MITIGATING SEISMICITY IN AZLE AND THE GREATER NORTH TEXAS AREA 1 M a tth e w J. H o r n b a c1 h 1 n H e a th e r D e S h o Br i a n S tu mp 1 Chr is Hay war d Be a tr i c e M a g n a n2 i C l i ff Fr o h l i c3h Jo n O l s o n 4 Bi l l E l l s w o r th and the N O R th Te X a s 1, 2,3,4 s e i s mi c i ty g r o u p 1 Southern Methodist University Dept. of Earth Sciences Dallas, Texas 2 The University of Texas Institute for Geophysics Austin, Texas 3 The University of Texas Dept. of Petroleum and Geosystems Engineering Austin, Texas 4 United States Geological Survey Menlo Park, Ca EARTHQUAKE RISK HAS INCREASED SUBSTANTIALLY IN NORTH TEXAS SINCE LATE 2008 Recent increase in Texas seismicity (Most occur in the Fort Worth Basin) For 2015, Texas seismicity is on track to be a factor of ~20 greater than historic levels. Incorporating Induced Seismicity in the 2014 United States National Seismic Hazard Model – Results of 2014 Workshop and Sensitivity Studies Pubs.usgs.gov/of/2015/1070/ “Seismicity Caused by or Likely Related to Hum an A ctivity” NRC, 2012 Little Linkage Between Hydraulic Fracturing and Felt Earthquakes INJECTION-INDUCED SEISMICIT Y: A WELL ESTABLISHED PHENOMENA  "Although only a very small fraction of injection and extraction activities at hundreds of thousands of energy development sites in the United States have induced seismicity at levels that are noticeable to the public" NRC, 2012  Studies on this topic in Texas date to 1918  Understanding the physical mechanisms remains an open question requiring data and collaboration from government, industry, and other subject-matter experts MECHANISMS FOR INDUCING EARTHQUAKES Changes in solid stress due to fluid extraction or injection (poro-thermoelastic effects, Direct fluid pressure changes in gravitational loading) effects of injection (fluid pressure diffusion) Permeable reservoir/aquifer I Volume and/or mass change Well Fault Increase in pore pressure along fault (requires. Change in loading Permeable high-permeability conditions on fault reseryoir/ pathway) (no direct hydrologic aquifer connection required) Fig. 3. Schematic diagram of mechanisms for inducing earthquakes. Earthquakes may be in- duced by increasing the pore pressure acting on a fault (left) or by changing the shear and normal stress acting on the fault (nght). See (4). Ellsworth, 2013 Did Injection Trigger Earthquakes? The 7 Question Approach Outlined in NRC Repor t (from Davis and Frohlich, 1993) 1. Are the events the first known earthquakes of this character in the region? 2. Is there a clear correlation between injection and seismicity? 3. Are epicenters within 5 km of wells? 4. Do some earthquakes occur at or near injection depth? 5. Are there known geologic structures that may channel flow to sites of earthquakes? 6. Are changes in fluid pressure at well bottoms sufficient to encourage seismicity? 7. Are changes in fluid pressure at hypocentral distances sufficient to encourage seismicity? A Score of 6 or greater = likely (RMA scored a 6) A Score of 3-5 = possible-to-plausible A Score of 2 or less = unlikely What data are helpful in addressing these questions? 1. ARE T HE EVENT S T HE FIRST KNOW N EART HQ UAKES OF T HIS CHARACT ER IN T HE REGION? Use ful data  Instrume nt-Re co rde d Earthquak e s.  Pre -Instrume ntatio n Earthquak e s (Fe lt Re po rts).  S ur face M aps o f Quate rnar y De fo rmatio n (ge o lo gic maps).  S e ismic Image s Indicating Quate rnar y De fo rmatio n. USGS Quaternary Fault Maps Quaternary deformation along the Meeman‐Shelby Fault near Memphis, Tennessee, imaged by high‐resolution marine and land seismic reflection profiles (Hao et al., 2013) 2. IS T HERE A CLEAR CORRELAT IO N BET W EEN INJ ECT ION AND SEISMICIT Y? Example: Rocky Mountain Arsenal (1) Prior to injection, the area was not seismically active. (2) The seismicity generally mimics the injection pattern, but not perfectly. (3) Aftershocks in the region continued following injection (including after attempts to depressurize the reservoir). (4) Largest EQ (M5) occurred year after injection stopped. (from Hesiah & Bredehoeft, 1981; NRC Report, 2012) Re quire d Data  We ll-co nstraine d inje ctio n v o lume s and pre ssure s.  Highe r-re so lutio n (<1 k m re so lutio n, M3 >160 recorded Irving Azle DFW Airport 5 Temp. Networks: DFW Airport (2008-) Cleburne (2009-) Venus (2011-) Azle (2013-) Irving (2014-) Venus Cleburne Earthquakes Report by National Earthquake Information Center since 2008 (2.0 – 4.0) AZLE EVENT LOCATIONS THROUGH 26 AUG, 2014  The last widely felt event was Jan 28 th, 2014  Last EQ recorded in May 2015  Complex faulting  The EQ sequences slowed as injection volumes reduced Hornbach, DeShon, et al., 2015, Nature Communications CAUSAL FACTORS • Natural Tectonic Stress Changes • Ground Water Changes • Lake Level Changes • Industry Activity • SWD Injection • Brine Production Hornbach et al., 2015, Nature Comm. IT IS IMPROBABLE THAT THE AZ LE EARTHQUAKES ARE TRIGGERED NATU RALLY 1. During the past 150 years of settlement, there had been no reported felt earthquakes in the Azle/Reno area prior to November, 2013. 