Meteorological Drivers of Radon Variability and Implications for Human Exposure

Radon is a key tracer of subsurface–atmosphere exchange and an important contributor to natural radiation exposure. Its temporal variability is strongly influenced by meteorological conditions; however, the relative roles of atmospheric and soil-related drivers remain insufficiently understood, complicating exposure assessment and prediction.

In this study, we analyze high-resolution radon concentration time series using a state-space modeling framework, in which observed concentrations are linked to an unobserved emission component. Meteorological variables, including precipitation, temperature, and humidity-related indicators, are incorporated as external forcings. Predictor selection is performed using LASSO regression, followed by a parsimonious dynamic model that captures latent emission variability and its response to environmental conditions. This approach enables explicit separation of atmospheric effects from subsurface controls.

The results indicate that variables associated with soil moisture conditions exert the strongest control on radon variability, while precipitation alone shows limited and indirect influence. Short-term fluctuations are primarily driven by atmospheric mixing processes, whereas longer-term variability reflects soil moisture dynamics affecting radon exhalation. These findings demonstrate that neglecting subsurface state variables can lead to biased interpretation of radon data.

Overall, explicit consideration of latent soil processes improves the characterization of meteorological drivers and provides a more reliable basis for assessing human exposure under changing environmental conditions.