Seismic imaging using earthquakes and

implications for earth systems Huaiyu Yuan123*, Mike Dentith2, Ruth Murdie3, Simon Johnson3 and Klaus Gessner3

* Currently in a week-long meeting in San Francisco. 1 2 3 [email protected] CCFS, Macquarie University; CET-UWA; GSWA Email [email protected] for questions. Or spend a good 15 mintues reading it through!

1. control of the mineral systems and seismic imaging 6. Capricorn passive source project

How giant magma-related ore systems form is vigorously debated. One view, displayed below, argues ascending magmas pick up ore-forming compo- The Capricorn orogen is a major tectonic unit that recorded the assembly of the Yil- nents (diamonds, gold) during their passage through the lithosphere (Griffin et al., Nature Geoscience, 2013): e.g. garn and Pilbara cratons and the Proterozoic terranes to form the West Australian (a) diamonds, formed deep from metasomatically introduced carbon zones, are brought to surface in magmas that take advantage of lithospheric scale craton (a; e.g. Johnson et al. 2013). Numerous mineral deposit types have been rec- weak zones; and ognized throughout the orogen, and recent studies (Johnson et al. 2013; Aitken et (b) Au- and Cu-rich deposits are found in the back-arc and the mantle wedge, which are associated with low-degree and hi-degree melting, respectively. al. 2013) have illustrated the connection between known mineral deposits and the crustal scale fault systems and corresponding crustal blocks (b). In all cases lithospheric scale weak zones facilitate the movement of metal-bearing fluids to the surface. Therefore 3D structure of the lithosphere helps to focus deposition of the ore, and entire lithosphere architecture through seismic imaging is preferably needed. We illustrate below how seismic imag- An exciting seismic array of our own, the Capricorn passive source project is funded ing, mostly using earthquakes, can be utilized for understanding craton lithosphere tectonics and also for the purpose of mineral exploration. to cover the Capricorn area. This passive source project is part of a multidisciplinary (a) We focus in Sections 2 to 5 on imaging techniques that are at various resolution scales: from continent wide (500-km lateral and 20-km vertical resolu- project, “Distal footprints of giant ore systems: Capricorn case study”, which aims tion) surface wave tomography, to local fault scale (10s meters both later and vertical resolution) ambient noise imaging. These techniques will be ap- to characterize the Capricorn Orogen from the to mantle using approaches in (a) Elements of the Capricorn Orogen and surrounding cratons and basins, showing the location of plied to the Capricorn passive source deployment (Section 6), WA’s first ever large scale earthquake seismic deployment in Capriconn region which is geology and geochemistry and geophysics. The overall goal of the seismic project is the Capricorn Orogen seismic transect. Modified from Martin & Thorne (2004). The thick grey lines to delimit the northern and southern boundaries of the Capricorn Orogen. Inset shows location of the going to commence early next year. produce 3D multiple scale seismic images across the orogen, which will constrain Capricorn Orogen and Paleoproterozoic crustal elements (KC, Kimberley Craton; NAC, North Australian geological models for the timing and kinematic evolution of faults and shear zones Craton; SAC, South Australian Craton;WAC,West Australian Craton) and Archean cratons (YC, Yilgarn in Craton; PC, Pilbara Craton; GC, Gawler Craton). Other abbreviations: MI, Marymia Inlier; YGC, Yarlar- the region and its 4D metallogenic history. weelor Gneiss Complex. Figure and captions from Johnson et al. 2013. Red wiggle lines are active source lines, who results are summarized in b). Interactions between magmas and the SCLM. Overall there will be 64 broadband seismographs covering a500x500 km region of a, Plume triggers kimberlite formation and lows to area of thinner SCLM the orogen (c), providing unparalleled coverage in the region for the earthquake where melting is focused. Variable interaction of melts with crust and SCLM (b) influences Ni-Cu and PGE deposit genesis. b, Generalized convergent-margin seismology to uncover the lithosphere architecture: What is the spatial extent of setting. Au-poor magmatic-related deposits form from dominantly astheno- spheric or crustal melts (such as Cu-rich or W Sn porphyry, respectively). the major lithosphere domains? Where are their boundaries? Are the boundaries Low-degree melting of , particularly in retro-arc settings, can linked to the crustal faults? How do they fit the mineral deposits on the surface? produce Au-rich metasomatic refertilization of the SCLM. Subsequent melt- ing (which may be much later) contributes Au to magmatic systems, forming Are there signs of wedge tectonics and polarity reversal? (b) Summary cross-section through the Capricorn Orogen, showing the spatial relationship between deposits of porphyry Cu-Au, epithermal Au, iron oxide Cu-Au, intrusive-relat- known mineral deposits and the location of suture zones and other major lithospheric-scale faults. ed orogenic Au, and possibly also Carlin-type Au and classic orogenic Au. Figure and captions taken from Griffin et al. 2013. This first ever orogen-scale 2D seismic array is going to roll in early 2014 (d). The first batch of results will be available a year from now. The three-year deployment will provide excellent opportunity for thesis projects for graduate students. Reference cited: Griffin, W.L., Begg, G.C., O'Reilly, S.Y., 2013. Continental-root control on the genesis of magmatic ore deposits. Nature Geosci 6, 905-910.

