Of the any geochemical parameters that provide predictove information about movement in the earth's mantle, radon is the most
useful and the most studied. Earthquake prediction is a relatively new field and has been the subject of three recent intenational meetings and two journals. Stuidies of examining radon movement as a possible predictor of earthquakes have occurred primarily in Russia, the Republic of China, and the United States.
Radon emanating from soil and measured in groundwater would seem to be a good indicator of crustal activity such as earthquake. Such movement would be expected to change the radon levels. However, the current literature describing the possible correlation between radon levels and earthquake activity uses such qualifying
and cautious words as possible, apparent, limited, could sometimes, may be, and suggestive. It is clear that in some cases there are precurser changes in radon levels, but that the causal relationship or mechanism relating these to eartquake activity is not understood.
The measurements of potential earthquake precursors involve radon from groundwater in deep wells of a few thousand meters or from soil within one meter of the surface. The measurement span time periods of months and sometimes years and are usually done in the vicinity of a fault line or area of high seismic activity. The earthquakes studies are generally of a magnitude greater than 4 on the Rishter scale, although some possible correlatio nave been reported for smaller magnitudes. Figure below shows a radon concentration spike preceding a Chinese earthquake.
figure below: Time series radon concentration at Guza station in China, showing an relevated value before eartquakes of 7.2 on the Richter scale. Reproduced by permission from C. Y. King, Radon studies in China, Earthquake Predict. Res. 3, 4768(1985).
An earthquake is the mailure of a rock under the stress of an applied force and results in a shock. The corresponding changes in radon levels can be long-term or can occur just before the short-term or peaked changes is sharper. A long-term or several years buidup of radon levels has been observed such as before the 1966 TTashkent earthquake. Some short-term anomalies have been seen in Japan and China. In some cases, an anomaly in radon concentration is seen within a day or so of the event.
Radon has a mean diffusion lenght (the averae distance it moves by diffusion) in dry porous soil of about one meter and a few centimeteres in groundwater, and thus a limited ability to migrate due to its short half-life. An eartquake may therefore eighter change the rate of release of radon and thus change the convectio gradient at shallow depths or the extra pressure on the pore volume may cause an increased emanation. On the atomic level the theory is based on the recoil energy of the radon progency of about 100 keV. These atoms travel through hundreds of crystalline lattice sites and often get lodged in a microcapillary in the mineral structure eighter in the aqueous or air medium.
Several complications arise in trying to relate radon levels with earthquake activity. Rain contains radon washed out of the atmosphere and can change levels at the surface and in groundwater. Pressure and temperature affect the emanation levels of radon and thus introduce a seasonal variation. The local geological structure also is a factor as is the conductivity through it. Additional complications include tidal motions and other general seismic events. In spise of all the problems, it has been claimed that anomalies in radon levels have been detected tens to hundreds of kilometers from the epicenter of a subsequent earthquake. It is predicted that they could be measured as far away as 4000 km for an M = 8 (Richter scale) eartquake.
Earthquakes clearly need to be predicted both in location and magnitude. The measurements in the past have brought some successes and some failurefor radon changes as an indicator. The equipmet is costly and measuring is tedious, but the rewards could be significant if many lives can be saved by predicting an eartquke.
