11

RISK ANALYSIS AND MANAGEMENT

This chapter summarizes information on population exposure to GWEN fields, compares GWEN exposures to existing standards and to exposures from other sources of nonionizing fields, draws on existing epidemiological evidence to derive upper bounds on GWEN-related risks, describes likely public concerns regarding risks associated with GWEN emissions, and suggests alternatives for addressing those concerns.

RISK ASSESSMENT

Scientific tools are too blunt to measure directly the human health risks associated with low-level exposures to most environmental agents1 Estimates of such risks are garnered instead by extrapolation both from studies at high exposure levels and from experiments with laboratory animals, cells, and tissues.

Risk assessment is an analytical tool for characterizing potential health risks to individuals or populations by combining existing epidemiologic and laboratory data on the potency of an agent with estimates of expected levels of human exposure. A National Research Council committee2 has identified four elements of a complete risk assessment: hazard identification, exposure assessment, exposure-response evaluation, and risk characterization. Hazard identification involves reaching scientific consensus on whether exposure to an agent can cause an increase in the incidence of some deleterious health condition. Exposure assessment is the characterization of the spatial and temporal patterns of exposure to the agent. Exposure response evaluation involves gleaning from epidemiologic studies and experiments on laboratory animals a functional relation between the level of exposure and the magnitude of the induced health hazard. Risk characterization combines estimates of exposure and estimates of effect per unit of exposure from the two preceding stages to derive quantitative statements about human health risk.

Each stage of this risk assessment process involves uncertainties. Agents shown to be hazardous to some animals might not be hazardous to humans. Agents that induce deleterious effects at high exposure levels might be harmless or even beneficial at very low exposure levels of exposure. Exposure assessments are hampered by limitations in data collection. Finite resources limit the number of exposure conditions and health endpoints for which health responses can be evaluated. Because these and other uncertainties in the risk assessment process can be large, it is important that uncertainty be explicitly incorporated in any risk characterization.

Uncertainties in risk assessment are particularly large in the case of exposures to subthermal levels of nonionizing radiation because of the combination of perplexing epidemiological and laboratory findings, the lack of a testable hypothesis for health effects,



The National Academies | 500 Fifth St. N.W. | Washington, D.C. 20001
Copyright © National Academy of Sciences. All rights reserved.
Terms of Use and Privacy Statement



Below are the first 10 and last 10 pages of uncorrected machine-read text (when available) of this chapter, followed by the top 30 algorithmically extracted key phrases from the chapter as a whole.
Intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text on the opening pages of each chapter. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

Do not use for reproduction, copying, pasting, or reading; exclusively for search engines.

OCR for page 143
ASSESSMENT OF THE POSSIBLE HEALTH EFFECTS OF GROUND WAVE EMERGENCY NETWORK 11 RISK ANALYSIS AND MANAGEMENT This chapter summarizes information on population exposure to GWEN fields, compares GWEN exposures to existing standards and to exposures from other sources of nonionizing fields, draws on existing epidemiological evidence to derive upper bounds on GWEN-related risks, describes likely public concerns regarding risks associated with GWEN emissions, and suggests alternatives for addressing those concerns. RISK ASSESSMENT Scientific tools are too blunt to measure directly the human health risks associated with low-level exposures to most environmental agents1 Estimates of such risks are garnered instead by extrapolation both from studies at high exposure levels and from experiments with laboratory animals, cells, and tissues. Risk assessment is an analytical tool for characterizing potential health risks to individuals or populations by combining existing epidemiologic and laboratory data on the potency of an agent with estimates of expected levels of human exposure. A National Research Council committee2 has identified four elements of a complete risk assessment: hazard identification, exposure assessment, exposure-response evaluation, and risk characterization. Hazard identification involves reaching scientific consensus on whether exposure to an agent can cause an increase in the incidence of some deleterious health condition. Exposure assessment is the characterization of the spatial and temporal patterns of exposure to the agent. Exposure response evaluation involves gleaning from epidemiologic studies and experiments on laboratory animals a functional relation between the level of exposure and the magnitude of the induced health hazard. Risk characterization combines estimates of exposure and estimates of effect per unit of exposure from the two preceding stages to derive quantitative statements about human health risk. Each stage of this risk assessment process involves uncertainties. Agents shown to be hazardous to some animals might not be hazardous to humans. Agents that induce deleterious effects at high exposure levels might be harmless or even beneficial at very low exposure levels of exposure. Exposure assessments are hampered by limitations in data collection. Finite resources limit the number of exposure conditions and health endpoints for which health responses can be evaluated. Because these and other uncertainties in the risk assessment process can be large, it is important that uncertainty be explicitly incorporated in any risk characterization. Uncertainties in risk assessment are particularly large in the case of exposures to subthermal levels of nonionizing radiation because of the combination of perplexing epidemiological and laboratory findings, the lack of a testable hypothesis for health effects,

