Sensitivity Analysis of Microseismic Moment Tensor Inversion

Microseismic monitoring is an important method for the quantitative evaluation of hydraulic fracturing effect during unconventional oil and gas exploration, mainly including source imaging and focal mechanism inversion. Fracture directionality, scale and stress state of target areas are very important for the calculation of stimulated reservoir volume. We can obtain their information through the study on microseismic focal mechanism. In current study, more and more people use the Moment Tensor model to represent the microseismic source and perform the inversion. However, there are many factors that may affect the microseismic moment tensor inversion, thus adding errors to the result. Through the sensitivity analysis of microseismic moment tensor inversion, we can deeply understand the moment tensor inversion process and quantitatively evaluate the reliability of inversion results.

In this paper, special attention is payed to investigate the sensitivity and reliability of the microseismic moment tensor inversion. The microseismic moment tensor inversion method we use is based on the anisotropic elastic wave equation and waveform matching technique. Using synthetic datasets and numerical tests, the impact of various factors that may affect the inversion results, e.g., errors in model (P- and S-wave velocity, anisotropy), noise in data (three-component vs. single component, background noise), and the effect of source location errors are quantitatively analyzed. We did several moment tensor inversions which have errors in models, data and source location, and compared the inversion results with the correct value. According to our tests, the microseismic moment tensor inversion can tolerate moderate errors in velocity models and data noise if stations' azimuth coverage is good, while the influence of shale reservoir's anisotropy and errors in source location is significant and these parameters should be well constrained in order to get credible microseismic source mechanisms.