Correlation of Hydraulic Fracturing Induced Seismicity with Operation Parameters of Shale Gas Extraction: Two Case Studies in Western Canada

Abstract:

Northeast British Columbia, specifically the Horn River Basin (HRB) and Montney Trend, are among the largest shale gas production regions in western Canada. In contrast, there has been no large-scale hydraulic fracturing (HF) operation in the Northwest Territories in the Norman Wells region of the central Mackenzie valley. In this study, we investigate the effect of injection pressure, operation duration and injected volume on the observed seismicity in the HRB and Norman Wells regions and compare our observations with the pre-HF records. In the HRB, we apply the single-station location and waveform correlation methods to establish a homogenous earthquake catalog (2006/12-2011/12). In the Northwest Territories, we combine data from a local seismograph network of 4 stations plus a dense array of 7 stations located from 1 km to 50 km from the operation wells to locate earthquakes (2013/09-2014/07). In the HRB, the initial effect of an increased injected volume is an increase in earthquake frequency but not magnitude. Local earthquakes gradually become larger in magnitude as the scale of HF in the region expands. While the injection pressure during HF operations has been regulated at a relatively constant level, the massive increase of injection volume in 2010 and 2011 coincides with a series of M_L>3 events. Relatively large seismic moment release (>10¹⁴ N m) occurred only when the monthly injected volume exceeded ~150,000 m³. In addition, we observe variable time lags, from days to up to 4 months between intense HF and the occurrence of a significant local earthquake. On the other hand, in the Norman Wells region, two small-scale HF were performed in 2014 with a total injected volume of ~ 14000 m³. We observed an increase in the number of micro-earthquakes (M < 2.0) during the HF period without a clear change in the overall seismic pattern. From these two observations, we conclude that HF operations do not necessarily result in an increase in the occurrence rate of larger earthquakes. A clear change in background seismicity pattern towards more frequent occurrence and/or greater magnitudes occurred only when the monthly total injected volume became significantly higher. The hydrologic properties of the source formations and local geologic conditions might also play important roles in induced seismogenesis.