Détectabilité et signification de la marée barométrique de 12h00 dans les enregistrements de radon-222, de ruisselement d'eau, de gaz carbonique et de température dans une cavité souterraine.
Titre du congrès :American Geophysical Union Fall Meeting
Ville du congrès :San Francisco
Date du congrès :10/12/2007
Radon concentration has been measured since 1995 in the Roselend dead-end tunnel, French Alps, together with other geophysical and geochemical parameters. Bursts of radon concentration, reaching 65 600 Bq×m-3 and up to several weeks duration, are observed over a background level of ca. 800 Bq×m-3. These bursts appear to be related to the bedrock deformation or to the hydrogeological processes associated with the yearly cycle of water
level in the nearby artificial Roselend Lake. In order to work out a generation mechanism for these bursts, we developed tools to characterize the transport properties in the host rocks.
Here, we focus on the 12h00 (S2) barometric tide. We first show, using real radon time series integrating synthetic signals, that a modified spectrogram method is more efficient than simple FFT to evidence weak periodic signals in such a context. Then, we apply this method to the radon concentration in the tunnel atmosphere measured by two different sensors:
the AlphaGUARDTM sensor based on volumetric detection in an ionizing chamber, and the BarasolTM sensor, based on surface detection. Using the time series recorded by the AlphaGUARDTM, the S2 line, difficult to observe with a simple FFT method, emerges clearly with our spectrogram method. This S2 line is not evidenced using the BarasolTM time series, illustrating the higher sensitivity of the AlphaGUARDTM for this particular purpose. Using a regular spectrogram analysis, we further show that the amplitude of the S2 line in the AlphaGUARDTM data depends on time, and appears particularly strong during the radon bursts. The presence of the S2 line reveals a high sensitivity of radon exhalation flux from the tunnel wall to changes of atmospheric pressure, and thus supports the advective transport mechanism for the radon bursts. A small but clear S2 component is also evidenced using our
spectrogram method in a dripwater flow rate time series, representing a flow averaged over a 6 m2 area of the tunnel ceiling, while it is not observed in the flow rate of a more localized dripping. This suggests that some water drippings can also be affected by atmospheric pressure variations. The temporal structures of the S2 component in the flow rate and in the radon concentration, however, are not similar, indicating that water dripping from the ceiling
cannot be the dominant source mechanism for the radon bursts. No S2 component is observed in the time series of carbon dioxide, but an interesting pattern is revealed by the S2 component of a temperature profile in the atmosphere. This study illustrates how a refined frequency domain analysis aiming to extract the S2 component in various geophysical time series can provide interesting clues on the complex processes affecting transport of fluids in unsaturated fractured media under multiple influences.