The Pennsylvania State University
The Graduate School
College of Earth and Mineral Sciences
SEISMIC ANISOTROPY ACROSS THE APPALACHIAN
MOUNTAINS AND PLATEAU
A Thesis in
Geosciences
by
Austin White-Gaynor
c 2015 Austin White-Gaynor
Submitted in Partial Fulfillment of the Requirements for the Degree of
Master of Science
August 2015 ii" " " " " " " The thesis of Austin White-Gaynor was reviewed and approved* by the following:
Andrew Nyblade Professor in Geosciences Thesis Adviser
Charles J. Ammon Professor in Geosciences
Sridhar Anandakrishnan Professor in Geosciences
Demian Saffer Professor in Geosciences Interim Associate Head of Graduate Programs and Research
*Signatures on file in the Graduate School iii Abstract
Few modern mountain ranges have as extensive of a deformational history as the
Appalachian Mountains. Running nearly the entire length of the eastern coast of North
America, the Appalachian Mountains are the result of a roughly 500 Ma Wilson Cycle.
Geochemical and geophysical evidence points toward a diachronous orogenic evolution of northern and southern segments of the mountain belt. I characterize seismic anisotropy in the mantle throughout the central Appalachian region, the boundary between the northern and southern segments, in order to identify any first-order lithospheric changes between the two segments. Shear-wave splitting measurements are made throughout the region, and while delay times are similar everywhere (average = 0.87 seconds), fast azimuth directions ( ) change from roughly E-W in the eastern part of the study area rotating smoothly to NE-SW in the western part of the study region, across the Allegheny
Front. Extended analysis was performed at three permanent stations located in the eastern, western and central part of the study area. These results show complex, or multilayered, anisotropy in the eastern portion of the study area and less complex, likely single layered anisotropy in the western portion of the study area. The top layer of the multilayered case has a of 60 , roughly normal to the continental margin, which can be associated with frozen-in anisotropy from Triassic rifting. This top layer does not extend into the western most part of the study area, but instead pinches-out or diminishes in magnitude to zero. The bottom layer, present at all three permanent stations, is modeled with = -60 , which is sub parallel to apparent plate motion and likely produced by iv
finite strain in the asthenosphere. Fast azimuth directions in the northern portion of the study show less change along longitude and remain mostly E-W oriented. These results are attributed to the large E-W shear-zone in southern New York that formed during the Alleghenian orogeny and could represent a dividing boundary between northern and southern segments of the Appalachian Mountains. v Table of Contents
List of Tables ...... vivii
List of Figures ...... viiviii
Acknowledgments ...... ixx
Chapter 1...... 1
1.1 Introduction...... 1
1.2 Tectonic Background ...... 4
1.2.1 TheGrenvilleProvince...... 4
1.2.2 The Taconic Orogeny ...... 5
1.2.3 The Acadian Orogeny ...... 7
1.2.4 The Alleghanian Orogeny ...... 9
1.2.5 Continental Break Up ...... 11
1.3 Anisotropy ...... 12
1.3.1 Seismic Anisotropy ...... 12
1.3.2 Mantle Anisotropy ...... 14
1.3.3 Shear-wave Splitting ...... 15
1.3.4 Global and Regional Studies ...... 17
Chapter 2...... 20
2.1 Data and Methods ...... 20 vi
2.1.1 Data ...... 20
2.1.2 Transverse Component Energy Minimization Method . . . . . 22
2.1.3 Multiple Layer Modeling ...... 28
Chapter 3...... 29
3.1 Results...... 29
3.2 Discussion...... 34
3.3 Summary and Conclusions ...... 48
Bibliography ...... 49
Appendices ...... 59
Appendix A List of individual events for each station ...... 59
Appendix B Summed Transverse Component Energy Grid for each station . . . 76
Appendix C Individual results for each station ...... 149
Appendix D Plots of individual results for stations MVL, SSPA, and ALLY . . . 166 viivi List of Tables
1:#List#of#stacked#results#for#each#station#.#.#.#.#.#.#.#.#.#.#.#.#.#.#.#.#.#.#.#.#.#.#.#.#.#.#.#.#.#.#.#.#.#.#.#.##32 viii List of Figures
1.1 Location of study area...... 3
1.2 Appalachian promontories and embayments...... 4
1.3 MapshowingextentofGrenvilleProvince...... 6
