A data set of long-term radon measurements in about 2200 houses in southern Belgium has been collectedin an on-going national Rn survey. The spatial variation of indoor Rn concentrations is modelled byvariograms. A radon distribution map is produced using the lognormal kriging technique. A GIS is usedto digitise, process and integrate a variety of data, including geological maps, Rn concentrationsassociated with house locations and an administrative map etc. It also allows to evaluate the relationshipsbetween various spatial data sets with the goal of producing radon risk maps.
The use of two mapping methods, kriging and moving average, in mapping the indoor radon risk, isinvestigated. Both methods are applied to three databases: simulated radon data, data collected inSouthern Belgium with low sampling density, and data collected in Luxembourg. We use commercialsoftware (SURFERR 6) for kriging, as well as softwares developed by the authors, especially for the radoncase, for kriging and for moving average. Simulated data prove to be very useful in this context. Weconclude that kriging as implemented in SURFERR 6 may not be well adapted to radon mapping.
Login