Harvard School of Public Health ENVIRON MENTAL SCIENCE AND ENGINEERING PROGRAM August 20, 1997 Mr. M. Royack Tank Operations Program Manager Tank Farm Operations U.S. DOE P.O. Box 550 Richland Operations Ofiice Richland, WA 99352 Dear Mr. Royack: In response to your request to the WHS to review the PNNL report entitled, ?Health Risk Assessment for Short- and Long-?term Worker Inhalation Exposure to Vapor-phase Chemical ?'om the Single-shell Tank 241-0103.? and evaluate its scienti?c merit and relevance to current operations, the enclosed critique has been prepared and approved by all the members. Their names are listed at the end of the report. If there are questions concerning the contents of the critique, or if you desire fuller explanations on any points, kindly direct them to me. I will share them with the other panel members and send you a consensus response. Sincerely, Melvin W. First, Professor of Environmental Health Engineering, Emeritus cc: C. Abrams M. Corn D. Oakley G. Schmauch .L. Kovach EC E. tr In 7 7 1997 Mailing address: HARVARD AIR CLEANING LABORATORY "0 665 Huntington Avenue, Boston, Massachusetts 02115 Telephone; 617.432.1164 FAX: 617-432-3349 A Critique of the PNN Draft Report ?Health Risk Assessment for Short? and Long-term Worker Inhalation Exposure to Vapor-phase Chemical from the Single-shell Tank By the Worker Health and Safety 20 August 1997 It is not stated why this study was commissioned by WHC or what instructions were given to PNNL. The Tanks Advisory Panel (TAP) and the Worker Health and Safety subpanel of TAP had for some considerable time prior to the report?s initiation been urging WI-IC to rescind its blanket order requiring supplied air respirator use at the tank farms infavor of a farm-by- farm and tank?by-tank policy based on demonstrated need. Although Tank 241-C-103 (hereafter referred to as C-103) was considered one of the least likely, on the basis of its history, to qualify for release of respirator requirements, this perception may have been the motivation for ordering the failed filter~fan?stack installation. Regardless, the requirement for exclusive use of supplied air respirators has been rescinded for Tank C-103 and a new policy, entered in the HASP, is currently in force. In consideration of the following facts; (1) the exhaust system is not now, and never was, Iimctional (2) the requirement for use of respirators in the vicinity of C-103 was never rescinded (3) the requirement for respirator use in the vicinity of 0103 will not likely be withdrawn until the tank is remediated, the labels the PNNL report, its ?ndings, and its conclusions unrealistic and irrelevant. Considering the ?merits of the report, there are serious errors of concept and process that challenge its value. With respect to the toxicity review, it is wrong to invoke and apply standards and methodology because they are based on exposure for 24-hrs per day, seven ?days per week, over a 7 0-yr lifetime. Different criteria apply for occupational exposures. Even when toxic risk assessments utilize only occupational criteria, the time of exposure is a critical component of exposure assessment and dose estimate. The PNNL report ignores the real exposure duration (as opposed to a hypothetical and incorrect exposure duration), when, in fact, it has been estimated reliably that routine visits to represent approximately 10-min. per week. From this consideration, alone, it should have been evident that risk assessments based on chronic diseases, including cancer are otiose. It has never been proposed, or practiced, that a level of protection less than supplied air respirators be used whenever working around any tank, including C-103, whenever covers have been opened, thereby reducing vapor inhalation exposures to a level of no concern during these periods. This work activity would never be a factor in vapor exposure estimates when not wearing a respirator. Therefore, it should have been obvious that the only exposures of interest would be peaks of short duration, not exceeding 15-min, if one were to analyze realistic exposures in the vicinity of C-103. Perhaps the most contradictory statement of the report occurs in the ?Discussion and Summary?, where the report is characterized as ?This qualitative evaluation.? This, in spite of pages and pages of tabular data incorporating 2 to signi?cant ?gures. How many signi?cant ?gures must one use to enter the quantitative zone? Equally troubling is a deliberate and consistent unwillingness to confront an absence of validation for a number of critical assumptions, as well as for the inherent arbitrariness of the chosen models; this-, despite frequent disclaimers of certainty