2. Layer-cake craton in the North American continent (d)

A surface waveform tomography of the North American continent shows provocative craton scale layered lithosphere structure (Yuan and Romanowicz, Nature 2010), owning to the rich (c) data coverage provided by the USArray trans- (c) The Capricorn passive source project. The deployment is composed of 3 parts: 1) a 8-station back- portable array deployment, a 400-seismometer bone network (red squares) that will cover the whole orogen for 3-years to provide overall control for broadband deployment sweeping the continen- the path coverage; 2) a 26-station short-term deployment (red dimands) will cover the western side of the orogen for the first 1.5 years to magnify the local coverage; and 3) another 26-site short-term

tal US in the past 10 years: Section 4 deployment in the eastern side, using pulling out the sites in part 2 and re-deploying in the east. Grey sites are Geoscience Australia’s permanent sites (gray boxes) and Australian National Universities earli- Section 3 er temporary deployment site, which are taken into consideration to maximize the area coverage. Inset (a) Craton lithosphere is seismically fast above shows the location of the Capricorn deployment. 250 km; References cited: Aitken, A.R.A., Joly, A., dentith, M.C., Johnson, S.P., Thorne, A.M., Tyler, I.M., 2013. 3D architecture, structural (d) The Capricorn passive source project field leaflet. Site permitting is currently undergoing with the evolution and mineral prospectivity of the Gascoyne Province. Geological Survey of Western Australia Report 123. Section 5 (b) Suture zones that welded Archean blocks are cooperation from the Geological Survey of Western Australia. The leaflet shows the passive source site Johnson, S.P., Thorne, A.M., Tyler, I.M., Korsch, R.J., Kennett, B.L.N., Cutten, H.N., Goodwin, J., Blay, O., Blewett, design adopted from the ANU’s earlier passive source deployment. (Photo courtesy of Professor Brian R.S., Joly, A., Dentith, M.C., Aitken, A.R.A., Holzschuh, J., Salmon, M., Reading, A., Heinson, G., Boren, G., Ross, J., characteristic of Vsv (vertically polarized shear Kennett from the ANU). Costelloe, R.D., Fomin, T., 2013. Crustal architecture of the Capricorn Orogen, Western Australia and associated wave) faster than Vsh (horizontally polarized metallogeny. Australian Journal of Earth Sciences 60, 681-705. shear wave) in the craton lithosphere (red dented into blue in b; see map for suture zone locations); 5. Long Beach has its own fault (c) A layer-cake like anisotropy structure is pres- Ambient noise can be pushed even further to reach a high resolution comparable to (a) ent (Layer 1, 2 and Asthenosphere). The asthe- active source seismic studies: tens of meters both vertically and horizontally. This exam- nosphere possess a direction of fast velocity ple (Lin et al. Geophysics 2013) shows how to use over 5000 geophones in a 8x6 km area symmetry axis that is parallel to the current ab- Layer 1 (f)(f) in Long Beach, California to image the Newport-Inglewood fault zone in the shal- solute plate motion (red in asthenosphere in c), Layer 2 Layer 1 low-most upper crust (above 1km depth). reflecting rock deformation in response to the asthenospheric flow. Layer 1 and 2 are two Asthenosphere Layer 2 (b) (c) lithospheric layers of frozen-in anisotropy, with layer 1 fast axis direction parallel to the suture strikes which are coincidentally parallel to the plate motion for the transect (red in layer 1 in (a) The array configuration and the regional fault lines in c). (e)(e) Southern California. The small circles show the 5204 stations used in this study. From Lin et al. 2013. Layer 1

