Imagerie de la Terre profonde avec le bruit sismique

Présentation de Michel Campillo (ISTerre) aux 2èmes Rencontres Scientifiques et Techniques Résif | 12-14 octobre 2015, La Grande Motte

Transcript

1. Imagerie de la Terre profonde avec le bruit sismique Michel Campillo (ISTERRE, Grenoble)
2. Body waves in the ambient noise: microseisms (Gutenberg, Vinnik ..) The origin of the noise in the period band 5-10s: Landès et al., 2010 VARIABLE SOURCE LOCATIONS
3. Body wave in high frequency (anthropic) noise Regional scale (N Hz) Local scale (N.100 Hz) Hillers, G., Campillo, M., Lin, Y.-Y., Ma, K.-F., Roux, P., 2012. Anatomy of the high-frequency ambient seismic wave field at the TCDP borehole. J. Geophy. Res. 117 Olivier, G., F. Brenguier, M. Campillo, R. Lynch, and P. Roux, 2015, Body-wave reconstruction from ambient seismic noise correlations in an underground mine: Geophysics, 80 , KS11–KS25.
4. Chaput et al., JGR 2015 Erebus volcano: icequakes
5. 44 ‘large’ events All 3318 events Coda Correlations ZZ correlations: reciprocity holds
6. POLENET/LAPNET array in Finland (RESIF-SISMOB) Comparison of high frequency (1Hz) 1-year noise correlation with earthquake data Poli et al. 2012a Z-Z noise correlations Z comp. actual earthquake Surface wave tomography  body waves (deep reflections) Finland: ‘homogeneous’ Archean crust=> simple seismograms
7. Comparison with synthetic Green functions CZZ (data) GFZZ (theory) Reconstruction of P and S multiple reflections CRR (data) GFRR (theory) Polarisation: noise correlation vs synthetics Good reconstruction of phase and relative amplitudes of the components of the reflected waves. (amplitude discussed by Prieto) A favorable context: distance vs. mean free path, amplitude in actual earthquake records  Deep phases Poli et al., 2012a
8. BODY WAVES: deep reflections/echography Short periods 5-10 s  strong scattering P and PcP Japan to Finland Finland to Japan Standard (surface-wave) pre-processing (Shapiro and Campillo, 2004; Sabra et al. 2005;….) eliminates the contamination by EQ ballistic waves.
9. PP PKP P T ransition zone slow[s/km ] Data 0 50 100 150 200 0.1 0 0.1 time[s] Model 0.1 A C B Upper M antle Station 2 Station 1 Low er M antle 72° 64° 68° Earth’s mantle transition zone discontinuities from ambient seismic noise ( phase transition  (P,T)) From Poli, Campillo, Pedersen and LAPNET WG, Science 2012
10. PP PKP P T ransition zone slow[s/km ] Data 0 50 100 150 200 0.1 0 0.1 time[s] Model 0.1 A C B Upper M antle Station 2 Station 1 Low er M antle 72° 64° 68° Earth’s mantle transition zone discontinuities from ambient seismic noise ( phase transition  (P,T)) In agreement with receiver functions (Alinaghi et al. 2003) From Poli, Campillo, Pedersen and LAPNET WG, Science 2012
11. From Poli, Thomas, Campillo and Pedersen 2014 Core phases PcP and PdP D’’: - different hypotheses for the nature of the layer - PdP difficult to observe - lack of earthquake data
12. From Poli, Thomas, Campillo and Pedersen 2014 Advantage of noise vs earthquake records: -surface to surface -impulsive wavelet -double beam forming Stacked vespagrams for: Earthquakes Noise A 5% increase of velocity at 2530 km depth….
13. Consider now the problem of long period body (P and S) waves at the global scale with noise sources at the surface: -Very weak scattering -A problem of a different nature, although indeed the uneven distribution surface noise sources is still there. This representation is not formally valid on the free surface: the integral vanishes. GF reconstruction would require a more complex procedure (Ruigrok et al., 2008) (Here also, the correlation of multiply scattered waves should lead to the Green function.)
14. GLOBAL TELESEISMIC CORRELATIONS (periods 25-100s vertical components) Boué, Poli et al., GJI 2013
15. Numerous phases can be identified Vertically incident S waves on the vertical component??
16. Long periods (25-100s) Processing: separating EQ and coda from ambient noise Low daily coherence High daily coherence (EQs) AXISEM synthetics High amplitude spurious High coherence=days following large earthquakes
17. From Boué et al., 2014 Long periods (25-100s) Spurious arrivals and numerical simulation CC (ambient noise) CC(EQ days) PREM synth Synth EQ CC Very clear pulses in the correlations, likely holding information about the deep Earth, but should not be interpreted directly as components of the Green function  Attempt to use a representation with a free surface condition boundary
20. Time after EQ + 600 in mn
21. Spectrum ScS© Identification peak: 102.4s: 17S14,16S12, 14S18 (ScS) Vapp =+/-30km/s Corresponding to the ray parameter of a ScS at 12° (3.62s/°) error=13s! Peak at 78.7s error 7s + question of bias due to anisotropy…. Reduces the error! Spectrum P’P’© in blue
22. Focusing pattern at USArray C(0,r)  Jo(kr)  kVapp=29+/-2km/s l=13-15 To be redone in narrow bands…