S11C-06
Remote Dynamic Triggering of Earthquakes in Three Canadian Shale Gas Basins Based on a Multi-station Matched-filter Approach with Dense Station Coverage

Monday, 14 December 2015: 09:15
305 (Moscone South)
Bei Wang1, Rebecca M Harrington1, Yajing Liu1 and Honn Kao2, (1)McGill University, Montreal, QC, Canada, (2)Geological Survey Canada, NORTH SAANICH, BC, Canada
Abstract:
Earthquakes triggered by remote, transient stresses may indicate critical ambient stress conditions on host faults, independent of their proximity to plate boundaries. Here, we investigate dynamic triggering of three sedimentary basins in Canada where seismic station coverage has been increased to monitor anticipated increases in fluid injection activity: northeast British Columbia and western Alberta, the Norman Wells area of the Northwest Territories, and northeast New Brunswick. We select triggering mainshock candidates satisfying the following criteria: Ms > 6, and local peak ground velocity exceeding 0.01 cm/s. We find 31 mainshocks in northeast British Columbia/western Alberta, 9 in Norman Wells, and 4 in New Brunswick during increased station operation. We will investigate seismicity rates in 10-day windows before and after each mainshock using local earthquake catalog data and uncataloged events detected using a multi-station matched-filter approach on continuous waveform data. The multi-station matched-filter method detects earthquakes by cross-correlating known earthquakes with continuous data and declaring events when correlation values of combined stations exceed a pre-set threshold.

After determining seismicity rates in the 20-day windows surrounding each mainshock, we will use aβ-statistic and p-value to quantify if statistically significant triggering has occurred. Where triggering occurs, calculations of triggered earthquake focal mechanisms may help explain how receiver pre-existing faults become critically stressed, and what physical factors are directly correlated with dynamic triggering. Cases of observed triggering may imply that the seismic response to injection activity could be more intense than in regions without remote dynamic triggering. Alternatively, if triggering occurs but the seismic response to injection activity is limited, it could imply that hydraulic communication with basement faults is key for inducing earthquakes.