OCR for page 143
ASSESSMENT OF THE POSSIBLE HEALTH EFFECTS OF GROUND WAVE EMERGENCY NETWORK and difficulties in reproducing effects in separate laboratories. The problem of assessing risk associated with nonionizing fields is further complicated by evidence that effects might not be monotonically related to field strength. The laboratory literature includes many examples that suggest that effects appear only within particular windows of frequency or field intensity.3, 4, 5 This evidence greatly enlarges the range of exposure conditions that an exposure-response evaluation must consider to assure completeness. Finally, GWEN LF frequencies are 3,000 times higher than those associated with the ELF epidemiological studies reviewed in Chapter 9. Risk extrapolations to GWEN LF exposures based on ELF epidemiologic results are, therefore, very sensitive to the choice of exposure measure. Because risk extrapolations with such large uncertainties are of little use in a policy context, they have not been pursued in this study. In summary, the application of traditional tools of risk assessment in the GWEN context are frustrated by (1) uncertainties over what, if any, human health end points are affected by exposure to GWEN fields, (2) a lack of evidence of a relationship between increasing risk and increasing levels of some measure of field exposure (i.e., a dose-response relationship) for nonionizing fields of any frequency, and (3) the uncertainty of the basis for extrapolating effects from one exposure regime (frequency, field strength, etc.) to another. Although it is premature to apply the traditional risk assessment process to GWEN exposure, a variety of analyses can provide valuable perspective on possible risks from GWEN emissions. In the remainder of this chapter, we summarize information on population exposure to GWEN fields, compare GWEN exposures to existing standards and to exposures to nonionizing fields from other sources of nonionizing fields, and draw on existing negative epidemiological evidence at or near GWEN frequencies to derive upper bounds on GWEN-related risks. We also discuss some important risk-management issues including likely public concerns regarding risks associated with GWEN emissions and ways to address those concerns. GENERAL DESCRIPTION OF GWEN FIELDS GWEN relay nodes (RNs) broadcast both low-frequency (LF) signals (150-175 kHz) and ultra-high-frequency (UHF) signals (225-400 MHz). GWEN's fixed input/output (I/O) terminals produce only UHF signals. GWEN LF and UHF transmitters are both frequency-modulated and transmit only periodically according to well-defined communication protocols. System design and protocol are such that time-averaged duty cycles for any LF transmitter can be no less than 0.0014 and no more than 0.27. At full RN implementation, typical LF duty cycles are expected to be about 0.014. Ground-to-air UHF transmitters would be constrained by design and protocol to time-averaged duty cycles of 0.019-0.27. Ground-to-ground UHF transmitters for the 16 RNs serving I/O nodes have a fixed duty cycle of 0.40. Unless otherwise noted, LF and UHF exposures reported below represent peak fields, not time-averaged fields.

OCR for page 143
ASSESSMENT OF THE POSSIBLE HEALTH EFFECTS OF GROUND WAVE EMERGENCY NETWORK GWEN LF broadcasts are made from a 299-ft vertical top-loaded monopole antenna and are isotropic in azimuth. The LF radiated power during LF broadcasts is nominally 2,000 W, but could be as high as 3,200 W, depending on the needs of the site. The intensity of the ground-level LF fields produced by a GWEN LF transmitter is shown in Figure 2-4. There are two types of GWEN UHF transmitters. UHF transmitters at most GWEN RNs are designed for airborne communication. They broadcast in a pattern that is omnidirectional in azimuth, but somewhat directional in elevation, with the strongest fields occurring in a shallow 8 ° cone extending outward and upward from the antenna position some 10 m above the ground. The radiated power from the UHF antenna is typically 50 W, but can be as high as 79 W with a duty cycle less than 0.27. The field profile of the antenna, in the cone of peak field strength with an assumed radiated power of 79 W is shown in Figure 2-5. Field strengths at ground level are somewhat smaller than shown in Figure 2-5, primarily because ground-level fields are strongly affected by ground reflections. For RNs constructed to date, UHF-transmitting antennae are outside the 4-ft fence that surrounds the LF antenna, a few tens of meters from the site boundary. Twelve GWEN RNs are equipped with a ground-to-ground UHF-transmitting antenna that is directional in both azimuth and elevation. This antenna is mounted on top of a pole 20-150 ft high with its main lobe of 9-dB gain directed horizontally. The radiated power from the antenna is typically 50 W with a 0.4 duty cycle and a narrow beam. In characterizing GWEN emissions in those two bands, it is important to distinguish between the far field, where the ratio of electric-field intensity to magnetic-field intensity is constant, and the near field, where the electric- and magnetic-field components of an electromagnetic wave are independent of one another. For GWEN transmissions, the near field extends to a couple of kilometers for LF transmissions and to several meters for UHF transmissions. Thus, much of the population exposure of interest occurs in the near field of the LF antenna and in the far field of the UHF antenna. Population exposures to UHF fields can be completely specified by the UHF electric field, by the magnetic field, or by power density alone. Descriptions of most LF exposures, however, need to include both the electric and magnetic fields. COUPLING OF ELECTROMAGNETIC FIELDS TO HUMAN BODY To the extent that nonionizing electric and magnetic fields have biological effects, the effects will depend on the fields induced in the body. In general, fields produced in the body are not the same as fields that one would measure in air at the same location if the body were absent. For that reason, nonionizing electromagnetic-field (EMF) exposures are usually estimated in two parts. First, the “ incident” fields are calculated or measured. These are the electric and magnetic fields that are produced in the absence of distorting influences, such as a human body. Second, various dosimetric principles are used to relate incident fields to fields that are induced in the body6 The relationship between incident and internal

