4D Integrated Study Using Geology, Geophysics, Reservoir Modeling & Rock Mechanics to Develop Assessment Models for Potential Induced Seismicity Risk

Transcription

1 1 4D Integrated Study Using Geology, Geophysics, Reservoir Modeling & Rock Mechanics to Develop Assessment Models for Potential Induced Seismicity Risk Jeremy Boak Oklahoma Geological Survey Best of RPSEA 10 Years of Research - Ultra-Deepwater and Onshore Technology Conference August 30-31, 2016 The San Luis Resort, Spa & Conference Center, Galveston, TX rpsea.org

2 Outline o Project Motivation o Project Elements Task 7: Seismic Monitoring and Network Improvements Task 8: Gravity Infill and Interpretation Task 9: Construction of Integrated Multi-Variate 3D Geologic Model Task 10: Reservoir Modeling Task 11: Integrated Rock Mechanics Study o Technology Transfer 2

3 Earthquakes per Year Project Motivation o Rapid rise in earthquake frequency in Oklahoma : 1-2 M3.0+ earthquakes per year o Earthquakes likely to be induced by oil & gas activity o Basement structure poorly understood in much of Oklahoma Projected M3.0+ Earthquakes

4 Earthquakes per Day Oklahoma M2.8+ Earthquakes M2.8+ Earthquakes 30 Day Mov. Avg. 180 Day Mov. Avg. 0 Jan-11 Jan-12 Jan-13 Jan-14 Jan-15 Jan-16 Jan

5 Earthquakes Alfalfa Grant Woods Noble Garfield Pawnee Logan Payne Lincoln Oklahoma 5

6 Corporation Commission Area of Interest Alfalfa Grant Woods Noble Garfield Pawnee Logan Payne Lincoln Okfuskee Oklahoma Seminole 6 6

7 Comparison of Damage Probabilities 7 7 Source:

8 Number of Earthquakes Earthquakes Occur in Basement, on Optimally Aligned Faults Earthquakes Active Fault Orientations Ave. = 5.4± Maximum Horizontal Stress Depth (km) OGS OF

9 Disposal Rate, Millions of Barrels Per Day (MMBPD) SWD in 684 Arbuckle Wells in Earthquake Area of Interest (AOI) Jan-13 Apr-13 Jul-13 Oct-13 Jan-14 Apr-14 Jul-14 Oct-14 Jan-15 Apr-15 Jul-15 Oct-15 Jan Arbuckle SWD OCC Reductions

10 Average Daily Earthquakes (M2.8+) Earthquake Rate is Declining S Central Central Northwest North Central Rest of OK Jan-10 Jan-11 Jan-12 Jan-13 Jan-14 Jan-15 Jan-16 10

11 Seismic Activity Shifts Regionally Through Time Seismic Moment by Region 400,000 Rest of OK Seismic Moment (Trillion N-m) 300, , ,000 North Central Northwest Central South

12 Galena Township Fault 70% of 2016 Seismic Energy Oklahoma Geological Survey Dr. Jeremy Boak, Director 11 July 2016 M4+ aftershocks in Woods County The recent M4.2, M4.2, and M4.4 that occurred on 2016 July 08 are considered aftershocks of the M5.1 earthquake that happened on 2016 Feb 13. This sequence brings the total of M4 or greater aftershocks on the Galena Township Fault Zone (GTFZ) up to eight. Empirical relations from global observations suggest that a few more aftershocks of M4+ are expected to occur on the GTFZ, as the crust adjusts to the displacement done by the M5.1 mainshock. Clustered aftershocks of similar magnitude are not uncommon in seismically active regions. The duration of aftershock sequences is typically longer in mid-continental settings than along plate boundaries. We derived three focal plane solutions (FPS or beachball diagrams) from each event that indicate strike-slip displacement dominated the motion along the fault. In case of further seismicity, we urge the public to protect themselves from falling objects: (1) drop to the ground; (2) cover your head and neck; and (3) hold on to any sturdy cover. Time: 2016 July 08 21:04:27.5 (CDT) Magnitude: 4.4 Mw Epicenter: , Depth: 7.1 km Time: 2016 July 08 17:29:38.0 (CDT) Magnitude: 4.2 Mw Epicenter: , Depth: 6.4 km Time: 2016 July 08 16:31:57.7 (CDT) Magnitude: 4.2 Mw Epicenter: , Depth: 7.3 km 12

