Robust seismic velocity change estimation using ambient noise recordings

We consider the problem of seismic velocity change estimation using ambient noise recordings. Motivated by Zhan et al., we study how the velocity change estimation is affected by seasonal fluctuations in the noise sources. More precisely, we consider a numerical model and introduce spatio-temporal seasonal fluctuations in the noise sources. We show that indeed, as pointed out by Zhan et al., the stretching method is affected by these fluctuations and produces misleading apparent velocity variations which reduce dramatically the signal to noise ratio of the method. We also show that these apparent velocity variations can be eliminated by an adequate normalization of the cross-correlation functions. Theoretically we expect our approach to work as long as the seasonal fluctuations in the noise sources are uniform, an assumption which holds for closely located seismic stations. We illustrate with numerical simulations in homogeneous and scattering media that the proposed normalization significantly improves the accuracy of the velocity change estimation. Similar behaviour is also observed with real data recorded in the Aegean volcanic arc. We study in particular the volcano of Santorini during the seismic unrest of 2011–2012 and observe a decrease in the velocity of seismic waves which is correlated with GPS measured elevation.