170 likes | 285 Views
Can we observe a non-shear pattern during 2003 fluid injection at Soultz -sous- Fôrets?. Zuzana Jechumtálová, Jan Šílený Institute of Geophysics, Prague. Motivation. Larger microearthquakes (ML ≥ 1.6) which were induced during and after the 2003 massive fluid injection are
E N D
Can we observe a non-shear pattern during 2003 fluid injection at Soultz-sous-Fôrets? Zuzana Jechumtálová, Jan Šílený Institute of Geophysics, Prague
Motivation • Larger microearthquakes(ML ≥ 1.6) which were induced during and after the 2003 massive fluid injection are occurring onpre-existing faults(Horálek et al., 2010). • nearlypure shear slips • clusteringon two fault segments • injectionfluid pressureremainedbelow the tensile strength of the material • It doesnot a priori excludethe existence oftensile fracturesduring injection. • However, to answer this issue there is necessary to investigate • weaker microearthquakes, whereopeningnewcracksdue to injection • is morerelevant.
Objectives • investigation of weak microearthquakes • significantly contaminated by noise • not detected by all stations • MTs of which need not always be well constrained and • may contain spurious non-DC components • comparison of the source mechanisms resulting from • the alternative approaches • moment tensor • model describing a slip along the fault • with an off-plane slip component • offers a clue to estimate the reliability of the shear vs. • non-shear source components
Outline • Geothermal HDR site at Soultz-sous-Forêts • characteristics of the fluids injection experiment in 2003 • Soultz seismic network • Source mechanism retrieval • previous analysis of mechanisms of 2003 injection • criteria of additional event selection • inversion methods using two alternative source models • resultant source mechanisms and their comparison • Conclusions
The Soultz 2003 injection • massive fluid injection • GPK2 & GPK3stimulated • duration11 days Adapted from Cuénot et al. (2006) • seismicity duration: 20 days • microearthquakes recorded • 5 000 events M - 0.9 • only 240 events M > 1.0 • three largest events M =2.7, 2.8, 2.9 • totalfoci volume: 2km x 2km x1km Adapted from Dorbath et al. (2009)
Soultz seismic network ▼8three-component stations ▲6one-componentstations frequency range of seismometers :1.0 – 40 Hz sampling frequency:150 Hz epicentraldistances:≈ 100m to 7km Fairlyuniform distributionof 14 stationsused for the MT estimationson the focal sphere.
Previous analysis of mechanisms of 2003 injection Horáleket al. (2010) • criteriaof eventselection • the1.6 ≤ ML ≤ 2.9events • coveringthe whole experiment • seismograms havinghigh signal to noise ratio • MTsmeeting these criteria • wellconstrained • stable to noisecontamination • stable tostructuremismodelling • source mechanisms • dominantly pure shear • dip-slip, oblique normal and strike-slip
Additional events processed • criteriaof eventselection • the magnitudeML ≥ 1.4events • from thefirst phaseof the injection in 2003 whenonlythe borehole • GPK3wasstimulated • weak events –not testedyet • consequencesof these criteria • seismograms havinglow signal to noise ratio • eventsnot detectedbyall 14 stationsof Soultz surface network • MTs need notalways be wellconstrained • sparsedatacoverage,noisecontamination and structure • mismodellingmay producespurious non-DC components • inversion of13 eventsusingtwo alternative source models
