Fracking and the injection of waste water from oil extraction into disposal wells are hot topics at the moment. The principal concerns are their possible human health and environmental impacts,particularly "man-made earthquakes." In each of the 33 years between 1967 and 2000, the central and eastern United States experienced an average of about 21 earthquakes of M3.0 or greater. In the 13 years from 2000-2013, during which these methods have become widely used, the number has risen to about 100.
Over the last few decades, worldwide data has suggested that increasing fluid pressure at depth could trigger earthquakes. The argument is that an increase in pore pressure may alter the state of friction and thus reduce the shear strength of pre-existing fractures, bringing them close to failure. The cascading sequence of three 2011 earthquakes (M5.0, 5.7, and 5.0) near the Wilzetta oil field in Oklahoma (after 20 years of injection) are probably the best-known events that fit this description.
An obvious question is: What is the largest possible magnitude we can expect from induced events?
Some seismologists argue that the size of the initial earthquake depends upon the area impacted by the change in pore pressure and the scale of pore pressure change. However, there are others who argue that changing the state of stress on pre-existing faults in the vicinity of injection wells can trigger earthquakes beyond the fluid reach and, under favorable conditions, can lead to larger earthquakes than have been observed so far.
No uniform delay time between the start of an injection process and a local increase in seismicity has been observed. This implies that the state of the stress close to many wells is below the critical state for shear failure. However, the increase in micro seismicity associated with some sites implies the existence of faults close to shear failure.
In such situations, if the size of the pre-existing fault can accommodate larger magnitude earthquakes, an induced earthquake could change the state of Coulomb failure stress on nearby fault segments and trigger larger magnitude earthquakes. This seems to be what happened in Oklahoma in 2011.
So, what is the frequency of induced earthquakes?
Without knowing the details of stress/strain conditions of faults at depth and the nature and scale of pore pressure changes it is impossible to answer this question. For one thing, stress changes at depth due to fluid injection could be transient,depending upon the state of fluid injection at the wells.
The three Oklahoma earthquakes suggest that the state of stress on certain segments of this complex fault system has reached a critical state due to the mixed effects of induced high pressure fluid injection and stress changes from induced earthquakes. A study of these temblors published in 2014 indicates that Coulomb stress changes from the first M5.0 event are consistent with the occurrence of the M5.7 event less than a day later. Accordingly, the study concluded that the volume of fluid injection is not the limiting factor for potential future earthquakes, either in terms of size or numbers.
Our understanding of the connection between high-pressure fluid injection and increased seismicity is still developing, but it is now accepted that fluids can play a major role in affecting the pressures acting on faults. While it is difficult to establish a clear relationship between fluid injection and earthquake activity,there is a growing consensus that the two may be closely related.
On April 21, 2015, the Oklahoma Geological Survey announced that it now "considers it very likely that the majority of recent earthquakes, particularly those in central and north-central Oklahoma, are triggered by the injection of produced water in disposal wells"-a dramatic reversal of its long-held position that they are due to natural causes.
Whatever their cause, induced earthquakes add to risk. How can this hazard be assessed? A recent report by the USGS provides some insight on the sensitivity of regional and local hazard estimates from fracking-related seismicity to different modeling assumptions. However, the report acknowledges the extreme uncertainty in formulating the seismicity rates and does not provide any guidelines on how to integrate the hazard from fracking-related seismicity into long-term earthquake hazard maps. Considering how difficult it is to quantify the frequency with which large earthquakes occur, a practical way of evaluating the induced earthquake hazard may be through scenario studies. These could be done using earthquakes in the AIR stochastic catalog that are as large as M6.7 and within 100 km of the Wilzetta oil field.