OCR for page 143
ASSESSMENT OF THE POSSIBLE HEALTH EFFECTS OF GROUND WAVE EMERGENCY NETWORK fields depends on frequency, orientation of the body relative to the field, body size, and the impedance of the body to ground. Comparisons of the intensity of exposure produced by various sources must account for any differences in those factors. For instance, UHF fields couple much more strongly to the body than do LF fields. A vertically polarized electric field at ultrahigh frequencies induces an internal electric field in an ungrounded standing adult that is 300 times that induced by an LF electric field of the same incident strength and polarization.6 In many areas outside the boundaries of GWEN sites, LF-field strengths at ground level can be expected to exceed UHF-field strengths by less than a factor of 300, so that the internal electric fields from a GWEN site will be predominantly from the UHF component. The incident LF and UHF fields of the GWEN system induce electric fields, magnetic fields, and currents in the bodies of exposed people. The strength of this coupling is described in Chapter 3 for a “worst case,” defined as a grounded person standing near the 4-ft perimeter fence of a GWEN site. Dosimetric calculations show that body currents induced by electric-field coupling greatly exceed those induced by magnetic-field coupling for GWEN exposures. Later analyses of LF exposure and risk in this chapter therefore focus on the electric field. SHIELDING BY BUILDINGS Buildings and trees attenuate the electric-field component of LF fields and both the electric- and magnetic-field components of UHF fields. Measurements suggest that houses shield LF electric fields by 12-30 dB (75-97%), UHF electric fields by 0-20 dB (0-90%), and UHF magnetic fields by 0-10 dB (0-68%).7 People spend much of their day indoors, so the shielding can have a significant influence on exposure from sources outside the home. POPULATION DISTRIBUTION AROUND GWEN SITES If there is a public-health impact of exposure to GWEN fields, it will be proportional to the number of people exposed. The population density around candidate locations at 40 RN sites was estimated by surveying U.S. Geological Survey topographical maps of the areas8 Survey locations were chosen randomly from the list of candidate locations for each site, unless a specific location had already been specified. Structures not specifically identified as schools, churches, or factories were counted as homes. Structures were counted in each of six contiguous concentric regions around the site, with outer radii of 200, 300, 500, 1,000, 2,000, and 4,000 m. The distance from the center of each site to the nearest house was noted. The raw data are listed in Table 2-3 and presented in Figure 11-1, Figure 11-2, and Figure 11-3. The data show that the density of houses within several kilometers of GWEN RN sites is typically about one house/km2. That is sparse, compared with the average of 11 households/km2 for the conterminous 48 states as a whole, and is about the same as the

OCR for page 143
ASSESSMENT OF THE POSSIBLE HEALTH EFFECTS OF GROUND WAVE EMERGENCY NETWORK average house density in Montana and the Dakotas. None of the 40 sites examined had a housing density exceeding 10/km2. The typical distance from site center to the nearest house is 330 m. Only a few sites have houses within 150 m. Several sources of bias are possible in these data. First, counting all structures as houses except those known to be something else will tend to bias housing densities upward. In contrast, the topographical maps used for this study had an average age of 15 yr (standard deviation, 8.7 yr); given general trends of increasing population density, this will tend to bias estimated housing densities downward. Finally, we do not know the extent to which the 40 RN sites examined are representative of the entire GWEN system. FIGURE 11-1. Cumulative distribution of number of homes in six contiguous concentric regions around sample of 40 RN sites.

OCR for page 143
ASSESSMENT OF THE POSSIBLE HEALTH EFFECTS OF GROUND WAVE EMERGENCY NETWORK FIGURE 11-2. Cumulative distribution of home density in six contiguous concentric regions around sample of 40 RN sites. FIGURE 11-3. Cumulative distribution of distance of center of 40 RN sites to nearest home.

OCR for page 143
ASSESSMENT OF THE POSSIBLE HEALTH EFFECTS OF GROUND WAVE EMERGENCY NETWORK LF AND UHF EXPOSURES OF POPULATION AROUND SITES Rough estimates of the numbers of people living in homes exposed to GWEN fields above various intensities can be derived from the population and field-strength data described above. Estimates are presented in Table 11-1. Estimates for RN sites are derived by assuming all structures identified in the survey of topographical maps are occupied at the national average rate of 2.6 persons/household.9 Estimates for the total GWEN system assume that existing thin-line connectivity capability (TLCC) relay nodes are similar to GWEN RNs. UHF exposures are assumed to occur at antenna height (10 m). At full RN implementation, time-averaged field strengths will be about 0.014 of peak values for LF and 0.019-0.027 of peak values for UHF. TABLE 11-1. Approximate Number of People Living in Homes Exposed to GWEN Fields of Various Peak Intensities. Peak Field Strength 40 RN Sites Total GWEN System 0.02 V/m < UHF < 0.05 V/m 17,000 40,000 0.05 V/m < UHF < 0.1 V/m 2,000 4,700 0.1 V/m < UHF < 0.2 V/m 420 1,000 0.2 V/m < UHF < 0.5 V/m 80 200 0.5 V/m < UHF 0 0 0.1 V/m < LF < 0.2 V/m 11,000 28,000 0.2 V/m < LF < 0.5 V/m 3,300 7,800 0.5 V/m < LF < 1.0 V/m 400 1,000 1.0 V/m < LF < 2.0 V/m 80 200 2.0 V/m < LF 30 80 5 µG < LF < 10 µG 6,600 15,000 10 µG < LF < 20 µG 1,300 3,100 20 µG < LF < 50 µG 300 800 50 µG < LF < 100 µG 50 110 100 µG < LF < 200 µG 15 40 200 µG < LF 25 55 The persons exposed to the strongest GWEN fields continuously are, of course, those who live closest to the sites. Someone living at a site boundary would be exposed to GWEN LF fields at ground level of up to 40 V/m and 0.7 mG.

