Probability mapping of indoor radon-prone areas using disjunctive kriging.

ABSTRACT

After a reference to the use of maps of radon-prone areas for indoor radon risk management, and to the methods used to produce them, there is a brief illustration of the geostatistical method of disjunctive kriging (DK) introduced by G. Matheron as a substitute for conditional expectation. There are some good reasons of using this method for the mapping of radon-prone areas as follows (1) spatial correlation is exploited; (2) unbiasedness is conserved even in the conditions of quasi-stationarity; (3) lognormality of the data is not required; (4) choosing the point estimation allows drawing up smooth probability maps. An application of DK is also presented for the production of probability maps in a campaign of indoor radon measurements conducted by Institute for Environmental Protection and Research, in the provinces of Rome and Viterbo (Central Italy). In the application, it is assessed in particular how much the spatial correlation, even though low, influences the results.