Layer 1 is interpreted as the old “scums” in the (d)(d) Layer 2 References cited: early Earth history when cratonic blocks first Lin, F.-C., Li, D., Clayton, R.W., Hollis, D., 2013. High-resolution 3D Asthenosphere shallow crustal structure in Long Beach, California: Application of formed in the vigorously convecting mantle, ambient noise tomography on a dense seismic array. Geophysics 78, (b) Inverted 3D shear velocity model at 100-, 300-, and 650-m depths. Black lines show the Newport-Inglewood fault system on the surface. Layer 1 Q45-Q56. which is remarkably in spatial agreement with The three north–south cross sections used in (c) are shown at 100-mdepth. (c) The three cross sections of the inverted 3D shear velocity model. the highly depleted nature of chemical layer Layer 2 found in xenolith themal-barrometric analysis (d and e; Griffin et al., Lithos 2004). Lithospheric Layering in the North American Craton. a), depth cross-section of isotropic shear velocity variation along AA’ (see map for location). Positive velocity variation is in blue; negative in red. b), radial Layer 2 is the metasomatized lower lithosphere anisotropy ξ (transverse isotropy) variation. ξ is defined as (Vsh/Vsv)2 where Vsh and Vsv are horizontally and vertically polarized shear waves, respectively. . Positive variation (blue) is Vsh>Vsv, while negative variation (red) is Vsv>Vsh. c), anisotropy fast axis symmetry axis direction, plotted against the absolute plate 4. Magmatic time machine in the Yellowstone hotspot layer (see a in Box 1). Ascending diamond-bear- motion (APM) direction such that red means APM parallel and blue APM normal. d), olivine Mg# contours w.r.t. depth inferred from thermal-barometric anal- ing magmas therefore can easily penetrate this ysis plotted at sample sites (shown as CC’ in the map), from Griffin et al. 2004. Contour number 92 is thought to represent the high depleted shallow chemical Yellowstone is a world-class hotspot system, the prototype of continental plume first proposed by layer (Layer 1), while contours 89-91 is for Layer 2. e), same as c), excepted plotted along CC’. Note the correlation of layer 1 & 2 between the geochemistry layer, but will stop at the Layer1/Layer2 bound- and seismic anisotropy. f), Sketch for forming a kimberlite layer between the Layer 1 and 2 boundary. Study regions in Sections 3, 4 and 5 are indicated. Morgan (1971): on the surface age-progressive volcanic (figure a) show the northwest- ary due to barrier imposed by the highly deplet- ward motion of the hotspot system (from 17Ma to the current time), indicating the passage of ed (ergo, viscously very strong) shallow Layer 1. A kimberlite layer is speculated to form at the boundary (f; Sleep G-cubed 2009). Those kimberlites will lithosphere over a underneath plume (hot upwelling rocks), which is confirmed by teleseismic escape to the surface when they find lithospheric scale weak zones (e.g., red dented region in the blue in b). tomography (figure b; Yuan and Dueker 2005).

