Petition for Correction of CSB Report 2014-01-1-WV, 11/1/16 Requester: Philip C. Price, PhD I meet the “affected person” requirements of “CSB Final Data Quality Guidelines”. I was (and am) a resident and homeowner at 1391 Nottingham Rd., Charleston, WV at the time of the Freedom Industries spill on 1/9/2014. My contaminated water was provided by West Virginia American Water. My wife and I both had inhalation-based irritation and rashes from dermal exposure. I am an Analytical Chemist with more than 40 years of experience investigating Superfund sites, chemical incidents, and cancer clusters. Request: A major revision or addendum must be issued for this seriously flawed report. It contains misstatements of fact, unsubstantiated allegations, and critical omissions. It would not pass standard peer review, nor be acceptable for publication in a refereed scientific journal (I review for several international journals). This incident record will have no historical use, if viewed as flawed. After two and a half years of work, the CSB Report fails to address the most fundamental questions of a chemical spill investigation: • • • • • What chemicals were spilled? How much of each chemical was spilled? When did the spill happen? How did the spill happen? Who received what relative exposures? (which residents' neighborhoods, census tracts) There is no definitive analysis of the spilled material. The CSB report only lists components provided by Freedom and Eastman. We know Freedom employees were mixing Crude MCHM with caustic PPh, adding hydrochloric acid to neutralize, removing an unknown precipitate, and removing water. It is certain that these reactions generated new components from the original chemicals. These starting mixtures were low purity products, so an additional 10,000-50,000 ppm of unknown components were also initially present. It would appear that government agencies (page 17-18 of CSB report) are “taking Freedom’s word for it” as to what spilled and how much. Page 109 of the report shows the measured Flash Point was lower than expected (volatiles present), yet no “new components” or inappropriate concentrations were detected? Page 124 reiterates that it was a 6-component spill. Toxicology work by the National Toxicology Program (http://ntp.niehs.nih.gov/results/areas/wvspill/studies/index.html ) was clearly done on the wrong material. Their report concedes that “There was uncertainty regarding the exact concentrations of chemicals in the Freedom Industries storage tank that leaked”. The work done by OSHA at their Salt Lake lab (footnote 5 of CSB report) was inappropriate and misrepresented by the CSB as a “bulk analysis”. It would not be possible to identify (or even separate) these alcohols, esters, acids, and ethers solely by EI-based GC/MS. Best practices would dictate additional LC and IC separations, coupled with CI, ESI, APCI, or other specialized ionization. Given the importance of clarifying a public health exposure to over 250,000 people, I would also expect accurate mass determinations for all identified components above a few ppm concentration. Without a clear definition of what the public was exposed to, how can one evaluate public health effects? How is it possible for the NTP to do “toxicology testing”? The size of the spill is listed as 38 tons; this information was provided by the convicted spillers' records, and never verified by a mass balance during cleanup and remediation. Total inventory of MCHM mixtures in Tanks 395, 396, & 397 at time of spill was more than 110,375 gallons; or 410 tons. http://www.dhsem.wv.gov/Documents/Freedom%20letter%201-23.pdf We know material was recovered from site and booms on the river. More material was stored on the GAC inside the WVAmW plant. Thousands of cubic yards of MCHM-contaminated soil were removed from site during remediation. At only 10 ppm concentration on dirt, a 14 foot deep excavation under the tank pad area would yield an additional 20,000 lb of MCHM. Finally, concentration modeling of WVAmW's distribution system could estimate the pounds of MCHM flushed to customers. Is it possible the spill was significantly larger? The leak timeline is not substantiated. January 9 “during the morning”, Robert Keatley observed about 1,000 gallons out on the ground. The CSB found 0.4-0.75-inch holes in the bottom of the 48,000-gallon tank 396; at least 38 tons of MCHM blend passed through the two small holes. The tanks were 20 feet high and partially full, so the initial head pressure forcing out fluid was 10-30 psi. The CSB postulates (with no supporting data) a maximum flow of 11.5 gallons per minute (or more than 15 hours to generate the spill). The report also alleges a tank emptying rate of “one inch per 17 minutes” (calculations withheld), with one day to empty the tank. We know there was a direct path from subsurface soil to the river. During the WV DEP's site remediation, significant MCHM concentrations were found more than 4 meters below the soil surface (Test Pit #6): http://www.dep.wv.gov/dlr/oer/voluntarymain/Documents/FreedomVRPSiteVisit_8-26-15-Final.pdf This may indicate a much more long-term leakage event. How long was MCHM flowing into the water intake? Is it possible the tank was leaking, undetected, for a very long time? The CSB report shows that