The majority of radon measurements in the built environment are made over sub-year periods and are then generally seasonally corrected (i.e. scaled by an appropriate seasonal correction factor (SCF)) to estimate the annual average radon concentration. SCFs are statistically derived and assume an underlying annual cycle, reflecting the widely observed seasonal variation in indoor radon concentrations. In the UK, Public Health England has pioneered the calculation and use of a national SCF set using an annual sinusoidal model for variations in radon concentration and averaging across the entire country.

To test the validity of that model, a 4 year record of weekly radon data from four houses in Brixworth (Northamptonshire, UK) was analysed in conjunction with corresponding weather data for the period from a nearby weather station. The radon data showed a statistically significant annual cycle comprising both annual sinusoidal and second harmonic (i.e. 6 month period) terms.

Two sets of SCFs were calculated: first, using a conventional annual sinusoidal model that explained 21.2% of the variance in the radon data; and, secondly, a second harmonic term was included in the model that explained 24.6% of the variance. This represents an improvement of 3.4 percentage points (15.9%) and, thus, will result in better SCFs.