Signal processing of soil gas radon, atmospheric pressure, moisture, and soil temperature data: A new approach for radon concentration modeling

The variability in space and in time of gases (He, Rn-222, CO2) in soils might be used for volcano or seismic fault surveillance or in search of hidden mineral deposits, The gases measured in soils or in the weathered layer of the substratum can, however, be strongly altered by environmental conditions, such as atmospheric pressure, soil temperature, or moisture, An accurate knowledge of the influence of environmental conditions is required to decipher information from deeper phenomena in the earth, Variations in radon concentration, used as a gas flow tracer, are modeled using a new approach based on signal processing techniques in order to express impulse responses from multivariable time series, A general formulation is proposed and the inverse problem is solved by calculating the impulse responses of the parameter to be explained versus the variables responsible for the variability of the parameter, Such problems are generally ill posed and regularization methods must be used to develop the pattern, Due to the block-Toeplitz structure of the correlation matrix, large orders can be used, enabling signal processing of time series controlled by the superposition of fast and slow phenomena, Two examples are processed to illustrate this approach. In the first example (Rn-222 in soil gas monitoring over a hidden sulphide deposit), radon concentration is controlled only by moisture and temperature, not by atmospheric pressure, The result seems to reveal a diffusive behavior. The impulse response of Rn-222 concentration versus rainfall exhibits a great variability, possibly due to the opening of cracks in the surface, the effect being still noticeable 20 days after the rainfall, In the second example (Rn-222 in a thermal anomaly at summit of Etna volcano), where radon concentration is controlled by soil temperature and atmospheric pressure, the result indicates a convective/advective flow. Radon flux collapsed for 1 day probably due to a break in the rising gas column, In 1993, dramatic increases of signal were recorded for a few hours; such increases are interpreted as gas pulses not related with local seismicity.