Rockfall forecasting at a British Columbia mine using meteorological and thermal imaging data

Authors: Reasoner, AE; Ortmann, CJ; Potter, JJ; McNabb, JC; Meyer, BJ; Bidwell, AK; Brown, LD Download Paper Open access courtesy of:

DOI https://doi.org/10.36487/ACG_repo/2535_09

Cite As:
Reasoner, AE, Ortmann, CJ, Potter, JJ, McNabb, JC, Meyer, BJ, Bidwell, AK & Brown, LD 2025, 'Rockfall forecasting at a British Columbia mine using meteorological and thermal imaging data', in JJ Potter & J Wesseloo (eds), SSIM 2025: Fourth International Slope Stability in Mining Conference, Australian Centre for Geomechanics, Perth, https://doi.org/10.36487/ACG_repo/2535_09



Abstract:
It is generally accepted within the mining geotechnical community that rockfall initiation is influenced by meteorological conditions, such as rainfall and freeze-thaw cycles. Rockfall typically occurs rapidly and without warning, making it difficult to predict the timing and location of individual events that would pose a risk to mine workers. However, quantifying the links between weather patterns and rockfall mechanisms provides an opportunity to forecast periods of increased risk based on prevailing meteorological conditions. Previous research from the University of Arizona’s Geotechnical Center of Excellence (GCE) analysed rockfall occurrences at a mine in southern Arizona, where strong solar irradiance, monsoonal rainfall, and diurnal temperature variations are likely drivers of rockfall activity. This study builds on that work by exploring the relationship between verified rockfall events and freeze-thaw processes at a steelmaking coal mine in British Columbia. Although data were collected between 27 January–13 April 2022, the analysis focused on a narrower window (24 March–11 April) selected for its pronounced freeze-thaw cycling. Rockfall events were identified from thermal imaging, using a computer vision algorithm and verified by a trained practitioner. Statistical and machine learning models, including logistic regression and tree-based classifiers, were applied to identify key meteorological variables associated with rockfall occurrence and to develop preliminary predictive models that forecast periods of elevated rockfall risk at this location. By comparing results from sites in contrasting climates, this study advances our understanding of how varying environmental conditions influence rockfall in open pit mines, and supports the development of predictive models that enable a more refined and higher-confidence approach to rockfall risk management and mitigation.

Keywords: rockfall, slope stability, machine learning

References: