Methods for identifying high radon areas of the US: The case of Minnesota
- Lawrence Berkeley Laboratory, CA (United States)
We have been developing methods for identifying high radon areas, based on the correlation of indoor monitoring data with information on housing structure, soil or geological, and meteorological parameters. Examination of short-term data in Minnesota using ordinary regression analysis indicates that surficial radium concentrations alone account for about 60% of the variance in county geometric-mean (GM) concentrations, but such analyses do not take proper account of the small numbers of houses sampled in many counties. In contrast, Bayesian analysis indicates that the variance in the true county GMs is less than that calculated directly from the monitoring data, and that the surficial radium data account for approximately 80% of the variation of the logarithm of the true GMs. We have carried out a new survey of year-long radon concentrations in 926 Minnesota homes, primarily to develop a denser set of data in a small number of counties in order to explore use of Bayesian modeling to predict concentrations for smaller areas than counties, such as census tracts. Initial analyses suggest the following: (1) The spatial distribution of the long-term data, and its predictability from the radium data, are similar to that of the short-term data, though the long-term concentrations are typically a factor of two less, since the short-term {open_quotes}screening{close_quotes} data were usually from detectors in basements. (2) The variation of census-tract GMs within counties is modest compared With the variability of county GMs across the state. (3) The soil and geologic information add only slightly to the predictive power provided by the radium data. We have also been applying the Bayesian approach 10 improve estimates of county-mean indoor radon concentrations across the United States.
- OSTI ID:
- 394063
- Report Number(s):
- CONF-9607135-; ISSN 0017-9078; TRN: 96:028741
- Journal Information:
- Health Physics, Vol. 70, Issue Suppl.6; Conference: 41. Annual Meeting of the Health Physics Society, Seattle, WA (United States), 21-25 Jul 1996; Other Information: PBD: Jun 1996
- Country of Publication:
- United States
- Language:
- English
Similar Records
Bayesian Prediction of Mean Indoor Radon Concentrations for Minnesota Counties
Predicting New Hampshire Indoor Radon Concentrations from geologic information and other covariates