corrosion caused holes in the inside bottom of the tank, which were thoroughly documented. They assumed (with no data) a fixed corrosion rate of 10-15 mils/year. Several researchers (Eastman documents, Whelton, et al.) have noted the corrosiveness of “Crude MCHM”. However, no Bronsted acids were identified anywhere in the CSB's report. The CSB report postulates that rain entered through the roof of the tank. No evidence (photos, drawings) was presented. They suggest “aqueous corrosion” (p 32), but are inconsistent as to the source: “no corrosion holes were found on the roof”- p 30 “holes identified in the roof” - p 32 Although the steel tank sat directly on the ground, “aqueous corrosion” only caused pitting on the inside? No analytical work was done to identify organic or mineral acids in liquid from the failed tank. What was the source of the acid that corroded the tank? Defining where people were exposed to chemicals from an incident is important for understanding symptoms in different groups. A nine-county area served by WVAmW's Charleston Plant was affected. Homeland Security developed a fragmentary database of MCHM analyses done for many weeks throughout the system. These data should be mapped to generate estimates of which residences (census tracts?) received the highest total loadings of the spill. 22% of households reported (p 60 of CSB report) a health effect. The CSB report has a “Toxicology” section (Appendix D), but omits any reference to this spill. Obviously, the NTP studies were not done with material matching the Freedom spill, since it has not been properly analyzed. The CSB report does not address oral vs. inhalation ingestion, and avoids any geographic specifics. Do we understand where the spill plume spread, and levels of human exposures? Finally, the CSB report only makes 3 formal recommendations (pages 113-4). This seems to indicate the CSB did not fully understand the chain of events that led to this spill's negative impacts on Kanawha Valley residents. Blame is not productive, but specific actions to prevent a future event are needed. There were no recommendations for: • • • • • • WV American Water Company, even though they agreed to pay $126 million for their lack of preparations and inappropriate response. WV DEP, who had a responsibility for tank inspections, stormwater permits, mixing & blending permits, SDSs (still missing) Center for Disease Control, although they communicated two conflicting “safety messages”, and used calculations that did not consider exposures correctly. Department of Homeland Security, although they selected inappropriate methods, used poor record-keeping, and were unaware of how to correctly report low level analyses. WV Bureau of Public Health, which still does not have SDSs for chemicals in the Zones of Critical Concern, and has not audited nor approved WVAmW's SWPP. State of WV – does not regulate nor record locations of brine waste from oil and gas operations. These represent known and observed toxic (fluoride) and radiological (radium & radon) hazards to surface water. Other potable water systems in WV have already had incidents of raw water fluoride plumes at or above EPA's 4-ppm MCL. Part 3 of R2 (to American Water Works) suggests “as modeled by WVAW's Kanawha Valley Water System June 2016 Source Water Protection Plan”. Obviously, the CSB never examined this document. It has not been approved, and has obvious deficiencies. A more egregious example is the omission of Yeager Airport as a water risk. An event there could release jet fuel, aviation gasoline, de-icer, or PFOA-based runway foam – and the Elk River entrypoint would be at the Freedom site. A Source Water Protection Plan should include source water monitoring. WVAmW has selected a monitoring method based on UV detection of organics, and not a true “chemical Total Organic Carbon” analyzer. WVAmW's cheaper monitor will not effectively detect diesel fuel, gasoline, aircraft de-icer, glycolbased anti-freeze, paint thinner, alcohol, naphtha, and many other commonly-spilled materials. A failure to effectively deal with these many issues negatively impacts the “affected persons”: • • • • • Incorrect and incomplete identification of spill components could have serious public health implications. Lack of these specifics decreases the CSB report's credibility. Failure to confirm the size of the spill affects our understanding of exposure amounts, remediation efforts, GAC in-use-properties, spill modeling, etc. Failure to clarify the corrosion source perpetuates a lack of understanding for other tanks and chemical systems. Empirical corrosion data is useful to prevent other events. Indeterminacy in the event's timing affects our understanding of water source monitoring (control charting), GAC properties, and exposures. Failure to understand the spatial, temporal, and concentration spread of contaminants through the water distribution system means that the actual public exposures are unknown. Unless many more specific actions are taken by many agencies, we will have a recurrence of a similar event. Actions will be taken based on correct information in the CSB Report. I would hope that the CSB would follow its Final Data Quality Guidelines, p7, and more fully apply “commonly accepted scientific... standards” to its work.