13 Task 7: Seismic Network Enhancement & Monitoring Amberlee Darold, Jefferson Chang 13

14 14 OGS Network Prior to RPSEA Project

15 15 OGS Network After RPSEA Project

16 16 Permanent vs. Temporary Stations

17 Characteristics of Larger Earthquake Clusters Wellston (2013) Prague ( ) 17 17

18 Cumulative Frequency Earthquakes per Day (M2.8+) Prague and Wellston Fault Earthquake Swarms 3 2 Earthquake Frequency (M2.8+) Lincoln Prague Wellston 1 0 Jan-11 Jan-12 Jan-13 Jan-14 Jan-15 Jan-16 Jan Lincoln County Earthquakes Lincoln Prague Wellston Magnitude 18

19 Task 8: Gravity Data Infill and Interpretation Kevin Crain, Stephen Holloway 98 W 97.5 W 97 W 96.5 W 96 W Legend Proposed RPSEA survey area. RPSEA GPS & gravity stations 6/2014 to 5/2016. o 160 field days from 6/2014 to 7/2016 o 3092 unique station locations o Stations were located with 2-mile spacing as a compromise between coverage and density. o Each station has three 1-minute gravity readings when are then averaged. o A total of 9020 unique gravity readings have been reduced, resulting in 3092 gravity stations with accompanying post-processed GPS coordinates

20 New Computational Algorithm for Gravity Gradients o o o o Normalized 3D gravity computation times (modified from Wild-Pfeiffer, 2008). Lengths of the horizontal bars are relative computation times for a single component (Gzz) gravity gradient computation, with varying methods of mass approximation. The width* of the vertical line to the right of the horizontal bars is the computation time for SIGMA (~5 seconds) normalized to the prism computation time (~52 hours). * at this scale, the line width is exaggerated x25 in order to make the line visible. SIGMA point-mass (tesseroid-taylor) tesseroid-taylor mass layer mass line point-mass (prism) Prism 20

21 Task 9: Construction of Integrated Multi- Variate 3D Geologic Model o Structural Geology Stephen Marsh o Potential Fields Gravity & Magnetic Kevin Crain, Stephen Holloway o o Seismic Tomography Chen Chen Hydrogeology Kyle Murray 21

22 22 Geologic Model of Study Area

23 Basement Structure is Complicated Depth (feet) Depth (feet) 5 Arbuckle 3 4 Igneous Basement Pluton Metamorphic Basement 1 6 Earthquakes Earthquakes 2 23 North West Industry collaboration has been essential South East

24 Optimally Oriented Faults in Oklahoma 24 OGS OF Preliminary Oklahoma Optimal Fault Orientations Sources: Literature & OIPA

25 Regional Basement Features Significant KS OK KS OK 25

26 Regional Basement Features Significant KS OK 26

27 27 Payne County Area

28 Development of Geologic Model Preliminary 3D Seismic Top Arbuckle 28 28

29 Seismic Tomography Project 1. Developed an automatic identification system, the PhasePApy, for determining the arrival times of seismic phases, extending previous work. 2. Employed the large number of earthquakes recorded by the OGS seismic network to derive a 3D velocity model of the central Oklahoma area via a tomographic approach. 3. Employed the resulting 3D velocity model to improve earthquake locations thereby advancing the understanding of, and aiding efforts to mitigate the earthquake hazards in the area, on both a regional and local basis. 29 McNamara et al., (2015) inferred a WNW-ESE fault based on their earthquake relocation study for the southern Cushing cluster and made a series of analyses. However, our relocation result shows NE-SW trending orientations for it, which is different from McNamara et al., (2015).

