Weingarten, Matthew听1听;听Ge, Shemin听2
1听University of Colorado-麻豆官网
2听University of Colorado-麻豆官网
An unprecedented increase in earthquakes in the central and eastern US (CEUS) began in 2009. Many of these earthquakes have been documented as likely induced by wastewater injection. To better understand the likelihood of an induced seismic event from a given injection well, we compare the location and timing of earthquakes and injection operational parameters across the CEUS. We compiled a database of more than 187,000 injection wells in the CEUS, both active and inactive. In combination with the Advanced National Seismic System's (ANSS) comprehensive earthquake catalog from 1973 through December 2014, we use spatial and temporal filtering methods to discriminate injection wells that may be associated with earthquakes from those that are not. Our goal was to understand whether or not well operational parameters such as injection rate, cumulative injected volume, injection pressure and injection depth affect the likelihood that a well is spatiotemporally associated with an earthquake.
We found more than 15,000 wells (~10% of all wells) that may be associated with earthquakes in the CEUS. Our spatiotemporal filter succeeded in capturing every suspected case of induced seismicity that we are aware of. We also found that the likelihood of an injection well being associated with an earthquake was dominantly controlled a well's injection rate. High-rate injection wells, injecting more than 300,000 barrels/month, are preferentially associated with earthquakes. The highest rate injection wells, wells injecting more than 1,000,000 barrels/month, were nearly twice as likely to be associated with earthquakes as wells injecting at lower rates. This phenomenon was observed over a wide range of geologic and hydrogeologic provinces in states such as Oklahoma, Colorado, New Mexico and Arkansas. Operational parameters such as cumulative injected volume, injection pressure and injection depth do not show a clear trend towards an increased likelihood of spatiotemporal association with an earthquake.
In all, we found that, since 2001, the entire rate increase of U.S. mid-continent seismicity was associated with injection wells. The year of 2014 was especially remarkable; 98% (658 of the 671) M3.0+ earthquakes in the U.S. mid-continent were associated with injection wells. Given that injection rate appears to affect the likelihood of wells being associated with earthquakes, industry or regulators can use this parameter to lower the likelihood of earthquakes associated with injection wells.