2016 PSPA – Pilar Sánchez - Advantages using Phase Cross-Correlation for searching temporal structural changes. Application to 2011 El Hierro eruption

Ambient seismic noise is a continuous vibration of the ground due to natural and artificial sources although the main origin is the interaction between the atmosphere, the ocean and the solid Earth. This noise is recorded by seismic stations located everywhere in the world at any time and it does not depend on earthquake activity. On other hand, the knowledge of the seismic wave field has been improving during the last decades; this allows it to study characteristics of its sources and the ground. For all of these reasons, its importance has increased in some areas of the Geophysics such as volcanic monitoring and tomography.

Figure 1. Summary sketch of a
seismic wave path  (grey  line).
The  red  triangle is  a  seismic
 station and the orange rectangle

 shows the recorded signal.        

Seismic waves created by noise sources travel through the Earth and suffer multiple scattering during their path until arriving to seismic stations (Figure 1). So we cannot know this path but if the ground and sources are equal in time we have to obtain coherent and stable signals after a while, i.e., we can obtain a stable trace that will be directly related to seismic response of the medium.

In order to obtain that trace, it is necessary a coherence measure that can be reached with the correlation between two temporal series. During this work, we present the advantages to use the phase correlation instead of classical correlation for different approaches and explain the constraints we have to take into account for the data processing.

We further apply these concepts to field data of nine seismic stations installed in El Hierro (Spain). We focus on finding temporal structural changes due to the 2011 eruption that happened in this island. It is important to work in a frequency band where the noise source not change its properties and where tremors not be recorded. To do so, we calculate spectrograms with which we can analyse the energy distribution in function of frequency and time (Figure 2).

Figure 2. Spectrogram for ten days of CCUM station. The white arrow marks the onset of the eruption. 

Once all the parameters are chosen, we calculate correlations with different methods of the whole available data and extract preliminary but promise results about the 2011 El Hierro eruption.

This work is supervised by Martin Schimmel (ICTJA-CSIC). Economic support has been provided through CGL2013-48601-C2-1-R project funded by the Spanish Ministry of Science and Innovation. We acknowledge National Geographic Institute for providing the seismic data for the study.