An Adaptive Probabilistic Imaging Location Method for Microseismic Monitoring

Microseismic event locations provide critical insights into fracture locations and stress conditions within rock formations, which are essential for seismic hazard monitoring. In practice, pick-based location methods are widely utilized due to their high computational efficiency. However, the accuracy of picking is compromised when the signal-to-noise ratio (SNR) is low. Source location utilizing equal differential time (EDT) surfaces between station pairs represents an effective strategy for mitigating picking errors. Typically, EDT surfaces are constructed using either a fixed width or a fixed probability density function (pdf), which presents challenges in simultaneously achieving high resolution and accuracy. To address this issue, we propose an adaptive probabilistic imaging location method that constructs EDT surfaces by incorporating an adaptive pdf linked to the SNR of arrival picking data. The location probability imaging function is defined as the product of the independent EDT surfaces used for locating sources. For high-SNR data, where picking errors are typically small, the adaptive pdf converges more rapidly than the Gaussian distribution, yielding higher resolution by assigning lower probabilities to locations distant from true positions. For low-SNR data, where picking errors are typically large, the adaptive pdf exhibits heavy-tailed characteristics with a slower rate of decrease, enhancing the accuracy by assigning higher probabilities to likely true locations. The effectiveness and stability of the method are validated through theoretical analyses and synthetic data tests. Application to mine microseismic data indicates that the proposed method improves the accuracy and resolution of microseismic event locations relative to other methods.