1.4 Cross-sectional diagram of the Taconic Orogeny...... 7
1.5 Map showing the major elements of the Taconic Orogeny...... 8
1.6 The Appalachian peri-Gondwanan realm and its major elements. . . . . 8
1.7 Schematic cross-section of the Acadian Orogeny...... 9
1.8 Schematic cross-section of the Alleghanian Orogeny...... 10
1.9 Map showing the major elements of the Alleghanian Orogeny...... 11
1.10 Map showing the extent and location of Triassic rift basins in eastern
North America...... 13
1.11 P-wave and S-wave velocities (in km/s) in a single olivine crystal. . . . . 14
1.12 Schematic diagram of an incident S-wave traveling through an anisotropic
medium(gold)...... 16
1.13 SKS raypaths at epicentral distances of 95 - 140 degrees...... 17
1.14 Global shear-wave splitting results...... 18
1.15 Shear-wave splitting results for central and southeastern North America. 19
2.1 Location of stations used in this investigation...... 21
2.2 Event (Mw 5.5) distribution for station MVL...... 23 ix
2.3 Example waveform and particle motion before and after splitting correc-
tions...... 25
2.4 Final stack for station L60A...... 26
2.5 Example waveform and particle motion for a null result...... 27
3.1 Map showing splitting results for individual stations...... 30
3.2 Consistency check for stations WRPS and SSPA with previously pub-
lishedresults...... 31
3.3 Fast Azimuth Direction ( ) vs longitude (a) and latitude (b) for all high-
qualityresults...... 35
3.4 Map of all shear-wave splitting results in eastern North America. . . . . 36
3.5 Fast Azimuth Direction ( ) vs. Distance from the Allegheny Front. . . . 38
3.6 Inversion results from Levin et al. [1999]...... 40
3.7 Observed vs Synthetic results at station MVL...... 41
3.8 Observed vs Synthetic results at station ALLY...... 42
3.9 Observed vs Synthetic results at station SSPA...... 43
3.10 Developed cross-section from modeling results...... 44
3.11 Previous model, new model, and observed results at station MVL. . . . 45
3.12 Previous model, new model, and observed results at station SSPA. . . . 45
3.13 Previous model, new model, and observed results at station ALLY. . . . 46
3.14 Moment tensor solutions for study area taken from Herrmann et al. [2011]. 47 x Acknowledgments
For this work I received much support from my faculty mentors, including Andy
Nyblade, family members and friends to whom I owe a big Thank You. xi
The mountains are the soul of the region.
To understand the mountains is to know ourselves.
– Sandra H.B. Clark 1
Chapter 1
1.1 Introduction
The Appalachian Mountains exhibit remnant features from a deep tectonic his- tory. Mountain building began in the Mesoproterozoic with the formation of the Grenville
Mountains. After subsequent delamination, these mountains collapsed and formed the underlying basement rock for eastern North America [Rivers, 1997]. Atop the east- ward dipping metamorphic basement now lie the Appalachian Mountains, the result of a roughly 500 Ma Wilson Cycle comprised of three distinct orogenic phases, the Taconic, the Acadian and the Alleghenian [Stanley and Ratcli↵e, 1985]. Figure 1.1 shows the study area for this investigation, including the Allegheny Front, an escarpment dividing the Valley and Ridge Province to the east from the undeformed Allegheny Plateau to the west. The Grenville Front is a demarcation of the western extent of the Grenville
Province. Substantial evidence points toward a diachronous evolution of northern and southern segments of the Appalachian Mountains which contain both throughgoing and nonthoughgoing elements (Figure 1.2). While the Taconic Suture extends the entire length of the orogen, geochronology shows a major distinction between the two segments with the northern segments incurring the oldest deformation in the west and youngest in the east [Hatcher and Odom, 1980]. The southern and central segment of the Appalachi- ans includes youngest deformation on the eastern and western flanks of the mountain 2 belt and oldest deformation in the center [Hatcher and Odom, 1980]. The boundary between the Northern and Southern Appalachians lies directly west of New York City and is also marked by the thinnest area of the orogen widening both to the north and to the south (Figure 1.2).
