Lack of temporal correlation between seismicity and injection in Arkansas, US

Differentiating between the induced (or triggered) and natural seismicity is an important part of understanding the effect of man-made processes. Should we call any seismicity in the vicinity of oil or gas fields induced (or triggered), we would never understand how to mitigate associated risks because we would mix unrelated processes. Temporal correlation is one of the most objective criteria to differentiate the two groups of seismicity we use. Seismic activity induced without proper objective evaluation leads to emotional discussions further reducing our ability to understand and mitigate hazards. Therefore we developed a technique based on temporal cross-correlation (Oprsal and Eisner, 2014) and carefully benchmarked the methodology on well-known cases of induced seismicity. This study shows a proper application of temporal cross-correlation to earthquakes which occurred in Central Arkansas in 2011-2012. The results show a lack of temporal correlation between injection rate and seismicity rates, implying that seismicity may not be triggered by injection at depth (Horton, 2012).