M.Sc. (Eng.),Faculty of Engineering and the Built Environment, University of the Witwatersrand, 2012 / Mining is a sensitive business that yields high returns and at the same time is associated with high risk of injuries/fatalities and potential losses of revenue. There is increasing intolerance for injuries and fatalities by governments and the other stakeholders involved in the mining business often resulting in mine closures and revenue loss. Chief among the mining risks is the occurrence of rockfalls where people work and access. The rockfalls are bound mainly by joints that intersect the rockmass thereby forming rock blocks that may fall once an excavation has been created. There are many methods that have been used over time to predict the occurrence of rockfalls. More recently probabilistic methods have gained more ground over deterministic methods. The properties of the joints that are identifiable on exposed excavations are the main inputs used in simulating rockfalls. To date there has been little work that has been done to compare predicted rockfalls to actual rockfalls. This dissertation presents a practical method for collecting rockfall and joint data in the stope hangiwall at two mines in the Bushveld Complex. The joint data has been used in simulating rockfalls using JBlock (a probabilistic keyblock stability programme). A comparison between simulated rockfalls and mapped rockfalls has been presented. Based on this comparison, a number of iterations were done to calibrate the JBlock results until near realistic rockfalls were achieved. Three case studies have been conducted to investigate the effectiveness of different stope support systems in reducing rockfall. The potential losses and injury risk associated with the different support systems have been quantified for all the individual rockfalls. In general the rockfall frequency is directly proportional to the risks associated with the rockfalls. Through this research it has been demonstrated that it is possible to use joint data found on excavation surfaces to statistically predict the occurrence of potential rockfalls in similar ground conditions. The optimum support system that has minimum injury and cost risk can also be selected from a comparison of a number of support systems. Armed with this information, rock engineers can now make strategic decisions versus the existing common tactical approach.