Seismic noise wavelet-based entropy in Southern California

Seismic noise properties in Southern California are considered. The initial data are continuous records of vertical oscillations with a sampling frequency of 1 Hz at 81 broadband stations for 12 years, 2008–2019. These data were converted to a time step of 1 min by calculating the average values in successive time intervals of 60 samples in length. The time series of low-frequency seismic noise with a time step of 1 min from each station was further converted to a sequence of daily values of the minimum normalized entropy of the distribution of the squared orthogonal wavelet coefficients. Since entropy estimates can be obtained from each station with a time step of 1 day, this makes it possible to construct a daily map of its changes in space. A general map of the distribution of entropy values is obtained by averaging daily maps. The main attention is paid to areas in which maximum entropy values are most often realized. These areas are identified by estimating the spatial probability density of the distribution of points at which a given number of maximum values are realized in each daily map. It has been shown that since 2012, the region of the most frequent realizations of entropy maxima is in direct contact with the epicenter of a strong earthquake on July 6, 2019, with magnitude 7.1. We consider the “secondary” entropy calculated for the probability densities of the distribution of the maximum values of the “primary” entropy of waveforms of seismic noise in a semi-annual moving time window. It is shown that the time intervals of increasing seismic activity correspond to a decrease in secondary entropy, which is interpreted as the concentration in the space of the distribution of the maxima of the primary noise entropy. Estimates of the change in the correlation coefficients between the daily values of the noise entropy in a semi-annual time window at network nodes of 12 reference points covering the region under study made it possible to study the spatiotemporal dynamics of strong correlations. The area of the future strong earthquake is characterized by high correlations of entropy values at the nearest reference points with other points.