Noisy Dispersion Curve Picking (NDCP): A Matlab package for group velocity dispersion picking of seismic surface waves

摘要

We developed a Matlab package, Noisy Dispersion Curve Picking (NDCP), which allows full control over several parameters to properly identify group velocity dispersion curves using two types of datasets: i) Green's Functions recovered from cross-correlations between pairs of stations and ii) earthquake surface waves.In seismic noise analysis, a non-homogeneous distribution of noise sources could lead to a non-symmetric reconstruction of the Green's function. In such a case NDCP allows to the user to pick the dispersion curve in the time domain both in the causal and anticausal part of the signal, allowing retrieval of more complete information. Furthermore, NDCP applies a narrow filtering to the signal allowing for better judgment of which periods can be reliably picked. Then the picked dispersion curve is displayed on the Frequency Time diagram in order to verify that it matches the maximum of the signal energy and avoids confusion with other spikes of energy or overtones. The same procedure is applied to earthquake data where the picking is also carried out in the time domain.To illustrate how NDCP performs, we show three examples: i) local scale correlograms functions obtained from sensors deployed into a geothermal-volcanic caldera, ii) regional scale correlograms functions between two stations installed by the Mexican National Seismological Service, and iii) the surface waves from an earthquake in the Pacific Ocean near the coast of Mexico.