30 Underground Injection Control (UIC) Class II Injection Salt Water Disposal (SWD) Wells (2013) Murray 2014, OGS OF

31 Underground Injection Control (UIC) Class II Injection Salt Water Disposal (SWD) Wells (2014) 31 Murray 2014, OGS OF5-2015

32 Most SWD occurs in Arbuckle Group, and in the AOI Salt Water Disposal (Million Barrels/Year) Salt Water Disposal (Million Barrels/Year) Arbuckle,Basement Dev-mid-Ordovician Mississippian, Woodford Atokan-Morrowan Desmoinesian Missourian Virgilian Permian Multiple, Other, Unspec AOI Disposal Non-AOI Disposal Source: Murray, K. (2015) Class II Saltwater Disposal for at the Annual-, State-, and County- Scales by Geologic Zones of Completion, Oklahoma, Oklahoma Geological Survey Open File Report OGS OF5-2015

33 Distance North (m) Fourier Component Task 10: Reservoir Modeling (Deepak Devegowda) o Objectives Interpretation of cross-well tests Identify preferential flow paths PRIOR GEOLOGIC MODEL Locate flow barriers/baffles 4D Reservoir Characterization FLOW SIMULATION Flow modeling and estimating stress/ pore pressure changes Prediction of seismicity and reconciling observations by updating geomechanical properties COMPUTE DATA MISFIT IN FREQUENCY DOMAIN HISTORY MATCH YES OUTPUT MODEL Frequency (milli-hz) 10 NO 4000 PERFORM INVERSION Well 1 Well 4 0 Distance East (m) Distance East (m) Sensitivity (Pa/md) 20

34 Task 11: Integrated Rock Mechanics Study Ahmad Ghassemi 34 o Simulation of effects of water injection near Wilzetta Fault 2D modeling 3D fully coupled finite element model; relatively simple geologic model o FLAC3D partially coupled model with more complete structural features & heterogeneity o Size & area of investigation chosen based on: Location of Wilzetta faults Location of large seismic events of Prague sequence Location of high volume injection wells in epicentral region. Area modeled is 5,940 m wide and 8,197 m long with a thickness of 911 m.

35 35 3-D finite Element Model

36 Task 11: Integrated Rock Mechanics Study 36 Some experiments are ongoing to measure geomechanical properties of the Arbuckle and other formations of interest.

37 Contacts Principal Investigator: Jeremy Boak Oklahoma Geological Survey Project Manager: Joe Renk NETL Technical Coordinator: Kent Perry RPSEA

38 Backup Slides Backup Material 38

39 Earthquake Activity Occurs Where Deep Injection Increases Substantially 39 Source: Walsh, F. R., and Zoback, M. D. (2015) Oklahoma s recent earthquakes and saltwater disposal. Sci. Adv. 2015; 1:e , 18 June 2015

40 Areas of Lower SWD Show Lower Earthquake Activity 40 Source: Walsh, F. R., and Zoback, M. D. (2015) Oklahoma s recent earthquakes and saltwater disposal. Sci. Adv. 2015; 1:e , 18 June 2015

41 SWD Includes Very Little Flowback Water from Hydraulic Fracturing 41 Source: Walsh, F. R., and Zoback, M. D. (2015) Oklahoma s recent earthquakes and saltwater disposal. Sci. Adv. 2015; 1:e , 18 June 2015

42 Cumulative Seismic Moment (Energy) by Year Cumulative Seismic Moment (Trillion N-m) 1.0E E E E E E J F M A M J J A S O N D J 42 42

43 Seismic Moment (Energy) Focus Shifts through Time 400,000 Seismic Moment by Region Seismic Moment (Trillion N-m) 300, , ,000 Rest of OK North Central Northwest Central South

44 M3.0+ Earthquakes M3.5+ Earthquakes Earthquake Year-to-Year Comparison /23/ /23/ J F M A M J J A S O N D J 0 J F M A M J J A S O N D J F 44