2. There is no clear evidence for fault surface expressions indicative of large-scale active faulting in the region. 3. Publicly available regional seismic data show no significant fault offsets in sediment deposited more than ~300 million years ago in the Fort Worth Basin. Additionally, Gutenburg-Richter Law Modeling suggest we should observe significant (~35 m) offset with depth if these faults have a M3 event only once every 10,000 years. 4. The seismicity pattern in Azle is not consistent with the typical foreshock-mainshock-aftershock sequence observed in most tectonic earthquake sequences, but is consistent with earthquake swarm patterns often associated with induced seismicity. FAULT OF FSET MODELING Assumptions: • Uniformly accumulated displacement since the Pennsylvanian (300Ma). • M max 5.6 (based on current seismicity) • Longest return intervals ~100,000 years Displacement calculations– three cases: 1. The Azle sequence displaced the fault by ~1.2mm. 2. A G-R sequence with single Mmax of 5.0 would slip the fault ~48mm. 3. A G-R sequence with single fault-filling M5.6 would slip the fault ~380mm. T ECTONIC OF FSET MODELED VS. OB SERVED Results: 1. One Azle seq. ~4m offset. 2. One Mmax 5.0 seq. ~140m offset. 3. One Mmax 5.6 seq. ~1200m offset. Khatiwada et al., 2013 CAUSAL FACTORS • Natural Tectonic Stress Changes • Ground Water Changes Unlikely. The region has been tectonically inactive for >200 million years • Lake Level Changes • Industry Activity • SWD Injection • Brine Production Hornbach et al., 2015, Nature Comm. CAUSAL FACTORS • Natural Tectonic Stress Changes • Ground Water Changes • Lake Level No anomalous water levels <1 kPa (<0.1 psi) Changeson the fault • Industry Activity • SWD Injection • Brine Production Hornbach et al., 2015, Nature Comm. QUANTIFYING SUBSURFACE INJECTION/PRODUCTION PRESSURES • Pressure modeling indicate injection/production caused pressure changes sufficient to trigger earthquakes. • pressure changes associated with drought are likely orders of magnitude lower • Faults near Azle/Reno area though historically inactive, appear near-critically stressed. • Currently, industry activities appear to represent the largest quantifiable stress driver on the fault system. AZLE EARTHQUAKES: INDUCED OR NATURAL? (Davis and Frohlich, 1993) Azle Answers 1. Are the events the first known earthquakes of this character in the region? YES 2. Is there a clear correlation between injection and seismicity? YES 3. Are epicenters within 5 km of wells? YES 4. Do some earthquakes occur at or near injection depth? YES 5. Are there known geologic structures that may channel flow to sites of earthquakes? YES 6. Are changes in fluid pressure at well bottoms sufficient to encourage seismicity? YES 7. Are changes in fluid pressure at hypocentral distances sufficient to encourage seismicity? YES Conclusion: It is likely that industry activity triggered the Azle/Reno EQs. PATH FORWARD NRC , 2012 “Current models employed to understand the predictability of the size and location of earthquakes through time in response to net fluid injection or withdrawal require calibration from data from field observations.” “The success of these models is compromised in large part due to the lack of basic data at most locations on the interactions among rock, faults, and fluid as a complex system.” BASIC DATA NEEDS (AS ALREADY OUTLINED IN THE AZLE STUDY)  Better Regional seismic data (TEXNET could improve this)  High quality, local seismic networks (TEXNET could improve this)  Bottom hole pressure and permeability measurements.  Brine production data and brine sources (geochemical data).  Better control on local subsurface structure.  Fault properties  In-situ stresses SUCCESSFUL EXAMPLES OF MITIGATION INVOLVE BETTER MONITORING AND MORE ACCESS TO DATA PARADOX VALLEY, COLORADO  BR adjusts injection strategies, to ma n a g e B o t t o m h o l e p r e s s u r e .  E Q s w a r m mo n i t o r i n g c o mb i n e d w i t h d o w n h o l e p r e s s u r e mo n i t o r i n g p r o v i d e s i n v a l u a b l e t o o l f o r mi t i g a t i n g h a z a r d a n d ma n a g i n g r i s k .  R e d u c i n g i n j e c t i o n v o l u me s / p r e s s u r e s r e d u c e d b o t t o m- h o l e p r e s s u r e s , w h i c h r e d u c e d e a r t h q u a k e s ( s i mi l a r t o w h a t we o bs e r v e i n Azl e ).  After changing injection strategies, r e d u c i n g i n j e c t i o n v o l u me : - - - f e l t s e i s mi c i t y i s r e d u c e d w i t h t i me . - - - e v e n t s s p r e a d s mo r e t h a n 8 k m a wa y ( a s s t r e s s d i f f u s i o n mo d e l s p r e d i c t ) . - - - b i g e v e n t s s t i l l o c c u r ( Li k e R M A ) .  C o n s t r a i n i n g “ a c c e p t a b l e ” s e i s mi c i t y r e q u i r e s h i g h q u a l i t y s e i s mi c / p r e s s u r e d a t a a n d a d e t a i l e d r i s k a n a l ys i s . Injection tests High Volume Injection (From Block et al., 2013) Injection at lower volumes/pressures