Seismicity dataset of Ghana obtained by Deep Learning

This data repository presents waveforms, earthquake bulletin, and supporting data for a set of detected earthquakes on the Ghana Digital Seismic Network (GHDSN), consisting of 6 broadband sensors deployed in southern Ghana from 2012 to 2014. The earthquakes and their corresponding P and S arrival phases were detected using the EQTransformer tool, a Deep Learning model for simultaneous earthquake detection and phase picking. The data set comprises waveforms of 73 detected local earthquakes and a bulletin of 559 arrival times (292 P and 267 S phases) in SEISAN format. Additionally, some preliminary results of processing this earthquake data are presented: 1) An accurate, updated crustal velocity model for the region obtained by performing a joint inversion on the hypocentral parameters and Vp and Vp/Vs for a 6-layers model, 2) 3D live figures graphically representing the hypocentral parameters of the detected earthquakes and active zones in this period, 3) an occurrence time sequence for the detected events, and 4) some statistical information of the earthquakes detected. This dataset is essential as it provides researchers interested in further processing the waveform of the detected events to derive more detailed information about the mechanism of the seismogenic sources and active faults in the area.