The aim of this study is to investigate the nature of seismic anisotropy in the upper mantle beneath the central Appalachian Mountains in order to identify any first-order boundary between the northern and southern segments. Any remnant anisotropy in the lithosphere will also provide information as to how anisotropy is created and retained through such a complex deformational system. Using shear-wave splitting analyses, I characterize anisotropic fabric in the upper mantle beneath 73 broadband seismic sta- tions spanning 10 longitude and 5 latitude encompassing Pennsylvania and portions of the surrounding states. The results of this analysis are used to further constrain deformational history of the Appalachian Mountain range.
This thesis is composed of three main chapters. The remainder of this chapter includes background information on the geology and tectonics of the Appalachian Moun- tain orogeny, seismic anisotropy, and how seismic anisotropy can be used to constrain regional tectonics and deformational history. Chapter 2 provides detailed information of the methodologies and data used in this investigation. Chapter 3 includes the results of the shear-wave splitting analyses, as well as a discussion about the splitting results and their implications for understanding seismic anisotropy and lithospheric structure beneath the orogen. 3
t n t o n r o F r F y n
Grenville lleghe A
Fig. 1.1. Map showing the study area outlined in the red box. Physiographic provinces are color-coded. The Grenville and Allegheny Fronts are drawn with dashed lines. Figure courtesy of Dr. Bruce Railsback. 4
P ennsyl
v ania
S alie n t
Fig. 1.2. Appalachian promontories and embayments. Study area for this investigation is shown in red box. Extent of the Northern and Southern Appalachians is noted. Figure taken from [Hibbard et al., 2010]
1.2 Tectonic Background
1.2.1 The Grenville Province
Proterozoic Laurentia formed through the amalgamation of orogenic belts sur- rounding the Superior, Rae, and Nain cratons (Figure 1.3. The Grenville Province, a
Mesoproterozoic orogenic belt, formed along the southern extent of Laurentia (Figure
1.3) [Rivers, 1997]. Pre-Grenvillian basement includes paleoproterozoic calc-alkaline arc magmatism and magmatic complexes that formed during the Labradorian ( 1680 Ma), ⇠ the Pinwarian ( 1500 - 1450 Ma), and the Elzevirian ( 1250 - 1190 Ma) orogenic events. ⇠ ⇠ These polycyclic basement components were later metamorphosed to become part of the current Grenvillian basement autochthon of eastern North America. The Grenvil- lian orogeny involved continent-continent collision between 1.19 - 0.98 Ga during the ⇠ formation of the supercontinent Rodinia. Orogenisis occurred in three distinct phases 5 separated by periods of extension and coeval emplacements of mafic magmatism and anorthosite complexes, suggesting possible delamination and orogenic collapse [Rivers,
1997].
Important Grenvillian lithotectonic elements within the study area include medium- grade metamorphic basement in western Pennsylvania and New York and in eastern
Ohio and West Virginia. After failed rifting 750 Ma, the supercontinent Rodinia ⇠ subsequently broke up 565 Ma during a 200 Ma rifting event forming the Iapetus ⇠ ⇠ Ocean between Laurentia and Gondwana [Aleiniko↵ et al., 1995], thus beginning the
Appalachian Wilson Cycle. Irregular rifting of Rodinia formed the series of promonto- ries and embayments defining the modern Atlantic margin (Figure 1.2).
