Törnman, W, Martinsson, J & Svanberg, E 2024, 'Enriching seismic data with noise and blasts and the importance of credibility', in P Andrieux & D Cumming-Potvin (eds), Deep Mining 2024: Proceedings of the 10th International Conference on Deep and High Stress Mining, Australian Centre for Geomechanics, Perth, pp. 207-216, https://doi.org/10.36487/ACG_repo/2465_08
(https://papers.acg.uwa.edu.au/p/2465_08_Martinsson/)
Abstract: Understanding the rock mass response to mining is fundamental for both safety and productivity in seismically active mines. Seismic data provides valuable insights beyond the rock surface such as hypocentres and source parameters, enabling mining engineers to find relationships between mining operations and rock movements.
Any conclusion drawn using seismic data is directly impacted by the credibility of the estimated seismic parameters. Unfavourable sensor placements, uncertain and biased estimates in combination with missing metrics to provide proper quality assurance and quality control (QA/QC) often result in consultations with experts to interpret the parameters and to assess the correctness. 
In this paper we present a robust processing solution that overcomes these challenges by utilising statistical techniques combined with self-learning capabilities. It present results from different case studies where the processing technique is applied and shows the impact of self-learning to improve the quality of the estimated parameters and in combination with easy to use tools to describe statistical uncertainty descriptions (credible regions) for intuitive QA/QC.
This paper shows that different types of data serve distinct purposes, and shows the importance of processing all triggered events, regardless of origin. Genuine seismic events are crucial for hazard assessments, while blasts and noise events generated by mining operations help evaluate and enhance the seismic system and offer valuable insights on mining activities.

Keywords: Bayesian seismic processing, self-learning capabilities, credibility, QA/QC, data inclusion