February 22, 2002

Clarifications on:

Data Evaluation Summary of 2001 Split Sample Analytical Results

In your letter, you state that the letter report posted on the COPA web site does not contain the polychlorinated biphenyl (PCB) result for the split sample collected by PSARA Technologies Inc. (PSARA), which correlates with Tetra Tech split sample MW-6I (the document on the web site erroneously states that Tetra Tech's designation for this sample was "MW-61"). The Tetra Tech letter report submitted to the U.S. EPA on November 16, 2001, actually contains the PCB analytical result that is missing from the document posted on the web site; the PCB concentration in PSARA sample BD9219 as reported by PSARA was 0.65 micrograms per liter (ug/L). It appears that when Tetra Tech's letter report was scanned and posted on the web site, some information was either lost or not properly transmitted. In the document posted on the web site, a number of errors involving sample designations, an example of which is cited above appear to be present. However, the errors are minor and do not affect the document's presentation of sample analytical information. Enclosed is a copy of the November 16, 2001 Tetra Tech report.

Differences Between Results Reported by Two Laboratories

In your correspondence, you disagree with Tetra Tech's explanation of the differences between split sample analytical results reported by the two laboratories and stated that additional explanation should be offered for the differences observed.

The U.S. EPA acknowledges that its letter report should not have expressed an equivalent level of confidence in the PCB analytical results reported by both laboratories. The U.S. EPA or Tetra Tech did not receive the data packages prepared by PSARA's laboratory, and therefore could not review those data packages to comment on the quality of the PCB results reported by PSARA. However, as discussed below, based on a review of the data packages prepared by Tetra Tech's subcontractor laboratory, TriMatrix Laboratories, Inc. (TriMatrix), the U.S. EPA and Tetra Tech believes that the PCB results reported by TriMatrix are of adequate quality.

The typical acceptance criterion for PCB analysis of aqueous samples, which was derived from an industry-wide analysis of performance evaluation (PE) samples, is a recovery ranging from 53 to 129 percent. Based on this criterion, if two PCB results for a given sample are within a factor of 2.4, the observed difference is within the typical variability. Of the PCB results reported by the two laboratories for 13 split samples, the results for 4 samples differ by more than a factor of 2. Specifically, the PSARA results for the 4 samples were greater than the TriMatrix results by factors of 2.8, 8.7, 9.2, and 14; such differences cannot be easily attributed to typical analytical method variability. For this reason, Tetra Tech reviewed the TriMatrix data packages again. The following observations made during the review support Tetra Tech's conclusion regarding the quality of the PCB results provided by TriMatrix:

• TriMatrix analyzed a total of 52 field and laboratory quality control (QC) samples for surrogates. The surrogate recovery acceptance criteria were met for 48 of these samples; the surrogate recoveries for the remaining 4 QC samples were not reported because these samples were analyzed at a dilution factor greater than 5.
• TriMatrix analyzed 11 blank spike or laboratory control samples for PCBs. The acceptance criterion was met for all 11 samples. The sample recoveries ranged from 76 to 101 percent, with mean and median recoveries of 90 and 92 percent, respectively.
• The results for all instrument and method blanks met the acceptance criteria.

In September 2001, TriMatrix met the performance acceptance criteria for two aqueous PE samples submitted as blind samples for PCB analysis. Specifically, for one sample that had an acceptance criterion ranging from 1.18 to 2.86 ug/L, TriMatrix reported a concentration of 1.7 ug/L. For the other sample, which had an acceptance criterion ranging from 5.10 to 13.7 ug/L, TriMatrix reported a concentration of 8.6 ug/L.

In summary, the U.S. EPA and Tetra Tech agree that the differences between the two laboratories' results for 4 of the 13 split samples are significant. However, these differences cannot be explained based solely on a review of the TriMatrix data packages. The U.S. EPA has requested that Viacom send the complete data packages for the four samples in question for review by the U.S. EPA. Because the TriMatrix QC results do not indicate a significant low bias, we can only speculate that the PSARA results for the four samples in question may be biased high or contain a large amount of sample variability. Because EPA is using a conservative approach to determine whether samples contain PCB concentrations above the cleanup goal, and because PSARA's results for the split samples collected in 2001 either were biased high compared to the TriMatrix results or were not significantly different from the TriMatrix results, the U.S. EPA believes that the cleanup decisions made based on the results are defensible. Once the U.S. EPA reviews Viacom's data packages for the four samples, the Agency will notify you of our findings.

Included with your letter was correspondence similar to a letter the U.S. EPA received from Mike Baker of the Citizens Opposed to PCB Ash (COPA) regarding a number of issues associated with the Bloomington PCB sites. The U.S. EPA responded to the COPA concerns in correspondence, dated August 21, 2000. The U.S. EPA response was sent to both COPA and the Monroe County Public Library. Enclosed is a copy of the August 21, 2000 letter addressing the concerns described in the COPA letter.

The U.S. EPA must respond to statements in your letter referencing laboratory misdeeds. First, no evidence exists that laboratory misdeeds have occurred for any of the sampling associated with any of the Bloomington PCB sites. As clarification, an individual who worked for the U.S. EPA and an U.S. EPA contractor were investigated for fraud concerning improprieties associated with running the U.S. EPA Region V laboratory. Based on the evidence, the Department of Justice indicated that data for any of the sites that these individuals worked on were questionable and in many cases resampling was required. These individuals were involved in the analysis of the 1996 Lemon Lane Landfill boring data and as a result the sampling was repeated in January 2000. The new sampling data demonstrated that the Lemon Lane Landfill data was consistent and that the original data did not appear to be questionable. The U.S. EPA laboratory has not been used for analysis since the 1996 sampling event. Again, to infer that laboratory misdeeds have occurred for samples associated with the Bloomington PCB sites is incorrect.

Finally, you infer in your letter that data used for the Bloomington investigations/cleanups has continued to raise data quality questions. The U.S. EPA takes data quality seriously and works closely with Tetra Tech and TriMatrix Laboratories. The U.S. EPA has taken measures such as laboratory audits and analysis of performance samples to ensure that good quality data is generated. The reports developed by Tetra Tech summarizing the split sampling with Viacom for Neal's Dump, Neal's Landfill, Winston Thomas Wastewater Treatment Plant, Bennett's Dump, and the Lemon Lane Landfill all showed consistent results. In situations where analytical results showed disagreement between Viacom's laboratory and U.S. EPA's laboratory, the higher PCB concentration was always used. We have no current evidence that data quality problems exist for the data generated for the Bloomington sites.

Sincerely,

Thomas Alcamo
Chemical Engineer


Thomas Alcamo
US EPA Region 5
77 West Jackson Blvd.
Chicago, IL 60604-3590