTL)
US EPA, Region 4
61 Forsyth Street, S.W.
Atlanta, GA 30303-8960
barnett.felicia@epa.gov
(404) 562-8659
Fax: (404) 562-8439
Facts Sheet for ProUCL 4.0 (PDF) (18 pp, 535KB)