We assess several of the important health and environmental risks associated with a reference geothermal industry that produces 21,000 MWe for 30 y (equivalent to 20 x 10{sup 18} J). The analyses of health effects focus on the risks associated with exposure to hydrogen sulfide, particulate sulfate, benzene, mercury, and radon in air and arsenic in food. Results indicate that emissions of hydrogen sulfide are likely to cause odor-related problems in 29 of 51 geothermal resources areas, assuming that no pollution controls are employed. Our best estimates and ranges of uncertainty for the health risks of chronic population exposures to atmospheric pollutants are as follows (risks expressed per 10{sup 18} J of electricity): particulate sulfate, 44 premature deaths (uncertainty range of 0 to 360); benzene, 0.15 leukemias (range of 0 to 0.51); elemental mercury, 14 muscle tremors (range of 0 to 39); and radon, 0.68 lung cancers (range of 0 to 1.8). The ultimate risk of fatal skin cancers as the result of the transfer of waste arsenic to the general population over geologic time ({approx} 100,000 y) was calculated as 41 per 10{sup 18} J. We based our estimates of occupational health effects on rates of accidental deaths together with data on occupational diseases and injuries in surrogate industries. According to our best estimates, there would be 8 accidental deaths per 10{sup 18} J of electricity, 300 cases of occupational diseases per 10{sup 18} J, and 3400 occupational injuries per 10{sup 18}J. The analysis of the effects of noncondensing gases on vegetation showed that ambient concentrations of hydrogen sulfide and carbon dioxide are more likely to enhance rather than inhibit the growth of plants. We also studied the possible consequences of accidental releases of geothermal fluids and concluded that probably less than 5 ha of land would be affected by such releases during the production of 20 x 10{sup 18} J of electricity. Boron emitted from cooling towers in the Imperial Valley was identified as a potential source of crop damage. Our analyses, however, showed that such damage is unlikely. Finally, we examined the nonpollutant effects of land subsidence and induced seismicity. Land subsidence is possible around some facilities, but surface-related damage is not expected to be great. Induced seismic events that have occurred to date at geothermal resource areas have been nondestructive. It is not possible to predict accurately the risk of potentially destructive events, and more research is needed in this area.
