ABSTRACT

Radon has continuously been monitored at the Roman spring of Tiberias, Israel since 2000 in the frame of an earthquake research project. However, there was no apparent earthquake related radon anomaly in 5 years of monitoring. Physical mechanisms behind periodic as well as transient radon variations were investigated. The radon signal contained periodic daily and non-periodic multi-day variations as well as seasonal patterns with maxima during winter. Spectral analysis showed diurnal and semidiurnal periodic constituents while tidal effects were absent. In 2003 the long-term average radon concentration dropped by 35%. Coevally, the diurnal and semi-diurnal radon variability considerably decreased. In contrast, the intensity of large-scale signals, corresponding to multi-day radon variability, increased. At this stage the level of the Kinneret Lake is suspected to be the driving force for the radon drop. Until 2003 the lake level hovered around –214 m below sea-level. In spring 2003 the lake level had risen by 4 m. The distance between the monitoring station and the lake shore is about 50 m. The radon concentration inversely followed the lake level with a time delay of about 3 months. Radon measured at a natural hot spring should depend on the flow rate of the hot water rising on the border faults of the pull-apart basin. Increased flow means less time for radon to decay and thus a positive correlation between the flow rate and the radon concentration is expected. Flow velocity is controlled by (i) the pressure at depth, and (ii) the fracture width. Both are affected by the loading forces of the graben filling to which the water column of the lake contributes. Due to the lack of data about the mass flow rates from the spring, a direct link between the flow rate and the radon concentrations cannot be proven. In fact, the hot water discharge seemed to be very stable in time. So, either minor changes of the flow rate affect the radon concentration or another mechanism is needed to explain the observations, e.g., the pressure-dependent gas solubility or the pressure-dependent mixing of different groundwater components. Nevertheless, this does not explain the appearance of long-periodic, intra-seasonal radon signals (with periods in the order of 1 month) which were practically absent before 2003. Such long-periodic radon signals were not reported till today.