regarding both throughout the report. For example, why select a head space vapor dilution ratio of 50 times? Why not 100, or 10? One searches in vain for con?dence limits for the derived risk assessment values expressed quantitatively (in spite of the disclaimer). One suspects that conservatively calculated con?dence limites would likely range from zero to in?nity. With regard to the basic information used by the PNNL reviewers, they state (para. 2.1) that, ?No effort was made in this risk assessment to validate the data.? We believe this decision makes their analysis and conclusions of no merit because the data they used so uncritically have been under serious question by both the TAP and since their first appearance, First, the identi?cation and quanti?cation of the 221 chemicals reported by Osborne have never been validated and those most familiar with this type of trace analytical methodology have expressed considerable doubt regarding the validity of the reported qualitative and quantitative results. Therefore, an exercise of critical judgement regarding the validity of the data in Table 2.1 should have been the ?rst step of the report. Second, both TAP and have been engaged for several years in an unsuccess?il effort to convince Hanford-based technicians that atmospheric di?iision models may not be used success?illy for determining stack emission concentrations close to the source, 1m, 2m, etc. The very least distance from the source at which data of any reliability whatsoever can be deduced is 100m, and several hundred meters would be more rational. A model output indicating that maximum contaminant concentrations occur at the base of a vertical stack should have alerted someone that something was wrong, inasmuch as everyone familiar with stack dispersion analyses knows that the concentration at that location will be zero. Furthermore, applying atmospheric dispersion models to a 10-in. stack 15-it above ground, carrying 1,000 CFM is, again, stretching the model to the point of nomapplication. The reference for the dispersion model used by PNNL has the title, ?Atmospheric Relative Concentrations in Building Wakes,? but there are no structures at the tank farm to form building wakes. Finally, with respect to stack dispersion studies, we cannot fathom why anyone concerned with workers? exposures would even consider a configuration where the e?luent is discharged horizontally 15?? above ground. Thus, their uncritical acceptance of the underlying data, combined with a non?scienti?c attitude toward subjecting their analyses to. statistical con?dence analysis, renders the conclusions of this report of no value. Although the report cites WHC as requesting analysis of maximum exposure, the driver behind this analysis appears to have been a concerted effort to maximize exposure and risk beyond all rational limits. For example, after calculating average annual exposures based on expected work locations, and ?nding the values to be in the nanogram and picogram per cubic meter range, the risk analysis did not use these values but, instead, assumed workers would be breathing undiluted stack effluent gases ?at the stack exit point," as noted in para. 2.5, for ?40 hr per week for 25 years, an absurd assumption. The authors also assume implicitly throughout that expose equals dose to a sensitive organ. The report omits many important details on the excuse that the ?report would be too volurninous?), and has been poorly edited, making it di?icult to follow much of the argument. In addition to textual errors, there are technical ones. For example: in para 2.2 it states, ?In the Table [21] an asterisk indicates that a surrogate chemical will be selected?, yet a footnote to Table 2.1 states, ?An asterisk indicates, tentatively identi?ed chemicals.? It is stated in the text (para 3.2) that the ?dose term? from Eq. 3.2 has units of m3/Kg-day when it should be mg/Kg-d. In para 2.2 it states that ?a newer listing of the chemical 1996. . was used for this assessment? yet in para 2.4 it states that the ?emission rate is based on the FY94 characterization of vapor.? Para 2.6 states that ?in-line I-IEPA ?lters [are used] to collect particles and gases? when it is well known that HEPA ?lters are perfectly transparent to gases. The deems the PNNL report to be unreliable, non-reflective of current operating procedures, and based on invalid assumptions. The is satis?ed that current procedures cited in the HASP are fully adequate for personnel safety. In summary, the believes this report should be withdrawn (retracted) in consideration of its non-relevance and the multiple errors. WI-IS C. Abrams, Chairman M. Corn M. First D. Oakley G. Schmauch