OCR for page 143
ASSESSMENT OF THE POSSIBLE HEALTH EFFECTS OF GROUND WAVE EMERGENCY NETWORK Plans for GWEN RNs call for 10-m-high UHF antennae to be near the corners of GWEN sites. A nearby house could theoretically be as close as 30 m from the base of the UHF antenna. That would expose residents on the second floor of a two-story house to power densities up to 1 µW/cm2. Exposures at ground level would be somewhat smaller. EXPOSURE COMPARISONS WITH EXISTING STANDARDS National and international scientific bodies have developed exposure guidelines for radiofrequencies (RFs), including the LF and UHF bands; these are reviewed in Chapter 10. The strongest fields that might be encountered at the boundary of a GWEN facility are LF fields of 40 V/m and 0.7 mG and UHF fields with power densities of 1 µW/cm2. Those are well below all existing guidelines for LF magnetic fields and UHF power densities. Although peak LF electric-field exposures at the fenceline are close to both International Radiation Protection Association and USSR standards, time-averaged exposures are well below the guidelines. EXPOSURE COMPARISONS WITH OTHER SOURCES The most prominent sources of population exposure to RF radiation at or near GWEN frequencies are UHF television (470-806 MHz) and AM radio (535-1610 kHz) broadcasts. Environmental Protection Agency (EPA) measurements of RF radiation in 15 U.S. cities10 enable comparisons of GWEN exposures to exposures from those broadcast sources. Figure 11-4 shows the numbers of people exposed above a given field level for both the 15-city urban population of the EPA study and populations near GWEN facilities. On the basis of the time-averaged electric field induced in body tissues, LF exposures of someone living at the boundary of a GWEN RN are comparable with the median exposure to AM-broadcast radiation in urban areas. A second source of LF frequencies is the LORAN navigation system. The LORAN system consists of 20 transmitters along U.S. coastlines broadcasting pulse trains on a carrier wave of 150-170 kHz. LORAN stations are generally in rural areas and operate at peak powers of 275-1,600 kW, with a typical time-averaged emission of roughly 15 kW. By comparison, GWEN LF transmissions have a peak power of 2.0-3.2 kW, with typical time-averaged emissions of 30 W. In countries other than the United States, a substantial source of exposure to LF fields is AM-radio broadcasts. These are made at frequencies of 150-175 kHz in France, Germany, Mongolia, Morocco, Turkey, and the USSR at 150-2,000 kW.11 Table 11-2 compares GWEN and other LF and UHF sources on the basis of broadcast power and number of emitters in the United States. GWEN sites are fewer, lower in power, and in more sparsely populated areas than many other broadcast sources at comparable

OCR for page 143
ASSESSMENT OF THE POSSIBLE HEALTH EFFECTS OF GROUND WAVE EMERGENCY NETWORK frequencies. The data in Table 11-2 and Figure 11-4 show that population exposures to GWEN fields are modest compared with those associated with broadcast sources in nearby bands. TABLE 11-2. RF Power and Frequencies of GWEN and Other Sources of Radiofrequency Radiation Source Frequency, MHz Range of RF Power Across Sources, kW Time-averaged RF Power of Typical Source, kW No. U.S. sites GWEN LF 0.150-0.175 2-3 0.028 127 LORAN LF 0.150-0.170 275-1600 15 15 AM radio 0.525-1.610 0.25-50 5 4972 GWEN UHF 225-400 0.020 0.00066 139a Cellular-phone base station 800-900 0.040-4 0.4 > 5,000 UHF TV 470-806 5-5,000 350 769 aDoes not include 34 airborne 1/O terminals. Source: See reference 10, reference 12, reference 13 and reference 14. Comparisons can also be made between exposures to GWEN fields and exposures to less similar sources of nonionizing fields. Figure 11-5 compares exposures at the perimeter of a GWEN site with other exposure situations involving sources in the extremely-low-frequency (ELF), very-high-frequency (VHF), and superhigh-frequency (SHF) bands. The situations include living next to a 500-kV transmission line, standing beneath a neighborhood distribution line, living in an urban area served by FM radio, using a cordless telephone, and standing 1 m away from a microwave oven. Comparisons are based on four measures of exposure: time- and body-averaged magnetic flux density, time-and body-averaged power absorption, instantaneous peak power absorption in any tissue, and fraction of time exposed. The data in Figure 11-5 demonstrate how sensitive exposure comparisons are to the choice of exposure measure. For sources that vary widely in frequency, for instance, comparisons based on power absorption are very different from those based on incident field strength, because power absorption is proportional to the square of the frequency.

OCR for page 143
ASSESSMENT OF THE POSSIBLE HEALTH EFFECTS OF GROUND WAVE EMERGENCY NETWORK FIGURE 11-4. Cumulative distributions of people exposed to GWEN and commercial broadcast fields above given electric-field strength. Broadcast exposures are derived by combining census data with EPA survey of RF-field exposures in urban areas.9, 10 GWEN data are taken from Table 11-1. Curves for GWEN system assume 100% duty cycle. Time averaging based on expected GWEN duty cycles would shift GWEN curves to left by factors of 70 and 50 for LF and UHF, respectively.

OCR for page 143
ASSESSMENT OF THE POSSIBLE HEALTH EFFECTS OF GROUND WAVE EMERGENCY NETWORK FIGURE 11-5. Comparison of exposures to GWEN fields at site boundary with five common exposure situations, based on four alternative exposure measures. Situations include living on edge of 500-kV right-of-way, standing beneath neighborhood distribution line, being exposed to ambient FM-broadcast radiation in urban areas, using cordless phone, and using microwave oven. SAR, specific absorption rate.

