Martinsson, J, Törnman, W & Svanberg, E 2024, 'Responsive short-term seismic forecasting: a web-based tool for mining efficiency and safety', in P Andrieux & D Cumming-Potvin (eds), Deep Mining 2024: Proceedings of the 10th International Conference on Deep and High Stress Mining, pp. 995-1002, https://doi.org/10.36487/ACG_repo/2465_63
(https://papers.acg.uwa.edu.au/p/2465_63_Martinsson/)
Abstract: Predicting the rock mass response to future mining is crucial for ensuring the safety and efficiency of operations in seismically active mines. Long-term forecasts often rely on historical seismic data and various inputs that may be available such as faults, geology, mining sequences, stress measurements and numerical modelling. Short-term predictions – spanning a few days to a week – demand a more dynamic and responsive approach. This paper introduces an online web application based on a hierarchical Bayesian prediction model tailored for continuously updated short-term seismic activity forecasts. Further, the activity can easily serve as inputs to post-processing analysis such as hazard assessments while preserving prediction uncertainties.
The approach integrates recent seismic activity, production plans, depth, size of production, and a seismic exposure term within different volumes of the mine and lets them share information through a hierarchical model structure. Notably, for short-term predictions, many covariates considered in long-term forecasts are constant and summarised by an overall seismic exposure term in each volume.
Implemented as an online web application, this model provides direct seismic insights to personnel across various devices, from control room displays to smartphones. Its dynamic interface visualises historical and predicted seismic activity with prediction intervals, empowering decision-makers for safe and efficient mine operations. This study highlights the effectiveness of combining online web applications and robust Bayesian methods, enhancing safety protocols and operational efficiency in seismically active mines.

Keywords: short-term forecast, mine seismic activity, mining seismicity, production plans, web application, hierarchical Bayesian model