Discrimination of earthquake and quarry blast data in different regions of Turkey

In this study, seismic events in Erikli and Enez districts (Edirne, Türkiye), Gelibolu district (Çanakkale, Türkiye), Kula district (Manisa, Türkiye), Lodumlu district (Ankara, Türkiye) and their vicinities are investigated. Thus natural and artificial seismic events are examined. 150 (Edirne), 77 (Manisa) and 156 (Ankara) digital vertical component velocity seismograms of seismic events with Md3.5 duration magnitude from seismic activity catalogs between 2009 to 2014 recorded by Enez (ENEZ), Erikli (ERIK), Gelibolu (GELI) , Kula (KULA) and Lodumlu (LOD) operated by Bogazici University, Kandilli Observatory and Earthquake Resarch Institute Regional Earthquake-Tsunami Monitoring Center (BU-KOERI-RETMC) were used in this study, respectively. The maximum S-wave and maximum P-wave amplitude ratio (Ratio) of vertical component velocity seismograms, power ratio (Complexity-C) and total signal duration of the waveform (Duration) are calculated. The earthquakes and the quarry blasts are discriminated by using the linear discriminate function (LDF) method and Back Propagation-Feed Forward Neural Networks (BPNNs) Learning Algorithm that is one of the artificial neural networks (ANNs) approaches by taken correlation between these parameters into consideration. For Edirne 69 (46%) of studied total 150 seismic events are determined as quarry blasts and 81 (54%) of them as earthquakes, for Manisa 39 (51%) of studied total 77 seismic events are determined as quarry blasts and 38 (49%) of them as earthquakes, for Ankara 63 (41%) of studied total 156 seismic events are determined as quarry blasts and 92 (59%) of them are determined as earthquakes, respectively. According to the results attained from this research, it is deduced that LDF method is as successful as ANNs approach.