OCR for page 143
ASSESSMENT OF THE POSSIBLE HEALTH EFFECTS OF GROUND WAVE EMERGENCY NETWORK TABLE 11-3. 30-Year Trends in Age-Adjusted Death Rates per 100,000 Population for Cancers that have been Associated with Electromagnetic-Field Exposure.   Age-Adjusted Death Rate per 100,000 Population   Cancer Sex 1955-1957 1985-1987 Change, % Breast Male 0.3 0.27 −11   Female 26.3 27.2 +3 Brain Male 4.0 4.9 + 22   Female 2.7 3.3 +22 Leukemia Male 8.5 8.2 −4   Female 5.6 4.9 −12 Non-Hodgkins lymphoma Male 4.7 7.1 + 51   Female 3.0 4.8 +60 Source: Cancer Facts and Figures - 1991, American Cancer Society,Atlanta, GA, 1991. TABLE 11-4. 14-Year Trends in Age-Adjusted Incidence per 100,000 Population for Cancers Possibly Associated with Electromagnetic-Field Exposure.   Age-Adjusted Incidence Rate per 100,000 Population   Cancer 1973 1986 Change, % Breast 44.8 57.6 +29 Brain and nervous system 5.0 6.1 + 22 Leukemia 10.5 9.3 −11 Non-Hodgkin's lymphoma 8.5 12.6 + 48 Source: Cancer Statistics Review 1973-1986, National Cancer Institute,May 1989.

OCR for page 143
ASSESSMENT OF THE POSSIBLE HEALTH EFFECTS OF GROUND WAVE EMERGENCY NETWORK PUBLIC HEALTH SURVEILLANCE AROUND BROADCAST FACILITIES Powerful commercial and navigation broadcast stations have been operating for many years at or near GWEN LF and UHF frequencies both in the United States and elsewhere11, 12 (see Table 11-2). Yet public health monitoring of spatial patterns of disease has identified no correlation of regional morbidity or mortality with distance from a broadcast antenna. That no such evidence has emerged suggests that, whatever the public health risks related to broadcast exposure might be, they are smaller than the detection threshold of the health surveillance apparatus. This observation can be used to place an upper bound on the possible health impacts of the GWEN system. LF Broadcasting in Europe. Inquiries were made on behalf of the committee to the operators of several LF-broadcast facilities in Europe. The results, shown in Table 11-5, suggest that any possible health effects of LF fields are not large enough to have been spontaneously noticed in populations of workers and residents near the facilities. On a time-averaged basis, the electric fields to which the European broadcast facilities expose populations are up to 10 times stronger than the LF electric field at the boundary of a GWEN RN site. LORAN Stations. The 15 LORAN-C navigation transmitters in the United States operate at LF frequencies and in a pulsed mode similar to that of GWEN LF transmitters.13 LORAN stations tend to be in areas of low to moderate population density, and there is no evidence of excess cancer risk associated with the fields from these facilities. MEADOWLANDS SPORTS COMPLEX The Meadowlands Sports Complex in East Rutherford, N J, is very near a number of AM radio broadcast antennae.14 A recent study by Kraut and colleagues of cancer incidence and cancer deaths in a population of almost 8,000 workers at the sports complex found no significant differences in cancer risk between worker and reference populations over the period 1978-1987.15 Electric fields measured at outdoor locations across the complex ranged from 1 to 16 V/m, with typical values of 3-10 V/m.16 Indoors, electric fields were much smaller, typically less than 0.3 V/m. No significant differences in cancer risk were noted between populations of 2,500 outdoor workers and 5,500 indoor workers employed over that limit the relevance of this study to short-latency cancers such as leukemia. The average employment duration for the Meadowlands cohort was roughly 5 yr. Only two leukemia cases and no leukemia details were identified in the entire 8,000 worker cohort. Both leukemia diagnoses were among indoor workers (Personal communication with Dr. Allen Kraut, University of Manitoba, Winnipeg, October 1, 1992). Among outdoor workers, 1.11 leukemia cases would have been expected in a

OCR for page 143
ASSESSMENT OF THE POSSIBLE HEALTH EFFECTS OF GROUND WAVE EMERGENCY NETWORK TABLE 11-5. Results of Inquiries to LF-Broadcast Facilities in Europe. Location Power kW Frequency kHz Period of operation Population Electric Field exposure V/m Beidweiler, Luxembourg 10,000 234 1972-1991 300 villagers 18 Junglinster, Luxembourg 200-1,200 234 1932-1972 1,500 villagers 7-16 Allouis, France 2,000 164 1961-1991 station up to 1,000 workers 16 Oslo, Norway 200 216 1954-1991 nearest residents in scattered farmhouses < 16 Tromso, Norway 10 154 1926-1991 airport town population 3 cohort of that size. The corresponding upper bound on the 95% confidence interval for proportionate incidence ratio for outdoor workers is about 3.3. This upper bound corresponds to an excess morbidity of roughly 3 chances in 10,000 per person per year. Since the 5 yr survival rate for adult leukemias is only about 30%, these morbidity bounds represent an upper bound on mortality risk of roughly 2 x 10−4 per person per year. BOUNDS ON EXCESS POPULATION RISK FROM GWEN FIELDS The upper bound on risk of excess cancer mortality associated with operation of GWEN LF and UHF transmitters can be estimated from four factors: (1) information on health surveillance of census tracts surrounding AM broadcast facilities; (2) relative exposures to RF fields of populations near AM broadcast stations and GWEN facilities; (3) population density around GWEN facilities; and (4) the magnitude of GWEN fields as a function of distance from the transmitter. In this calculation, it is assumed that any elevation in health risk is associated with absorbed energy from GWEN fields, and hence

