Seismic hazard analysis using mapping techniques for underground, narrow-vein metal mines

Abstract

This thesis investigates the ability to forecast the occurrence of large events through node-based hazard mapping at KGHM’s Morrison Mine located in Levack, ON. Nodebased hazard maps function by pasting a heat-map onto a solid model of mine workings to identify, visually, the areas that have experienced high levels of seismicity. The notion within this research is that it is possible to forecast the occurrence of large seismic events with a reasonable degree of effectiveness using a trailing period of 6-months and a forecast period of the following 2-months. Using the following parameters, each hazard map has been analyzed over three different forecast periods: • Cumulative Seismic Energy • Cumulative Seismic Moment • Number of Events > Defined Magnitude o M > 0 o M > -0.5 o M > -1 • Number of Events > Apparent Stress, 80th percentile The forecast periods analyzed were April – June 2016, June – August 2016, and August – October 2016 with a success rate (the number of events successfully forecasted) between 38% and 67% and false alarm rates (the number of denoted hazard areas that did not experience large events in the forecast period) between 41% and 65%.