Time series analysis and risk assessment of domestic radon: Data collected in dwellings along fault lines

Abstract

Indoor radon-in-air (IR) concentrations in dwellings situated on a fault line passing underneath Muzaffarabad City were measured with a continuous radon monitor. Measurements were taken at half an hour intervals at 11 sites for 48 h. The values ranged from 14 ± 1 to 155 ± 4 Bq.m^− 3 with an average value of 42 ± 2 Bq.m⁻³. The geometric mean (GM), standard deviation (STD) and geometric standard deviation (GSD) were 32 ± 2, 40.2 ± 0.8 and 1.5 Bq.m⁻³, respectively. Annual mean effective doses ranged between 0.35 ± 0.03 and 3.9 ± 0.1 mSv.y⁻¹ with an overall mean value 1.05 ± 0.05 mSv.y⁻¹. The experimental data were analysed by comparing algorithms to extract phase space dynamical information in the radon time series. Such approaches are potentially useful for deriving methods of extrapolation of short-term measurements to annualized average predictions and for earthquake detection. The computed values for Hurst exponent (H) from the rescaled range analysis and the Lyapunov exponent (L) provided estimates of the degree of chaotic regime. The Hurst exponent (0.877) for the radon time series indicates that the radon time series is a persistent time series. The value of Lyapunov exponent (0.0021) shows deterministic chaos in the data. The permutation entropy value was 0.6894 reflecting the irregularity in IR time series. The value of entropy obtained confirmed that the predictability of the IR time series is restricted to approximately one time step into the future. The mean value of IR concentrations corresponds to excess relative risks (ERR) of radon-induced lung cancer for 35- and 55-year-olds at all sites of 0.25 ± 0.01 and 0.205 ± 0.01, respectively. The observed 48-h average level (42 ± 2 Bq.m⁻³) was less than those recommended by the World Health Organization (WHO) (100 Bq.m⁻³), US Environmental Protection Agency (148 Bq.m⁻³) and Health Protection Agency, UK (200 Bq.m⁻³).