OCR for page 143
ASSESSMENT OF THE POSSIBLE HEALTH EFFECTS OF GROUND WAVE EMERGENCY NETWORK is proportional to the square of the electric field intensity in air. The effects o the calculation of choosing an alternative exposure metric, namely, the electric field in air, is also discussed. In our approach, an upper bound estimate of the health risks associated with GWEN LF radiation is computed from information on the absence of public health effects associated with exposure to AM broadcast fields. Public health statistics are regularly reviewed at the level of the census tract, which can be smaller than a city block in urban areas and larger than a county in sparsely populated areas. On the average, there are about 70 census tracts for every AM radio station in the United States with an urban location. The spatial resolution of health surveillance is therefore fine enough in many areas to detect any strong risk gradients around AM broadcast facilities. Health surveillance studies to date have not indicated the existence of risk gradients around broadcast facilities for the cancers which previous epidemiological studies have associated with EMF exposure. These cancers include leukemias, non-Hodgkins lymphomas, nervous system tumors, and breast cancers, which collectively impose an average background mortality risk of 3 x 10-4 per person per year in the United States. For purposes of risk estimation, let k be the smallest excess relative risk that would be detected by the health surveillance system and consider a continuously broadcasting AM station that creates a spatially-averaged electric field Eam in an adjacent census tract. If Ro is the background risk of cancer death in the region, then the largest EMF effect that could go unnoticed in the census tract is a risk of kRo. That no such effects have so far been established has implications for EMF risks from GWEN facilities that can be quantified by considering the relative exposures of populations around AM broadcast stations and GWEN RNs. If we assume that the risk is proportional to RF energy absorbed in tissue (and hence proportional to the square of the electric field intensity), then the largest excess risk, Rg, from exposure to peak GWEN LF fields of strength Eg that is consistent with the lack of evidence for populations living near AM broadcast facilities is: Rg(r) < kRoγ2 δ Eg(r)2/Eam2 (11-1) where r is distance from a GWEN LF antenna in km, γ is the ratio of electric field coupling to the body at GWEN LF frequencies compared to that at AM frequencies, and δ is the duty cycle of GWEN LF emissions. The AM broadcasting is assumed to be continuous. This expression does not account for electromagnetic shielding from houses or other objects, nor does it account for time spent away from home (i.e., when exposures may not occur). We assume that these two factors have equal influence on risk from commercial AM broadcast fields and GWEN fields, and therefore cancel out of Equation (11-1).

OCR for page 143
ASSESSMENT OF THE POSSIBLE HEALTH EFFECTS OF GROUND WAVE EMERGENCY NETWORK To first order, the average electric field (V/m) in a census tract of radius C directly adjacent to an AM antenna of power P (watts) with a surrounding exclusion zone of radius ε (meters) is: Eam = σP1/2/(C + ε) (11-2) where σ is a factor in the range 4-10 ohm1/2 that depends on ground conductivity, antenna height, and wavelength. Combining Equations (11-1) and (11-2) yields Equation (11-3) for the largest estimate of cancer death from exposure to GWEN fields that cannot be ruled out from health surveillance data: Rg(r) < kRoγ2 δ Eg(r)2(C + ε)2/(σ2P) (11-3) If ρ is the average population density in the region in persons/km2, then the upper bound on excess population risk for a single GWEN facility for those living between r1 and r2 kilometers from the antenna is: Equation (11-4) was evaluated using typical values of all parameters to estimate the total population risk from a single GWEN facility out to a distance of 10 km. Beyond this distance, GWEN fields are negligible relative to typical AM broadcast exposures in urban areas. The values assigned to the parameters in Equations (11-3) and (11-4) are as follows: k = 1. This value of k assumes that a doubling of EMF-related cancer mortality in census tracts around a typical AM broadcast antenna would have been detected, at least at one site. Ro = 3 x 10−4 per person per year. This risk estimate includes only those cancers that the epidemiological literature has suggested might be associated with EMFs, namely, leukemias, non-Hokgkins lymphomas, nervous tissue tumors, and breast cancers. γ = 0.22. This value is the ratio of electric field coupling of GWEN LF fields to that of AM broadcast fields at the center frequency of each band. δ = 0.014. This value is the best estimate for the typical duty cycle of GWEN LF emissions at full peacetime implementation.

OCR for page 143
ASSESSMENT OF THE POSSIBLE HEALTH EFFECTS OF GROUND WAVE EMERGENCY NETWORK Out to a distance of 0.3 km from a GWEN LF antenna, the electric field beyond the boundary fence (at 0.1 km) is approximately: Eg(r) = 0.04/r3 (0.1 km < r < 0.3 km) (11-5) where Eg(r) is the electric field intensity in V/m and r is the ground-level distance from the antenna in kilometers. Beyond 0.3 km from the LF antenna, the electric field is approximately: Eg(r) = 0.5/r (0.3 km < r < 10 km) (11-6) C = 150 m. This value of the census tract radius is typical of larger cities. ε = 100 m. The exclusion zone is assumed to be as large as a typical AM antenna height. σ = 7 ohm1/2, an average value for this parameter. P = 10,000 watts, the power of a typical AM station. ρ = 5 persons/km2. Data from the survey of aerial photographs of prospective sites for GWEN RNs indicate that the average density of homes is less than 2 houses/km2. Assuming the national average of 2.6 persons per household, the average population density around GWEN RNs would be about 5 persons/km 2. Using these values in Equations (11-3) and (11-4) yields an upper bound estimate of population risk for a single GWEN site of 3.6 x 10-6 excess cancer cases per year. During a 70-yr lifetime, the upper bound risk estimate would be 2.5 x 10-4 cases per site. For the full complement of 125 GWEN sites, the total upper bound estimate on the number of possible excess cancer cases is 0.03 over a 70-yr lifetime. It should be noted that the expected lifetime of the GWEN system is only 15 yrs, so that this upper bound risk estimate is probably high by a factor of about 5. A completely analogous set of calculations can be made using the electric field as an exposure metric. In this case, Equation (11-3) becomes Rg(r) < kRo γ δ Eg(r) (C + ε)/(σP1/2) (11-7)