1.2.2 The Taconic Orogeny
The formation of the Appalachian Mountains was a result of three major orogenic events that began in the Middle to Late Ordovician ( 465 Ma) with the Taconic Orogeny. ⇠ This mountain building event involved collision of volcanic island arcs and microterrains with the eastern margin of rifted Laurentia (Figure 1.4).
Taconian deformation was intense and polyphase in the interior of the Appalachian orogen in New England [Stanley and Ratcli↵e, 1985], however the ”Taconian Suture” con- sists of a major fault and terrain boundary that is traceable along the entire extent of the Appalachian orogeny with arc accretion throughout [Hatcher, 2010]. Many granitoid plutons appear in the Piedmont and Blue Ridge in the Southern Appalachians (Figure
1.5), as well as in western New England [Wilson, 2001; Stanley and Ratcli↵e, 1985]. Early
Silurian felsic volcanic rocks were extruded in central western Newfoundland, suggesting 6
Fig. 1.3. Map showing the extent of the Grenville Province. Study area is shown in red box. Figure taken from Rivers [1997]. 7
Fig. 1.4. Schematic cross-section of the Taconic Orogeny. Figure adapted from Hatcher Jr [1987]. that some rifting followed the orogeny [Van Staal et al., 1998]. Late Ordovician to Early
Silurian molasse formed outboard of the eroding Taconic Mountains across most of the central Appalachians [Rodgers, 1971].
1.2.3 The Acadian Orogeny
The peri-Gondwanan superterranes Avalon, Carolina and Gander (Figure 1.6) collided with Laurentia and the previously accreted early Paleozoic terranes during the zippered north-to-south closing of the Rheic ocean, thus resulting in the Acadian and
Neoacadian orogenies [Hatcher, 2010]. More specifically, the Acadian orogeny was a re- sult of docking of the Carolina superterrane in the south (Figure 1.7) and Avalon/Gander superterranes in the north beginning in the Middle Devonian.
The collision of the Carolina and Avalon superterranes produced plutons in both the allochthon and autochthon, however these plutons are much more abundant in the 8
Fig. 1.5. Map showing the major elements of the Taconic Orogeny. Figures taken from Hibbard et al. [2010].
Fig. 1.6. The Appalachian peri-Gondwanan realm and its major elements. Figure taken from Hibbard et al. [2007]. 9
Fig. 1.7. Schematic cross-section of the Acadian Orogeny. Figure adapted from Hatcher Jr [1987]. northern Appalachians close to the location of the more westward subduction [Butler and Fullagar, 1978; Phinney, 1986]. Also due to the transpressional nature of the colli- sion, deformation and high-grade metamorphism associated with this event raised large northern Appalachian Mountains and was the most significant mountain building event in the Appalachian Mountains of New England.
1.2.4 The Alleghanian Orogeny
The final stage of the Appalachian Wilson Cycle, the closing of the Theic Ocean and formation of Pangea, was the continent-continent collision of Laurentia with Gond- wana resulting in the Alleghanian orogeny c. 325 Ma (Figure 1.8). Similar to previous orogenic events, the Alleghanian orogeny involved north-to-south transpressional closing of the Theic ocean, with sparse deformation mainly near the coastal edge of the Northern 10
Appalachians [Rast, 1984]. Rotation of Gondwana would later cause head on collision in the southern Appalachians, resulting in the formation of the Blue Ridge-Piedmont megathrust sheet, and later, the Early Permian Alleghanian fold-thrust belt and the approximately 160 km of shortening in the region [Hatcher, 2010; Arthaud and Matte,
1977]. Significant dextral faulting occurred from New England to Nova Scotia [Hatcher,
2010]. It has also been suggested that the narrow, strongly curved segment of the Penn- sylvania Salient (Figure 1.2) is a result of the Reguibat Promonotory in West Africa colliding with the eastern margin of Laurentia prior to the main African collision [Lefort and Van der Voo, 1981].
