Analysis of chaotic behavior of indoor radon concentrations

Indoor 222Rn concentrations are influenced by several factores which may change with time,thereby causing temporal fluctuations of radon concentrations in rooms. Different chaos basedmeasurements (fractal analyses) were applied to radon time series in three different rooms (kitchen,working room, radiation laboratory) in Austria to determine the degree of chaotic behavior, topredict fluctuations in the future, and to investigate correlation's with meteorological parameters.Our fractal analyses of these indoor radon time series demonstrated that indoor 222Rnconcentrations do indeed exhibit features which are characteristic of chaotic systems. Thecomputed fractal dimensions, such as Hurst exponent, Lyapunov exponent, capacity dimension,embedding dimension and attractor dimension, provided estimates of the degree of chaoticbehavior, such as low dimensional chaos for the kitchen, high dimensional chaos for the workingroom and chaos with additive noise for the radiation laboratory. Application of a nonlinearprediction algorithm revealed that the predictability of radon time series is restricted toapproximately three relative time steps into the future and into the past, e.g. Monthly radonmeasurements cannot be reliably extrapolated beyond a period of three months. If the similarity offractal dimensions between radon concentrations and meteorological parameters is used as ameasure of the degree of correlation, our analyses suggest a strong correlation between radonconcentrations and negative pressure differences.