in the range of 1–63 Bq m-3, with the exception of Maine. Both Iowa and Maine had higher average outdoor radon than the other areas, and the geometric mean for Iowa was significantly higher than all the others.
In the original charge to the committee, EPA requested "central estimates for a population-weighted average national ambient concentration for radon, with an uncertainty range. Comparisons of the contribution of radon in water to other sources of indoor radon will be made and comparisons will be made to outdoor levels." The charge was amended to include a discussion of alternatives to population-weighted averages and of spatial and temporal variation. The ambient-radon data of Hopper and others (1991) are the only data that provide some portion of national coverage over an extended period, but the committee has concerns about the appropriateness of using these data to develop a population-weighted average for the United States. Hopper and others contend that the data cannot be used to represent the ambient radon of the state that contains the sampling site. However, they do think that the ambient radon measured at each site is representative of that site. Using population data from the 1990 census, one can calculate a population weighted average (APW) by summing the products of each site's seasonal average (Ai) and the city population (Pi) and then dividing by the sum of the city populations.
That assumes that the ambient radon measured at the site is representative of the ambient radon of the city. The population-weighted average radon concentration is given by:
where N is the number of sites. The population-weighted average radon concentration is 14.0 Bq m-3. The total population of the cities for all the sites in the national outdoor radon survey was about 24 million, or slightly less than 10% of the US population. The calculation is dominated by New York City, which has a large population and a lower than average ambient radon measurement, and therefore the population weighted average is less than the unweighted average of 14.8 Bq m-3.Further, the sites were not chosen to be a statistical representation of the population across the United States, nor were the sites chosen to be representative of ambient radon within each state or sample the geology of the country. The overall distribution of the seasonal data for each site in the Hopper and others (1991) data is given in figure 2.8. The committee feels that it is more reasonable to recommend an (unweighted) arithmetic average radon concentration of