Towards an early warning system for volcanoes
Mount Etna in Italy and several other volcanoes throughout Europe are still active and pose a threat to the local population. The challenge to scientists is to understand the signals generated by volcanoes and to forecast when the next eruption might occur. One of the more common antecedents to eruptions is increased frequency and magnitude of earthquakes. Experts from the Universitat de Barcelona applied advanced statistical modelling techniques to analyse seismic data from volcanoes in a project entitled MULTIMO. Seismograms must be carefully interpreted since the medium through which the signal travels may not be fixed as magma protrudes into the volcano over time. In order to minimise the possibility of incorrect forecasts, a non-linear model was implemented employing spectral analysis. Since early warning and evacuation, if necessary, is of utmost importance, the Spanish scientists came up with a way to analyse the time series in quasi real-time. The Universitat de Barcelona tested the novel system in real-world seismic monitoring stations. In addition, seismic data collected from volcanic monitoring sites was processed offline to further validate the system's algorithms. Additional functionality includes early detection of malfunction and simultaneous spectral analysis for multiple components. In combination with other tools, the new system can contribute to improving early warning systems and minimising the volcanic threat.