45 Earthquakes Alfalfa Grant Woods Noble Garfield Payne Logan Lincoln Oklahoma 45

46 Latitude Earthquakes per Day (M2.8+) Prague and Wellston Faults Lincoln County Earthquakes Earthquake Frequency (M2.8+) 36 Wellston Prague Lincoln 3 Lincoln Prague Wellston Longitude 0 Jan-11 Jan-12 Jan-13 Jan-14 Jan-15 Jan-16 Jan

47 Cumulative Frequency Prague and Wellston Faults 1000 Lincoln County Earthquakes Lincoln Prague Wellston Magnitude 47

48 Latitude Earthquakes per Day (M2.8+) Galena Township Fault Location Woods & Major Galena Twp Fault 3 Earthquake Frequency Woods & Major Galena Township 30 Day Average Day Average Longitude 0 Jan-13 Jan-14 Jan-15 Jan-16 Jan

49 Cumulative Frequency Galena Township Fault 1000 Gutenberg-Richter Plot Woods & Major Galena Twp Fault Magnitude (M L ) 49

50 Latitude Earthquakes per Month (M2.8+) Payne County Fault Systems 36.3 Payne County Cushing 2015 Cushing 2014 New Fault Blackwell Towns Earthquake Frequency (M2.8+) Payne County Cushing 2015 Cushing 2014 Blackwell Coyle Stillwater Yale Coyle Cushing Langston Perkins Longitude 0 1-Jan-13 1-Jan-14 1-Jan-15 1-Jan-16 50

51 Cumulative Frequency Payne County Fault Systems Gutenberg-Richter Plot Payne County Cushing 2015 Cushing 2014 Coyle Blackwell Magnitude (ML) 51

52 Preliminary 3D Seismic Basement 52 52

53 53 Basement Geology Model

54 OF3-2015: Preliminary Fault Map of Oklahoma RON 6 OKLAHOMA GEOLOGICAL SURVEY Richard D. Andrews, Interim Director 1E 2 CIMAR TEXAS BEAVER N 29 N 27 28E N W HARPER 24 ELLIS 25 ROGER MIL LS WOODS DEWEY KIOWA WASHITA CUSTER WOODWARD JACKSON GREER BECKHAM 25W HARMON 26W N TILLMAN ALFALFA 11 BLAINE MAJOR 13 CADDO 12 COMANCHE 13 COTTON GRANT LOGAN 3 1W 29 N E KAY E NOBLE 1W OKLAHOMA 1W N 2 GARVIN 2 1S 2 3 1E LOVE CARTER PAWNEE PAYNE CREEK OSAGE COAL BRYAN TULSA NOWATA WAGONER ROGERS E N ADAIR DELAWARE OTTAWA 22 OPEN-F ILE REPORT OF Prelim in ary Fault Map of Okla homa CRAIG KILOMETERS N 1S E 36 26E LE FLORE S MCCURTAIN MILES 22 SEQUOYAH CHEROKEE MAYES LATIMER 30 SCALE1: HASKELL MUSKOGEE PUSHMATAHA 10 CHOCTAW MCINTOSH 15 PITTSBURG 13 ATOKA 12 OKMULGEE 11 HUGHES OKFUSKEE SEMINOLE 6 PONTOTOC 5 JOHNSTON MARSHALL 4 LINCOLN 3 MURRAY CLEVELAND 3 MCCLAIN 4 GARFIELD 6 6 JEFFERSON STEPHENS 7 GRADY CANADIAN 8 KINGFISHER N O T G N I H S A W Inset Map showin g all faults used to compile the preliminary fault map. 3,418 fault segments from industry contributions and more than 6,000 fault segments fr om the published literature were used. The complicated representation for some major fault s is due to the large number of maps showing the same faults with each map showing a dif erent location. Additionally, fa ult intricacy may represent true fault complexity. Summary This preliminary fault map is compiled from oil and gas in dustry data and published lit erature. Most of the fault s from the published lit erature are documented in the supplement to this Open-File Report. This map is compiled fr om the Oklahoma fault database, which continues to be updated as addit ional fault in formation is available. This map is bein g released in a timely manner and should be considered preliminary. The fault s shown here represent one interpretation of all the fault s in the Oklahoma fault database bein g compiled by the OGS. This preliminary version identifies surface and subsurface fault s on one map. Interpretation and Compilation Methods This fault map does not show addit ional information associated wit h the fault database such as the formation the fault was mapped in, method used to constrain fault location, etc. Faults mapped at the surface were given highest precedence and most were included on this map. Fault segments that are mapped close to each other are interpreted as a sin gle fault. Faults mapped using 3D seismic were given precedence over faults constrained from 2D seismic data, and subsurface faults not constrained by seismic data were given the lowest precedence. True fault complexity in some areas may be misrepresented by this effort of producing sin gle fault locations for mapped faults. In areas with complex fault s, individual fault locations are uncertain and may be in error by as much as several kilometers. Some questionable faults and small parallel faults are omit ted. Future efforts wil document the choices made in developing a fin al fault map of Oklahoma. 3,418 fault segments from in dustry contributions and more than 6,000 fault segments fr om the published lit erature were used to make this map. All these fault segments are shown on Inset Map and give an indication of where faulting is most complex. Future efforts wil make more metadata available and wil include such in formation as the fault slip-sense and dip direction. Acknowledgements We would like to acknowledge the Oklahoma Independent Petroleum Association for it s leadership efforts in makin g the fault database contributions from industry possible. In addit ion, I would like to thank all those who have worked and continue to work on the Oklahoma Fault Database inclu din g Stephen Marsh, Stephen Holloway, Dr. Randy Keller, Russell Standridge, Amberlee Darold, and Jacob Hernandez. This work was partial y fu nded by Research Partnership to Secure Energy for America project number Open-file Report Disclaimer Open-File Reports are used for the dissemination of information that fil s a public need and are in tended to make the result s of research available at the earliest possible date. Because of their nature and possibility of being superseded, an Open-File Report is intended as a preliminary report not as a final publication. Analyses presented in this article are based on information available to the author, and do not necessarily represent the views of the Oklahoma Geological Survey, the Universit y of Oklahoma, their employees, or the State of Oklahoma. The accuracy of the in formation contained herein is not guaranteed and any mention of trade names are not an endorsement by the author, the Oklahoma Geological Survey, or the Universit y of Oklahoma. PRELIMINARYFAULTMAP OFOKLAHOMA By Austin A. Holland April 21, E I M O T A W A T T O P

