1
INTRODUCTION
THE Air Force has developed a probabilistic health-risk model, the Launch Area Toxic Risk Analysis (LATRA) model, to assist commanders in determining the risks to military personnel and civilians from exposure to emissions from normal and failed missile and space rocket launches. The model estimates the mean number of persons who might experience mild or serious health effects and the probability of exceeding each possible number of affected individuals. The Air Force Space Command requested that the National Research Council (NRC) independently review the toxicological components of LATRA to ensure their appropriateness. Specifically, the NRC was asked to focus on the toxicity of the three major rocket emissions—hydrogen chloride (HCl), nitrogen dioxide (NO2), and nitric acid (HNO3)—and several characteristics of the exposure-response components of LATRA, including identification of sensitive populations; definition of mild, moderate, and severe health effects; selection of independent variable in the exposure-response model; choice of analytic form for the exposure-response model (e.g., lognormal or probit) for each of the emissions; quantification of exposure-response model for each of the emissions; and representation and propagation of the uncertainties associated with the models. The NRC assigned this project to the Committee on Toxicology (COT), which convened the Subcommittee on Rocket-Emission Toxicants to respond to the request. Subcommittee members were chosen because of their expertise in inhalation toxicology, pharmacology, biostatistics, risk assessment, and environmental health. This report presents the subcommittee's evaluations, conclusions, and recommendations.
The remainder of this chapter provides a brief background of the development of LATRA and describes the organization of this report.
DEVELOPMENT OF LATRA
To assist commanders in making decisions on whether to launch a rocket given the weather conditions at the time of launch, the Air Force developed an atmospheric dispersion computer model, the Rocket Exhaust Effluent Diffusion Model (REEDM), which simulates the dispersion of a rocket's emissions under prevailing weather conditions. Specifically, REEDM predicts an isopleth, or ''footprint," of the concentrations of specific emissions at ground level downwind of the specific launch site.
Initially, the Air Force compared the exposure concentrations predicted by REEDM for each of the emissions with acceptable human exposure levels, called tier limits. Three different tier limits were developed for military and civilian base personnel and for the communities located around the launch centers. The derivation of those tier limits is described in more detail in Appendix A. If REEDM predicted that specific populations would be exposed at concentrations higher than the appropriate tier limits, the commander would be advised to hold the launch. The Air Force later decided that acceptable human exposure levels should not exceed one tenth of the National Institute for Occupational Safety and Health (NIOSH) immediately dangerous to life and health (IDLH) levels, for occupational exposures on base, and short-term public emergency guidance levels (SPEGLs) developed by the NRC, at the breathing zone for the public (U.S. Air Force 1994). (See Appendix B for definitions of IDLH and SPEGL values.)
That policy remained in effect until November of 1994, when a Peacekeeper launch was delayed several times, and then postponed, because REEDM predicted, based on forecasted lift-off weather conditions, that a nearby town would be exposed to HCl at concentrations that would exceed the SPEGL. That cancellation cost the Air Force hundreds of thousands of dollars. The Air Force subsequently estimated that the use of the 1-hr SPEGL of 1 ppm as a maximum allowable concentration (i.e., a ceiling limit value) for HCl reduced the probability of
a rocket being launched as scheduled from nearly 100% (prior to 1994) to about 27%, substantially increasing the overall costs of rocket launches. That information resulted in a reevaluation of the HC1 toxicity criteria and increased emphasis on implementing a probabilistic model, LATRA, to replace the simple comparison of acceptable human exposure levels with REEDM isopleths.
The LATRA model is designed to estimate the probabilities of various adverse health effects from exposing specified human populations to specific toxic emissions during rocket launches. It includes two major components: (1) a version of REEDM to predict downwind exposure concentrations and (2) exposure-response functions (ERFs) that relate the estimated exposure concentrations to expected health effects. LATRA estimates the mean number of people affected and the complete risk profile (the curve of the probability of exceeding each possible number of individuals affected) resulting from the rocket emissions from normal and failed rocket launches.
ORGANIZATION OF THIS REPORT
The remainder of this report is organized in six chapters with accompanying appendices. Chapter 2 describes the LATRA model and issues concerning its exposure-response components in more detail. Additional information on the rocket emissions, the derivation of the Air Force tier limits, and the relationship of those limits to the toxicity values used to develop LATRA are described in Appendix A. Definitions of established toxicity reference values that the Air Force considered in developing LATRA are provided in Appendix B. Chapter 3 provides the subcommittee's evaluation, conclusions, and recommendations concerning the identification of sensitive populations in light of the data available for the three rocket-emission toxicants and similar compounds. Chapter 4 provides the subcommittee's evaluation, conclusions, and recommendations concerning the definition of severity of effects. (Appendix C provides supplementary information for Chapter 4). Chapter 5 provides the subcommittee's evaluation of the structure of the LATRA-ERF model with respect to the characteristics the Air Force identified for review (noted above) and identifies possible alternative approaches to establishing exposure-response relationships or estimating incidence of
effects. Chapter 5 concludes with suggestions for improving the LATRA-ERF model and for alternative approaches to estimating health risks for rocket emissions. Chapter 6 provides examples of implementing the suggested alternative approaches for developing ERFs for HCl, NO2, and HNO3 on the basis of the available exposure-response data for those compounds, which are presented in Appendices D, E, and F, respectively.