How should we react to induced seismicity? Accounting for natural uncertainty within traffic light systems

Abstract:

Shale gas and geothermal power have become important energy sources worldwide, and the global demand for mined resources continues to rise. However, public concerns over risks of earthquakes in the vicinity of operational sites has increased. Traffic light systems (TLS) are often used to control our reactions to induced seismicity, and to reduce the probability of larger future events. Under a TLS, operations must stop if an earthquake with a local magnitude M_L above a given threshold is detected. Assessing accurate microseismic local magnitudes is challenging and differences in M_L estimates for the same event may be larger than a whole magnitude unit. This may make the difference between a decision to continue carefully (“amber” TLS zone) and an immediate stop of operations (“red” zone). In addition, site-specific uncertainties are both difficult to quantify and to include in the decision-making process, so how we should react in itself becomes uncertain. Some operations will incorrectly continue, increasing the risk of larger triggered earthquakes, while others will be wrongly halted.

Here we show how to calculate local event magnitudes with realistic uncertainties. Results show clearly that velocity uncertainties and station site effects in M_L have similar magnitude, and cannot be neglected in TLS systems - otherwise the risk of larger future events remains unquantified. We demonstrate how to include these uncertainties in the choice of TLS thresholds such that the certainty of an earthquake being in either TLS zone is above any level with which a regulator is comfortable. We apply the new TLS both to mining-induced seismicity, and to the hydraulic fracturing induced seismicity at Preston New Road, UK, where a deterministic monitoring approach and unaccounted-for uncertainties led eventually to earthquakes with local magnitudes that resulted in the immediate cessation of fracking across the UK. We demonstrate that a red-light threshold would have been encountered earlier if uncertainty had been accounted for in the TLS, which possibly could have prevented the UK-wide shut-down. We argue that it is critical to establish systems which permit regulators to manage risk while accounting for uncertainties.