Fig. 1.8. Schematic cross-section of the Alleghanian Orogeny. Image adapted from Hatcher Jr [1987].
Pull-apart basins formed in the Northern Appalachians were later destroyed by the Alleghenian orogeny which resulted in a 1000 km wide dextral shear zone with the ⇠ southern extent at the Central-Northern Appalachian Boundary [Arthaud and Matte, 11
1977]. Right-lateral wrench-faulting dominates the Northern Appalachians, the southern extent of which marks the southern boundary of the New York Promontory and runs along the latitude of Long Island [Arthaud and Matte, 1977]. This large right-lateral transcurrent fault is known as the Kelvin, or Cornwell-Kelvin, Fault [Arthaud and Matte,
1977]. Similar shearing occurred south of the Southern Appalachians [Arthaud and
Matte, 1977].
Fig. 1.9. Map showing the major elements of the Alleghanian Orogeny. Figure taken from Hibbard et al. [2010].
1.2.5 Continental Break Up
Beginning in the Late Triassic ( 200 Ma), the eastern Gondwanan continent ⇠ began to pull apart from western Laurasia and successful rifting occurred along the
Atlantic margin. Thus, following the 490 Ma Appalachian orogeny, the supercontinent ⇠ Pangea proceeded to rift and drift, completing the Appalachian Wilson Cycle. Former thrust faults shifted to low-angle normal faults creating the roughly 40 northeast-trending 12 elongate rift basins (Figure 1.10) [Manspeizer, 2013]. Prior to drifting, this margin underwent tectonic uplift and crustal thinning. Cooling and subsidence occurred in the o↵shore basins of the Atlantic margin starting in the Middle Jurassic, thus marking the drift period and formation of the Atlantic Ocean [Manspeizer, 2013].
1.3 Anisotropy
1.3.1 Seismic Anisotropy
Anisotropy can be defined as the property of being directionally dependent. In the case of an elastic medium, this applies to physical properties such as seismic wave speed.
Seismic anisotropy is a direct result of the intrinsic mineralogical structure, or mineral lattice, which can be and is altered during mineral deformation [Stein and Wysession,
2009]. While some lattice systems are inherently isotropic, it is quite common for a crystal to have an anisotropic structure with orthorhombic or hexagonal lattice systems.
Anisotropic lattice systems are the natural order for many minerals, such as olivine, however in lattice preferred orientation, LPO, the orienting of anisotropic min- erals is produced through high strain processes such as dislocation creep, or plastic flow in the mantle [Zhang and Karato, 1995; Christensen, 1984]. This should not be confused with di↵usion creep, which occurs at low stress and/or small grain size but does not result in seismic anisotropy [Savage, 1999]. In general, the seismically fast axes are in the plane of the flow with the a-axis (Figure 1.11), the fastest direction, pointing in the direction of flow and the b-axis, the slow axis, normal to the direction of flow [Dziewonski and Anderson, 1981]. While measurements of seismic anisotropy of individual minerals 13
Fig. 1.10. Map showing the extent and location of Triassic rift basins in eastern North America. Figure created from the USGS database by the National Park Service. 14
Fig. 1.11. P-wave and S-wave velocities (in km/s) in a single olivine crystal. The a, b, and c axes are label. Figure taken from Kumazawa and Anderson [1969]. can be conducted in a laboratory environment, a large assemblage of statistically aligned
LPO is necessary for the anisotropy to be detectable on the scale of propagating seismic waves; it has been shown that olivine is easily oriented by the ambient stress or flow field over su ciently large areas [Anderson and Dziewonski, 1982; Silver, 1996].
