Microearthquake mechanism from wave amplitudes recorded by a close-to-surface seismic array at Ocnele Mari, Romania

Abstract

The resolution of event mechanism is investigated in terms of the unconstrained moment tensor source model and the sheartensile crack source model representing a slip along the fault with an off-plane component. Data is simulated as recorded by the actual seismic array installed at Ocnele Mari (Romania), where sensors are placed in shallow boreholes. Noise is included as superimposed on synthetic data, and the analysis explores how the results are influenced (i) by data recorded by the complete seismic array compared to that provided by the sub-array of surface sensors, (ii) by using threeor one-component sensors, and (iii) by inverting Pand S-wave amplitudes vs. P-wave amplitudes only. The orientation of the pure shear fracture component is resolved almost always well. On the other hand, the noise increase distorts the non-double-couple components (non-DC) of the moment tensor unless a high quality dataset is available. The shear-tensile crack source model yields considerably less spurious non-shear fracture components. Incorporating recordings at deeper sensors in addition to those obtained from the surface ones allows for the processing of noisier data. Performance of the network equipped with threecomponent sensors is only slightly better than that with uniaxial sensors. Inverting both Pand Swave amplitudes compared to the inversion of P-wave amplitudes only markedly improves the resolution of the orientation of the source mechanism. Comparison of the inversion results for the two alternative source models permits the assessment of the reliability of non-shear components retrieved. As example, the approach is investigated on three microseismic events occurred at Ocnele Mari, where both large and small non-DC components were found. The analysis confirms a tensile fracturing for two of these events, and a shear slip for the third.