Changes in the chemical make-up of underground water may be a warning that an earthquake is on the way.
This is not a new idea, but it is a difficult one to prove. Seismologists have seen changes in levels of radon gas, concentrations of dissolved elements and the ratios of stable isotopes in water. But the studies typically did not have enough data to confidently link any shifts to subsequent earthquakes. Now new data from Iceland offers a detailed picture of how water changed ahead of two earthquakes of magnitude 5.6 and 5.5 in 2012 and 2013.
of Stockholm University in Sweden and his colleagues analysed weekly water samples from a borehole in northern Iceland for 5 years. Four to six months before each of the earthquakes, they observed a change in the mixture of hydrogen isotopes, and a shift in the levels of sodium ions.
Advertisement
Because they monitored water chemistry for a long time, and spotted specific changes before two earthquakes, the team had enough data for a statistical analysis to show that the chemistry changes were most likely associated with the earthquakes and not the result of random fluctuations.
鈥淭he whole idea of chemical changes in groundwater before earthquakes is highly disputed,鈥� says Skelton. 鈥淥ur study makes a pretty convincing case that these really were related to the earthquake.鈥�
Early days
So far, Skelton has not determined what caused the chemical changes. It could be that other sources of water had started mixing with the water the team were sampling, perhaps as a result of shaking or stress-induced cracking of the underground rocks. Because of these uncertainties, says Skelton, 鈥渨e鈥檙e not ready to use the 鈥榩rediction鈥� word yet鈥�.
鈥淭he results shown give cautious optimism,鈥� says of King鈥檚 College London. 鈥淏ut there is still a lot more evidence that needs to be gathered, in other study sites and for longer time spans, before we can consider large changes of groundwater chemistry to always be a precursor to medium-to-large seismic activity.鈥�
Seismologist of the University of Edinburgh, UK, agrees: 鈥淭here is a long way to go before observations such as these could be turned into diagnostic operational tools for forecasting earthquake probabilities.鈥� Most geochemical signals fluctuate anyway, so he says it is 鈥渧irtually inevitable鈥� that some signals will coincide with earthquakes.
鈥淔or this reason,鈥� says Main, 鈥渢he global community is now focused on prospective tests: making a public forecast in real time, and only testing in retrospect when a realistic large sample of events can be analysed fully for hits, misses and false alarms.鈥� Only after experiments like these will we be able to say whether possible predictors really work.
Journal reference:
