Reservoir-Triggered Seismicity in the Canadian Shield as an Analogue to the Activity Induced by the Deep Injection of Wastewater

English Abstract:

There are numerous examples of seismicity induced by wastewater injection. Although the injection of fluids is made within Palaeozoic sedimentary layers, it appears that most injection-induced earthquakes occur along pre-existing faults of the Precambrian basement. In these areas of deep injection, little is known about the distribution of Precambrian faults and their susceptibility to reactivation when subjected to an increase in pore-fluid pressure. An interesting analogue is the impoundment of large hydro-electric reservoirs that triggers earthquakes in the Canadian Shield.

In the Canadian Shield, the Precambrian basement is found beneath every hydro-electric reservoir and lineaments and faults are clearly visible in remote sensing imagery. The history of reservoir-triggered seismicity (RTS) can help define some characteristics of fault reactivation. Out of the tens of reservoirs with water depth greater than 50 m, only a small proportion caused RTS, similar to the very few deep wells that led to seismicity. Also, RTS does not necessarily occur in the deepest reservoirs. For example, the Manic-5 reservoir (the largest in terms of volume and depth; so large that it is visible from space), has not led to RTS.

Lessons learned from four decades of RTS in the Canadian Shield are: 1) Although 50 m of water depth appears to be the minimum, one cannot predict the likelihood of RTS for greater depth. 2) Although the increase in pore-fluid pressure at the base of the reservoir can exceed 1 Mpa; this does not necessarily translate into RTS. 3) Prominent regional-scale faults are not necessarily those reactivated. 4) The few meters of seasonal water level variations are not sufficient to create RTS. 5) RTS is generally within a few km of the reservoir shoreline. 6) Prior to the 1980’s, it is possible that RTS cases with earthquakes of magnitude less than 3 were not detected due to poor monitoring.