References cited: Griffin, W.L., Begg, G.C., O'Reilly, S.Y., 2013. Continental-root control on the genesis of magmatic ore deposits. Nature Geosci 6, 905-910. Taking advantage of several earthquake seismology tech- Griffin, W.L., O'Reilly, S.Y., Doyle, B.J., Pearson, N.J., Coopersmith, H., Kivi, K., Malkovets, V., Pokhilenko, N., 2004. Lithosphere mapping beneath the North American plate. Lithos 77, 873-922. eastern Snake River Plain Sleep, N.H., 2009. Stagnant lid convection and carbonate metasomatism of the deep continental lithosphere. Geochem. Geophys. Geosyst. 10. niques good for local scale resolution, several studies address how much the plume has modified the Archean lithosphere: e.g. ambient noise (figure c; Stachnik et al. 2008), body wave 3. Subduction polarity reversal and wedge tectonics tomography (figure b; Yuan and Dueker 2005), receiver func- (a) tion CCP stacking (figured; Yuan et al. 2010). A) Map showing the northeastward apparent motion of hotspot migration along the eastern Snake River Plain and the ages of the various calderas. With teleseismic P- (a) and S-wave (b) body-wave tomography, a regional scale transect from the CDROM deployment in the US Rocky Mountain From http://geology.isu.edu/Digital_Geology_Idaho/Module11/mod11.htm Ambient noise tomography is a new technique that uses the reigon (c) targeting the Archean Wyoming craton and its southern accreted Proterozoic terranes (c and d) yields interesting lithospheric scale velocity earth noise (instead of earthquakes) as its signal to image the Mid-crustal sill (MCS) anomalies along the craton margin (Yuan et al. AGU Monograph 2005): (a) crust and shallow mantle shear wave structure. Its resolution 1) A north-dipping high velocity anomaly, the “Cheyenne slab” (a and b) that rep- (b) reaches sub-tens of km (this section) and even tens of meters 7.x layer 7.x layer Craton Craton Craton P-wave Cheyenne Slab resents a piece of relict slab segments entombed in the Archean lithosphere. B) 3-D rendition of the Yellowstone plume[Yuan and (next section). Unprecedented resolution provided by this Lithosphere Lithosphere Lithosphere Dueker, 2005]. The gray line denotes the Eastern method reveals extremely interesting features in the hotspot 2) A deep seated low velocity structure, the “Aspen anomaly” (a and b) that may Snake River Plain (ESRP). Red contours are the Rhyo- lite calderas (also see a). Note a deep low velocity crust (c and e), a mid-crust high velocity sill (MCS, see profile A Aspen Anormaly fuel the Jemez volcanic field on the surface, and may be responsible for the mineral feature, the Yellowstone plume extends from the in (c)) that deepens from the youngest (0.8 Ma) out- (c) deposits near the surface (d); Yellowstone Park down northwesterly to over 500 km depth. wards to the oldest (17Ma). In the Archean lower crust, a high 7.x layer

