Temporal Relationship Between Injection Rates and Induced Seismicity

Normalized cross-correlation of effective functions was originally developed to differentiate induced and natural seismicity. We apply this methodology to data sets acquired during hydraulic stimulation of shale gas to study temporal relationship between seismicity and the injection. Delays of seismicity induced by hydraulic fracturing are compared with delays of seismicity related to changes of the water level in a dam. A new observation is made; we show that seismicity induced by hydraulic fracturing and measured by the normalized cross-correlation can have very positive to even short negative delay after the injection. We explain this observation by rapid decline of induced seismicity after decrease of injection. On the contrary, we show that the seismicity triggered by water impounding in a dam is characterized by longer delays due to much greater distance between lake and the induced seismicity. We show how it is possible to measure the delay by the cross-correlation methodology despite periodicity of the signals.