55 Earthquakes per Day Daily Injection/Production (Millions of barrels) Earthquakes, Oil & Water M2.8+ Earthquakes 684 Arbuckle Well Injection 14 County Injection Oklahoma Oil Production Jan-11 Jan-12 Jan-13 Jan-14 Jan-15 Jan-16 Jan-17 55

56 Daily Injection/Production (MMB) Earthquakes, Oil and Water Daily Seismic Moment (Trillion N-m) Daily Seismic Moment 684 Arbuckle Well Injection 14 County Injection Oklahoma Oil Production Jan-11 Jan-12 Jan-13 Jan-14 Jan-15 Jan-16 Jan

57 Average Daily Seismic Moment (Trillion N-m) Seismic Energy is Declining 2,500 2,000 1,500 S Central Central Northwest North Central Rest of OK 1, Jan-10 Jan-11 Jan-12 Jan-13 Jan-14 Jan-15 Jan

58 Number of Days A Change in the Seismic Weather >1

59 Earthquake Forecasting November 2015 Magnitude (M L ) Duration Year Year months days days Probability of one or more earthquakes of magnitude (m) over the specified time Probabilities expressed as values from 0 to 1. To transform probabilities to percent, multiply by 100. A forecast, not a prediction Shorter the period, the more uncertain the estimate

60 Earthquake Forecasting August 2016 Magnitude (M L ) Duration Year Year months days days Probability of one or more earthquakes of magnitude (m) over the specified time Probabilities expressed as values from 0 to 1. To transform probabilities to percent, multiply by 100. A forecast, not a prediction Shorter the period, the more uncertain the estimate

61 61 Residual Free-air Gravity Anomaly