(b) ve- locity feature (7.x layer in profiles A-C in (c)) is observed, a feature first seen in the active Craton Craton South-dipping subduction was proposed for the suturing of the southern Proterozoic Lithosphere Lithosphere source studies (proposed as the 7.x-layer because of its 7.x P-wave velocity) and also confirmed by S-wave Cheyenne Slab terranes to Archean Wyoming, due to lack of subduction related rocks on the Archean receiver function CCP stacking (Yuan et al. 2010) as the double Moho-like signal (see figure (d)). side and the dip direction of mylonite zones across the suture. However the north-dip- (C) Ambient noise imaging from Stachnik et al. 2008. Profiles A-C show shear ping Cheyenne slab may call for an episode of subduction reversal after the suture wave velocity variations. Note the feature discussed in the text: MCS in profile Aspen Anormaly zone was formed (see Yuan et al. 2005for details). A; 7-x layer in A-C. Also note the low velocity plume has clearly cut through (e) Taken together, the hotspot the high velocity Archean lithosphere in profiles B-C. CD-ROM body-wave tomography system illustrates a magmatic time (d) (a) The P-wave image. (b) The S-wave image. Blue and red denote high and low velocity varia- 3) The wedge tectonics near the margin of the Archean craton is also inferred (e and machine that recorded the tions. Triangles at zero depth are stations. Note the 3 major anomalies, the Cheyenne slab, the f), where a wedge of Archean lithosphere extrudes into the more juvenile (Proterozoic) Jemez body and the low velocity pipe (more obvious in the S-wave image). Labels see (c). plume/lithosphere interaction block. Top of the 7.x layer through time (figure e): (c) (d) Moho Wedge Tectonics • Impinging plume erodes the (e) Cartoon of lithospheric-scale (e) (f) base of Archean lithosphere, and deformation during convergence Laramie Array (e) tectonics that may result in prom- i- emplaces a sheared plume head nent seismic reflections. The underneath (Plume conduit and upper crustal part of the section E) Magmatic time machine. See text for interpretations. plume layer); Top of the 7.x layer illustrates ‘flake tectonics’, ‘Croco- Archean Moho diles’ and doubly vergent orogens • Decompression melting creates hot basalt magma, which rises through dyking into the deplet- Moho reflecting deformation in the entire crust. Deformation within Proterozoic Moho ed lithosphere and crust due to their lighter density, and eventually stagnates in the middle of the the uppermost mantle illustrates crust when reaching neutral buoyancy; wedge tectonics as based on deep (D) Receiver function CCP stack images form • Differentiation of stagnated melts produce denser Fe-rich residual magma in the mid-crust, Yuan et al. 2010. Profiles A, B location shown Jemez Volcani field seismic reflection sections in Canada nd Scandinavia (Snyder et al., 2002). which eventually cools and forms the large mid crustal sill complex (MCS). in the map. Note above the brightest signal (f) Imbricated Moho across the Archean Wyoming/Proterozoic boundary, the Cheyenne belt (CB) in the receiver function Com- Moho the presence of another positive veloc- Yellowstone mon-conversion-point stacking image from the Laramie array (see c for location). • The MCS becomes a positive load on the lower crust which could force the low velocity lower ity gradient signal, interpreted as the top of caldera Crustal Provinces of the southern Rockies and mineral deposits crust (LVLC) to flow outwards on millions year time scale. the 7.x layer. (c) Crustal provinces of the southern Rockies and CD-ROM transects (black triangles). Geographic labels are: GF, Geochron Front [Chamberlain, 1998]; CB, Cheyenne belt; FM-LM, Farewell Mountain-Lester Mountain shear zone that separates the Green mountain and Rawah arcs; SL, San Luis basin; JT, the Jemez volcanic trend/suture; GP, the Great Plains; CM, Cedar Mountains; LH, Leucite Hills; NVF, Navajo Volcanic Field; and MP, Middle Park. Tomographic • Differentiation also creates a small amount of lower density rhyolitic magma that ascends to the upper crust to form silicic chambers, with further images are presented along A-A’. (d) Vp tomography (90 km depth; note the color reversal: red is high velocity and blue is slow velocity; Griffin et al. 2013) of western USA, showing main Precambrian to Tertiary (magmatic-)hydrother- fractionation the silicic magmas eventually reach the surface to form age-progressive calderas. mal ore deposits by size (supergiant, giant and major) and dominant metals (yellow, Au; green, Cu-Au-Mo; pink, Cu-Mo-Ag-Au; orange, Mo; light blue, REE; light grey, W(-Sn); dark grey, Fe). Main lithospheric blocks, defined at sub-crustal depths from multidisciplinary data, are outlined. Deposits concentrate along prominent lithospheric structures, particularly in lower-velocity regions (blue) or on the flanks of highs, where lower velocities reflect refertil- ization of the sub-continental lithosphere mantle and/or higher temperature. Figure (d) and captions taken from Griffin et al. 2013. References cited: Morgan, J.W., 1971. Convection plumes in the lower mantle. Nature 230, 42-43. Stachnik, J.C., Dueker, K., Schutt, D.L., Yuan, H., 2008. Imaging Yellowstone plume-lithosphere interactions from inversion of ballistic and diffusive Rayleigh wave dispersion and crustal thickness data. Geochem. Geophys. Geo- References cited: Griffin, W.L., Begg, G.C., O'Reilly, S.Y., 2013. Continental-root control on the genesis of magmatic ore deposits. Nature Geosci 6, 905-910. syst. 9, Q06004. Snyder, D.B., 2002. Lithospheric growth at margins of cratons. Tectonophysics 355, 7-22. Yuan, H., Dueker, K., 2005. Teleseismic P-wave tomogram of the Yellowstone plume. Geophys. Res. Lett. 32, L07304. Yuan, H., Dueker, K., 2005. tomographic Vp and Vs images of the in Wyoming, Colorado and New Mexico: Evidence for thick, laterally heterogeneous lithosphere, in: Randy, G., Karlstrom, K.E. (Eds.), Yuan, H., Dueker, K., Stachnik, J., 2010. Crustal structure and thickness along the Yellowstone hot spot track: Evidence for lower crustal outflow from beneath the eastern Snake River Plain. Geochem. Geophys. Geosyst. 11, The Rocky Mountain region--an evolving lithosphere: tectonics, geochemistry, and geophysics. American Geophysical Union, Washington, DC, pp. 329-345 Q03009.