a + / Inversion methods Moment tensor (MT) • general dipole source Shear-tensile/implosion (STI) • physical source BUT toogeneral shear slip + tensile crack / cavity closure • includes unphysical sources Dufumier & Rivera 1997, Vavryčuk 2001 • 6inversionparameters: M11, M22, M33, M12, M13 a M23 • advantage: • linear inverse problem • disadvantage: • spurious non-DC components • decomposition of MT non-unique • 5inversionparameters: 4 angles (,,,), magnitude (M0) • advantage: • pure physical source • less parameters, i.e. more robust • disadvantage: • non-linear inverse problem
Mechanisms : MT vs. STI histograms of slope angle : focal spheres : 05/28 – 01:42 =0.25º P T DC59% V(I) 6% CLVD(P)35% 121º 38º -147º strike dip rake 112º 46º -147º P T N N ‘confidence zones’ : the NRMS remains below • 120% - dark • 150% - medium • 200% - light colour percentage of the best value traditional fault-plane solution plots : • nodal lines of DC part of MT • principal axes T, P and N • decomposition of the MT • principal axes T, P and N • source lines & the direction of slip / fault normal vector 05/29 – 04:28 =1.25º DC85% V(E)3% CLVD(T)12% 244º 74º -51º strike dip rake 243º 74º -52º T N T N P P Moment tensor Shear-tensile/implosion
Mechanisms : MT vs. STI 05/29 – 03:44 =-1.25º DC96% V(I) 3% CLVD(P)1% 6º 31º -33º strike dip rake 8º 28º -30º 05/29 – 14:25 =3.75º 191º 29º -33º strike dip rake DC74% V(E)11% CLVD(T)15% 190º 29º -35º 05/30 – 18:46 =1.0º DC45% V(E)10% CLVD(P)45% 231º 66º -51º strike dip rake 246º 74º -46º 05/30 – 21:11 =2.25º 239º 72º -60º strike dip rake DC79% V(E)9% CLVD(P)12% 242º 74º -55º
Mechanisms : MT vs. STI 05/31 – 02:23 =4.5º DC74% V(E) 13% CLVD(P)13% 186º 39º -52º strike dip rake 181º 40º -54º 05/31 – 03:29 =0.25º 2º 56º -35º strike dip rake DC52% V(E) 6% CLVD(T)42% 5º 58º -30º 05/31 – 11:33 =-0.25º DC70% V(I) 5% CLVD(P)25% 15º 38º -41º strike dip rake 7º 37º -47º 06/01 – 05:42 =-3.25º 145º 51º -144º strike dip rake DC32% V(I) 4% CLVD(P)64% 162º 41º -155º
Mechanisms : MT vs. STI 06/01 – 11:57 =1.0º DC83% V(E) 9% CLVD(T) 8% 199º 51º -50º strike dip rake 220º 59º -46º 06/01 – 14:27 =-1.25º 138º 49º -173º strike dip rake DC67% V(E) 5% CLVD(P)28% 145º 49º -179º 06/02 – 07:44 =1.75º DC54% V(I) 2% CLVD(P)44% 125º 69º -131º strike dip rake 118º 68º -134º
Mechanisms : MT vs. STI traditional fault-plane solution plots : • equal-area lower-hemisphere projection • nodal lines of DC part of MT source lines of STI • principal axes T - a triangle apex up P - a triangle apex right N - a triangle apex left • compression – colour area dilatation – white area • direction of slip / fault normal, off-plane angle – yellow circle Moment tensor Shear tensile/implosion
Discussion • orientationof double-couple part of MTinvery good agreementwith fault orientation in STI • all13 MTscorrespond to shear-slipSTI models • MTs withnon-DC partsbetween 4% and 68% • STI withslope anglesbetween -3.25º and 4.5º which are not significant non-DC componentsof MTspuriouscaused by sparse data coverage, noise contamination and structure mismodelling • direct parametrizationof shear/non-shear displacement in the STI straightforwardquantitative assessmentof fracture modes
Conclusions investigatingthe 13 inducedevents with ML ≥ 1.4 • source mechanisms • mechanismsdominantly pure shears dip-slip and oblique normal • orientation of T-axes and P-axes stable directions of T-axes (sub-horizontally in E-W direction) variations of P-axes (from vertical to horizontal in N-S direction) • comparisonwith45 largest events(1.6 ≤ ML ≤ 2.9) • all 58mechanisms in agreement • orientation of all T-axes and P-axesin agreementwith thestress pattern from in-situ measurements Even weak microearthquakeswith ML ≥ 1.4 were pure shear slips on pre-existing faults.