NOISY DISPERSION CURVE PICKING (NDCP): a Matlab friendly suite package for fully control dispersion curve picking

We developed a Matlab suite package (NDCP, Noisy Dispersion Curve Picking) that allows a full control over parameters to identify correctly group velocity dispersion curves in two types of datasets: correlograms between two stations or surface wave records from earthquakes. Using the frequency-time analysis (FTAN), the procedure to obtain the dispersion curves from records with a high noise level becomes difficult, and sometimes, the picked curve result in a misinterpreted character. For correlogram functions, obtained with cross-correlation of noise records or earthquake's coda, a non-homogeneous noise sources distribution yield to a non-symmetric Green's function (GF); to retrieve the complete information contained in there, NDCP allows to pick the dispersion curve in the time domain both in the causal and non-causal part of the GF. Then the picked dispersion curve is displayed on the FTAN diagram to in order to check if it matches with the maximum of the signal energy avoiding confusion with overtones or spike of noise. To illustrate how NDCP performs, we show exemple using: i) local correlograms functions obtained from sensors deployed into a volcanic caldera (Los Humeros, in Puebla, Mexico), ii) regional correlograms functions between two stations of the National Seismological Service (SSN, Servicio Sismológico Nacional in Spanish), and iii) surface wave seismic record for an earthquake located in the Pacific Ocean coast of Mexico and recorded by the SSN. This work is supported by the GEMEX project (Geothermal Europe-Mexico consortium).