OCR for page 143
ASSESSMENT OF THE POSSIBLE HEALTH EFFECTS OF GROUND WAVE EMERGENCY NETWORK Combining Equations (11-4) and (11-7) and using the same parameter values as those listed above yields an upper bound estimate for population risk at a single GWEN site of 5.4 x 10-5 cases per year. For a 70-yr lifetime and the full complement of 125 GWEN sites, the total upper bound estimate on the number of possible excess cancer cases is 0.5. Similar arguments can be made for the UHF emissions from the GWEN system. Again, less than one excess cancer death is predicted for the entire population living in the proximity of GWEN UHF transmitters for 70 yrs. The estimated maximum increase in cancer risk associated with GWEN fields is therefore well below the detection capability of the epidemiological survey system. LIMITATIONS OF GWEN RISK ASSESSMENT The approach taken in estimating cancer risk from exposure to GWEN fields was based on a bounding argument using negative data from public exposure to fields of similar frequency emitted by broadcast stations. In this approach the committee has estimated the magnitude of risks from GWEN fields in a manner that is consistent with the observation that existing RF broadcast facilities have had no apparent effect on the cancer risk of surrounding populations. Because of the lack of positive data on cancer risk and the resulting uncertainties in the parameters used for our calculation, the true risk is likely to be smaller than the upper bound on risk derived by the approach taken here. Thus, while the committee has provided an upper bound estimate of the health risks from GWEN fields, it should be recognized that this value is likely to be an overestimate of the true risk and should be used in a policy sense only to exclude actions that might be warranted by risks greater than the upper bound estimated here. RISK PERCEPTION Science can provide evidence, albeit uncertain, on how large the potential risks related to GWEN facilities might be. But the acceptability of those risks is a value judgment, not a technical issue. Evidence from social psychology shows that the depth of public concern about a real or potential public-health risk depends on many factors other than morbidity and mortality. They include the plausibility, familiarity, and fear of the hazard; the population distribution of risks and benefits of the technology in question; the fairness of the process by which the risk is imposed; and the credibility of the group charged with managing the risk.17 A number of nonrisk factors might pertain to the perception of risks associated with the GWEN system: Recent studies of the public perception of the risks associated with nonionizing fields show that the lay public believes that health risks related to such fields are plausible.18

OCR for page 143
ASSESSMENT OF THE POSSIBLE HEALTH EFFECTS OF GROUND WAVE EMERGENCY NETWORK Studies of the qualitative attributes of the risk related to ELF fields from high-voltage transmission lines show that, although people rank transmission lines as among the least risky of 16 common known or potential hazards, they nonetheless view transmission-line risks as less familiar and more dreaded than many other risks19, 20 and therefore perceive a greater need for regulation. If GWEN risks are similarly perceived, the public might view them as less acceptable than commonly accepted hazards that have higher actual risks. We do not know whether nonthermal levels of nonionizing radiation present a significant risk. If they do, however, it seems likely that the public-health impact of the GWEN system will be negligible, compared with that of UHF-TV and AM-radio broadcasts. But, because the acceptability of risks depends, in part, on the distribution of benefits from an activity,21 GWEN risks might be judged less acceptable than those related to broadcast sources. That is because urban populations exposed to UHF and AM broadcast signals derive direct benefits (news and entertainment) from those activities, whereas the risks associated with the GWEN system are encountered by the few dozen households in the vicinity of each site and the benefits of GWEN are both indirect and spread over the entire U.S. population. Data from polls conducted over the last 25 yr have shown a steady decline in the trust that the public places in government and in risk management institutions.22, 23 This lack of trust is likely to affect the public's perception of the safety of the GWEN system. EXPOSURE REDUCTION Available evidence precludes assurances that exposures to GWEN emissions are without risk. In fact, no amount of scientific research can guarantee such assurances for any agent. The goals of risk management are to identify rationales and fair processes for balancing residual risk against the costs of mitigative actions for reducing that risk. Options for Reducing Exposures to GWEN Fields. A number of actions might be taken to reduce population exposures to GWEN LF and UHF emissions. They include choosing sites for the RNs that avoid people, placing the UHF transmitting antenna in each site so as to minimize UHF exposures of those living nearby, placing the UHF transmitting antennae on higher poles, increasing the size of each GWEN RN site to reduce field strengths at the site boundary, and reducing the number of RNs in the final GWEN system. Each of those has costs--some major, some minor. Implications for Other Broadcast Activities. Because their risks and benefits are distributed differently, the public acceptability of risks from GWEN facilities is likely to be smaller than that from radio and television broadcasting. Taking steps to mitigate exposures from GWEN facilities need not imply, therefore, that mitigation should be applied to radio and television broadcasting as well.

OCR for page 143
ASSESSMENT OF THE POSSIBLE HEALTH EFFECTS OF GROUND WAVE EMERGENCY NETWORK RESEARCH NEEDS Given the extent of population exposure to broadcast sources in the United States, the lack of epidemiological data relevant to the question of RF exposure and human cancer is surprising. The economic costs surrounding delays in the siting and construction of GWEN and other new broadcast sources are difficult to quantify, but they undoubtedly far exceed the few millions of dollars per year that the United States currently spends on RF biological effects research. Additional federally supported RF-effects research would not only further our scientific understanding of the interactions of EMFs with biological systems, but also provide a basis on which the public could generate informed opinions concerning possible health risks associated with RF emitters.

OCR for page 143
ASSESSMENT OF THE POSSIBLE HEALTH EFFECTS OF GROUND WAVE EMERGENCY NETWORK REFERENCES 1. Hattis, D., and D. Kennedy. 1986. Assessing risks for health hazards: an imperfect science. Technology Review, Vol. 89. Pp. 60-71. 2. National Research Council, Committee on the Institutional Means for Assessment of Risks to Public Health. 1983. Risk Assessment in the Federal Government: Managing the Process. Washington, D.C. : National Academy Press. 3. Bawin, S. M., and W. R. Adey. 1976. Sensitivity of calcium binding in cerebral tissue to weak environmental electric fields oscillating at low frequency. Proc. Natl. Acad. Sci. (USA) 73 : 1999-2003. 4. Blackman, C. F., et al. 1979. Induction of calcium-ion efflux from brain tissue by radio-frequency radiation: effects of modulation frequency and field strength. Radio Science 14(6S) : 93-98. 5. Schwartz, J., D. E. House, and G. A. R. Mealing. 1990. Exposure of frog hearts to CW or amplitude modulated VHF fields: selective efflux of calcium ions at 16 Hz. Bioelectromagnetics 11 : 349-358. 6. Durney, C. H., H. Massoudi, and M. F. Iskander. 1986. Radiofrequency radiation dosimetry handbook. Report USAFSAM-TR-85-73. University of Utah for the United States Air Force School of Aerospace Medicine. 7. Smith, A. A. 1978. Attenuation of electric and magnetic fields by buildings. IEEE Transactions on Electromagnetic Compatibility EMC 20 : 411-418. 8. United States Air Force (USAF). 1991. Procedures and data for the survey of housing density around GWEN sites are available from Lt. Col. Stephen T. Martin, Headquarters, Electronic Systems Division, Hanscom Air Force Base, Massachusetts. 9. United States Bureau of the Census. 1990. Statistical Abstract of the United States:110th edition. Washington, D.C. 10. Hankin, N. N. 1986. The radiofrequency radiation environment: environmental exposure levels and RF radiation emitting sources. Report EPA 520/1-85-014, July. Washington, D.C. : U.S. Environmental Protection Agency. 11. World Radio and TV Handbook. 1987.

OCR for page 143
ASSESSMENT OF THE POSSIBLE HEALTH EFFECTS OF GROUND WAVE EMERGENCY NETWORK 12. Taishoff, S. 1990. The Broadcasting Yearbook, S. Taishoff, ed. New York, NY : Broadcasting Publications Inc. 13. United States Coast Guard. 1981. Specification of the Transmitted Loran-C Signal. COMDTINST M16562. July 4. 14. United States Federal Communications Commission. 1991. Personal communication with M. Ferrante, Mobile Services Division, December 13. 15. Kraut, A., E. Chan, P. J. Lioy, F. B. Cohen, B. D. Goldstein, and P. J. Landrigan. 1991. Epidemiologic investigations of a cancer cluster in professional football players. Environ. Res. 56 : 131-143. 16. Tell, R. A., and W. Van Pelt. 1988. An investigation of radiofrequency field strengths in the vicinity of the meadowlands sports complex, East Rutherford, NJ. Prepared for the NJ Sports and Exposition Authority, July 29. Las Vegas, NV : R. E. Tell Associates. 17. Hohenemser, C., and R.E. Kasperson. 1982. Risk in the Technological Society. Boulder, CO : Westview Press. 18. Morgan, M. G., H. K. Florig, I. Nair, C. Cortes, K. Marsh, and K. Pavlosky. 1990. Lay understanding of low-frequency electric and magnetic fields. Bioelectromagnetics 11(4) : 313-335. 19. Slovic, P., B. Fischhoff, and S. Lichtenstein. 1980. Facts and fears: understanding perceived risk. Pp. 181-216 in Societal Risk Assessment: How Safe is Safe Enough? R. C. Schwing and W. A. Albers, Jr., eds. New York, NY : Plenum Press. 20. Morgan, M. G., P. Slovic, I. Nair, D. Geisler, D. MacGregor, B. Fischhoff, D. Lincoln, and K. Florig. 1985. Powerline frequency electric and magnetic fields: a pilot study of risk perception. Risk Analysis 5(2) : 139-149. 21. Slovic, P. 1987. Perception of risk. Science 236 : 280-285. 22. Lipset, S. M., and W. Schneider. 1983. The decline of confidence in American in American institutions. Political Science Quarterly 98(3) : 379-402. 23. Renn, D., and D. Levine. 1991. Credibility and trust in risk communication. Pp. 175-218 in Communicating Risks to the Public, R. E. Kasperson and P. J. M. Stallen, eds. Boston, MA : Kluwer Academic Publishers.