SITE INFORMATION AND CORRESPONDENCE

Home Browse Search

be stated before an adequate number of sampling locations can be determined. The Work Plan fails to address the statistical parameter of power and does not adequately describe the rationale for establishing a 200% minimum detectable difference. CVs are likely to vary with contaminant type. Exhibit 3-12 on page 33 of Data Useability in Risk Assessment (1990) gives a range of CVs for various contaminants at 8900 Superfund sites; these values range from less than 1% for some SVOCs to over 100% for some metals (beryllium, cadmium, and mercury) . It should be noted that a CV of 200% would necessitate at least four or five samples, rather than three samples as recommended. Reasonable values for all parameters (CV, confidence level, power and minimum relative difference) should be discussed more fully, with respect to background values. According to this same guidance document, minimum recommended performance values are 80% for confidence level and 90% for power. 15. The Work Plan considers environmental contamination, if found, will be lognormally distributed (the data) . While lognormality is often assumed for environmental data, the validity of the assumption should be tested. It is not clear how DDRW-Tracy will test for lognormality and what they will do if the results show a different distribution. EPA recommends use of the 95% upper confidence level of arithmetic mean in deriving the Reasonable Maximum Exposure (RME) concentration as this generates a conservative estimate to be used in the baseline risk assessment. The decision to use arithmetic or geometric means (normal distribution versus a lognormal distribution) needs to be based on fit tests for normal distributions or for lognormal distributions using the Kolmogorov Smirnov statistic. 16. The adequacy of the detection limits should be evaluated in the Work Plan by presenting a table listing and comparing the method of detection or quantitation limit for each compound with the ARARs or, preferably, risk-based goal for each chemical in each medium. 17 . The initial numbers of soil borings may determine the nature of contamination but not the total extent of contamination for each SWMU area. The additional soil borings mentioned will probably be necessary to determine both the nature and extent of contamination as is mandated in the National Contingency Plan. The Work Plan needs to address both the nature and the extent of contamination. Describe in the Work Plan a provision for interim status reports for the SWMUs and for the USTs and some provision for regulatory input along the way. I-5