1.3.2 Mantle Anisotropy
Anisotropy is a fundamental characteristic of Earth materials ranging radially in space from the core to the crust [Shearer and Toy, 1991; Silver and Chan, 1991; Vinnik et al., 1989; Crampin and Booth, 1985]. Fluid filled cracks are likely the dominant source of anisotropy in the continental crust [Crampin and Booth, 1985; McNamara and Owens,
1993], however the major source of anisotropy, up to five times the magnitude [Silver,
1996], recorded teleseismically, derives from LPO of olivine. Peridotite, the primary 15 rock type found in the mantle, is composed of 60% olivine [Christensen, 1984; Nicolas and Christensen, 1987]. Orthopyroxene, which is also susceptible to LPO, makes up another 20% [Nicolas and Christensen, 1987; Babuska and Cara, 1991; Christensen and
Lundquist, 1982]. As mentioned before, olivine-rich aggregates are susceptible to finite strain on kilometer scales and deformation causes the a-axes of the LPO to concentrate in the direction of flow [Ribe, 1989; Silver, 1996; Mainprice and Silver, 1993]. Upper mantle anisotropy has been measured by numerous long-period surface wave investiga- tions [Tanimoto and Anderson, 1985; Montagner and Tanimoto, 1991; Anderson, 1965;
Forsyth, 1975], many studies of the Pn body phase [Hess, 1964; Shearer and Orcutt, 1986;
Raitt et al., 1969; Bamford, 1977], and also through free oscillation inversions [Anderson and Dziewonski, 1982], with an anisotropic layer necessary to a depth of at least 200km.
Because intrinsic anisotropy requires both anisotropic crystals and preferred orientation, the anisotropy of the mantle contains information about both composition as well as long-term strain deformation [Anderson and Dziewonski, 1982].
1.3.3 Shear-wave Splitting
One important and useful manifestation of anisotropy is shear-wave splitting, a phenomenon analogous to birefringence observed in an optically anisotropic medium, such as a calcite crystal [Silver, 1996]. When a single shear-wave encounters an anisotropic fabric, it will split into two orthogonal shear waves that propagate through the anisotropic medium at di↵erent velocities. One S-wave will travel polarized along the a-, or fast, axis while another will travel with displacement along the b-, or slow, axis (Figure 1.12) 16
[Babuska and Cara, 1991]. Shear-wave splitting expresses itself over a range of frequen- cies and exhibits a range of sensitivities to structures on di↵erent length scales in di↵erent parts of the Earth [Long and Silver, 2009].
Fig. 1.12. Schematic diagram of an incident S-wave traveling through an anisotropic medium (gold). Figure taken from W¨ustefeld et al. [2008].
The polarization of the fast component, , and the time delay, t,betweenthe components provide simple measurements to characterize anisotropy [Savage, 1999]. In order to simplify the calculations, core phases, such as SKS and SKKS (Figure 1.13), are utilized due to their known initial radial polarization, equal to back-azimuth, for the P-to-S converted wave exiting the core; an unaltered shear-wave radial polarization depends on the focal mechanism which complicates computation. 17
Fig. 1.13. SKS raypaths at epicentral distances of 95 - 140 degrees. Figure taken from Dziewonski and Anderson [1981].
1.3.4 Global and Regional Studies
Early work in the study of seismic anisotropy was done by Christo↵el (1877) and
Lord Kelvin (1905), however it was not until the last 15 years that shear-wave splitting studies were commonly made (Figure 1.14) [Savage, 1999]. It has been shown that the orientation and magnitude of the anisotropy varies widely between geologic settings and that there is no simple way to interpret shear-wave splitting results [Wenk et al., 1991;
Zhang and Karato, 1995; Ribe and Yu, 1991].
Nonetheless, observations of shear-wave splitting have led to a better understand- ing of deep structural geology in cratonic, oceanic rift, and subduction zones settings, among others. In orogenic environments, when uniaxial compression is the driving force,