BLM Toxicologist's Questions and Alternative Suggestion

Interestingly there is not one mention of these questions or concerns in the DEIS and it does not appear that Hollister staff consulted with or considered Karl Ford's question, concerns or suggestions during the preparation of the DEIS 

February 8, 2008

Memorandum

To: Rick Cooper, Hollister Field Office Manager, Tim Moore, Hollister Field Office

From: Karl Ford, Ph.D, Toxicologist, National Operations Center

Re: Review of “Clear Creek Management Area Asbestos Exposure and Human Health risk Assessment,” Region 9 EPA

I have been aware of this ongoing study, and have seen some interim briefings. I have briefly reviewed the report and have the following comments.

1. I note that EPA uses similar activity-based sampling as has been performed previously at the CCMA. The major difference in this report is the use of a

different analytical method TEM (ISO 10312). Neither laboratory nor it’s certifications were identified. Conversations with experts and even this report

suggests that TEM reports higher concentrations than does PCM. Does EPA have comparable datasets to show have the two different methods compare?

The iRIS cancer slope factor is probably based on PCM, so use of TEM may overestimate risk.

2. Data validation of the useability of the results was not provided. Appendix C referred to on page 7 was not provided. What did the blanks show?

3. The data quality objectives were not identified for the risk assessment. If the sampling had been correlated to soil moisture levels, it would have been more

beneficial than just precipitation. For example, the November 4 period designated by EPA as “moist” had greater asbestos concentrations than some “dry” periods.

4. The locations of the sampling were not identified. There are clearly areas within the CCMA that have much higher serpentinite than others and these areas have

been mapped on geologic maps.

5. Soil sampling is mentioned on page 9, but the Appendix F was not provided.  Where did the samplers ride and how do asbestos concentrations correlate to soil

samples? Did they ride in areas of lesser asbestos concentrations or just high areas? Page 24 indicates this as a possible source of uncertainty.

6. Page 10 and computation of means and 95% UCL. It is understood that EPA guidance has a preference for arithmetic means and 95% UCL, however, the 1

limited information available to me suggests the data are probably log-normally distributed. If true, the arithmetic mean and 95% UCL may overestimate the true

concentrations. I would like to see the distributions evaluated and, if appropriate, log means and log UCLs using the Land H method used.

7. Inspection of Figure 4 shows the skewed distributions for the moist condition ATV rider and SUV rider. Most measurements were in the <0.25 fibers/cc range,

a significant bimodal distribution between lead and trailing child riders. Some other potentially useful variables important to BLM are the following distance,

number of riders, speed, etc.

8. Page 17 indicates the calculations to derive the UCL are in Appendix E. I could not find these calculations. If there were any non-detects, how were these

handled?

9. In Tables 1 and 2, were means computed including lead, middle and trailing?  Were samples from weather conditions, “dry,” “moist,” and “wet,” similarly

merged? It appears obvious that most of the samples were from dryperiods and the moist period (that apparently was not so moist), thus skewing the mean

during tile wet season. Skewing the dataset towards dry conditions does not help us answer the question. I would like to see the risk computed for the wet period.

the opposite trend.

14. A combination of some additional analyses I have suggested above (evaluate dataset distributions and log means, compute risk during wet conditions, evaluate