Before an earthquake hits the Iberian Peninsula, specific patterns in seismic behaviour can be detected, Spanish researchers have found. Presented in the journal Expert Systems with Applications, the study could prove vital for researchers seeking to determine when an earthquake will strike. The team recognises, however, that the chances of forecasting an earthquake with 100% accuracy are slim. The research is based on data obtained by the Instituto Geográfico Nacional, specifically from 4,017 earthquakes measuring between 3 and 7 on the Richter scale whose epicentre was on the Iberian Peninsula or in the surrounding waters between 1978 and 2007. The researchers from the Universidad Pablo de Olavide (UPO) and the Universidad de Sevilla used clustering techniques to estimate various seismic movements, medium and large in size, under specific circumstances. 'Using mathematical techniques, we have found patterns when medium-large earthquakes happen, that is, earthquakes greater than 4.4 on the Richter scale,' explains UPO's Francisco Martínez Álvarez, one of the authors of the study. By applying clustering techniques to the data, the researchers discovered similarities and patterns, thus giving them better clues on when earthquakes can occur. The seismologists focused their research on two regions that offered the most information: the Alboran Sea and the Western Azores-Gibraltar fault region. They paid particular attention to three key elements: the magnitude of the seismic movement; the time elapsed since the last earthquake; and the change in a parameter, called the b-value, from one earthquake to the next. The b-value reflects the tectonics of the region under review. In a nutshell, a high b-value signifies that earthquakes are mostly small in size, and so the land has a low level of resistance. A low b-value indicates a relatively similar number of small and large seismic movements, suggesting that the land is more resistant. 'We have discovered the strong relationship between earthquakes and the parameter b-value, recording accuracy rates of more than 80%,' says Antonio Morales Esteban of the Universidad de Sevilla, a co-author of the study. 'After the calculations had been performed, providing the circumstances and sequences we have determined to be forerunners occur, we obtain a significant success probability.' The researchers say the technique enables two types of forecasts: sensitivity and specificity. Sensitivity is the probability of an earthquake happening after the patterns detected occur, and specificity is the probability of an earthquake not occurring when no patterns have emerged. The team found a sensitivity of 90% and specificity of 82.56% for the Alboran Sea region, and 79.31% and 90.38% respectively for the Western Azores-Gibraltar Fault. Their results show that the chance for an earthquake to occur in these regions right after these patterns occur (high sensitivity) is high, and that they largely occur only after the seismic patterns (high specificity). Further, the researchers are using their own algorithms to evaluate the same data. These algorithms are based on 'association rules', which are altnerative mathematical methods used to determine common events, or those which satisfy specific conditions within a set of events. 'The results are promising, although I doubt we will ever be able to say that we are capable of forecasting an earthquake 100% accurately,' Dr Martínez Álvarez says.