effect on air quality because high concentrations often occurred in small systems with low flow rates, which yielded lower overall emissions. The evaluation concluded that the resulting health risk posed by radon release into the atmosphere via aeration-system off-gas was much smaller (by a factor of about 100 to 10,000) than the risks that would result if radon were not removed from the water.
The EPA Science Advisory Board (SAB) reviewed EPA's report (EPA 1989; EPA 1988b) and found that the uncertainty analysis needed to be upgraded to lend more scientific credibility to the air-emissions risk assessment. However, the SAB also stated that revisions in the modeling would not change EPA's conclusion that the risk posed by release of radon from a water-treatment facility would be no more than the risk posed by using drinking water that contains radon at 11,000 Bq m-3. The SAB also noted that EPA's assumptions were conservative.
EPA also had its Office of Radiation and Indoor Air (ORIA) review its 1988 (EPA 1988b) and 1989 (EPA 1989) air-emissions studies for consistency and to provide a simple quantitative uncertainty analysis (EPA 1994b). The ORIA review indicated that the early studies had overstated the risk; it estimated an incidence of cancer of 0.004 cases per year, less than the 0.016 case per year initially estimated.
EPA (1991b) acknowledged in the proposed rule for radon that ''some states allow no emissions from PTA systems, regardless of the downwind risks.'' Indeed, on the federal level, EPA has, under the Clean Air Act, established National Emission Standards for Hazardous Air Pollutants (NESHAPs), including radon. Under NESHAP, an average radon-emission rate of 0.74 Bq m-2 s-1 is allowed from radium-containing facilities, and an individual member of the public can have a maximal exposure of 1.00 × 10-4 Gy y-1. It is not clear whether NESHAPs will ever affect aeration systems, inasmuch as they are applicable only to industries, not to drinking-water treatment facilities. Drago (1998) has noted that if the radon NESHAP were applied to a 93 m3 d-1 water-treatment plant (serving about 250 people) with an influent radon concentration of 18,500 Bq m-3 in its water, radon would be emitted in the off-gas at 222 Bq m-2 s-1. This analysis suggests that if the NESHAP were extended to include drinking-water plants, many aeration systems would have to treat the off-gas to remove radon. The Nuclear Regulatory Commission also regulates radon releases from the facilities that it licenses, but it is doubtful that its limits would ever apply to water-treatment facilities. Some states have adopted NESHAPs for radionuclides or Nuclear Regulatory Commission limits by agreement, but none directly applies them to water-treatment plants.
It is possible that some states may conduct or require risk screening of new water treatment aeration facilities for radon emissions. Although some states allow specific incremental lifetime risks associated with hazardous air pollutants (for example, California one in 1 million), it is not clear that they will be applied to radon. Drago (1998) noted that only Nevada, California, New Jersey, and