Theory of the Earth Don L. Anderson BLACKWELL SCIENTIFIC PUBLICATIONSTheory of the Earth Don L. Anderson Seismological Laboratory California Institue of Technology Pasadena, California BLACKWELL SCIENTIFIC PUBLICATIONS Boston Oxford London Edinburgh MelbourneEatorial Offices ‘Osney Mead, Oxford, OX2, OEL, UK 8 John Street, London WCIN 28S, UK 23 Ainslie Place, Edinburgh, EH3 6AJ, UK 3 Cambridge Center, Suite 208, Cambridge, MA 02142 107 Barry Stet, Carton, Victoria 3053, Australia Sponsoring Editor: Yohn Staples Manuscript Etor: Andrew Alden, Production Coordinator: Partners in Publishing/Pat Waldo and Susan Swenson [Interior and Cover Design: Gary Head Design Compositor: G&S Typeseters, In. Printer: Arata Graphics/Kingsport Distributors USA and Canada Blackwell Scientific Publications loRBS, PO. Box 447 Brookline Village, MA 02147 (617) 524.7678 Australia Blackwell Scientific Publications (Australia) Pty Ltd 107 Barry Stret, Carton Victoria 3053 UK Blackwell Scientific Publications Osney Mead Oxford OX2 OBL, 11 44 865-240201 (© 1989 Blackwell Scientific Publications All sights reserved. No par of ths publication may be epro- ‘duced, stored in a retrieval system, or transmitted, in any form or by any means, electronic, mechanical, photocopying, recording ‘ot otheewise without the prior permission ofthe copyright owner. 129876543 Library of Congress Cataloging-in-Publication Data ‘Anderson, Don L. ‘Theory ofthe earth Includes bibliographies and index. 1. Eanh 2. Geophysics 1. Title QC806.A515 1989 551 88,7908 ISBN 0-86542-335-0 (hardcover) 0-86542-123-4 (paperback)1 The Terrestrial Planets 1 Theories of Planetary Formation 1 Meteorites 4 Planetary Atmospheres 8 Composition of the Terrestrial Planets Comparative Planetology 24 References 24 2 Earth and Moon 27 Bulk Composition of the Earth 27 Evolution of the Earth’s Interior 28 Origin of the Crust 31 Origin of the Mantle and Core 31 Mineralogy of the Mantle 34 Moon 36 References 43 3 The Crust and Upper Mantle The Crust 45 ‘The Seismic Lithosphere or LID 49 u 45 Contents ‘The Low-Velocity Zone or LVZ 51 Mineralogical Models of 50-400 km Depth 56 ‘The Transition Region 57 References 61 4 The Lower Mantle and Core 63 ‘Composition of the Lower Mantle 63 Region D" 66 The Core 67 Mantle-Core Equilibration 73 ‘The Magnetic Field 74 References 76 5 ‘Thermodynamics and Equations of State 79 ‘Thermodynamics 79 Theoretical Equations of State 82 ‘The Griineisen Relations 83 Effect of Temperature on Bulk Moduli 85 ‘Thermal Expansion and Anharmonicity 86 Isothermal-Adiabatic Transformations 89vi contents Calculation of Density in the Earth 90 Finite-Strain Equation of State 91 Zero-pressure Values of Lower-Mantle Seismic Properties 92 ‘The Equation of State 92 The Seismic Parameter D 95 Effect of Composition and Phase 95 ‘The Repulsive Potential 97 Shock Waves 99 References 102 6 Elasticity and Solid-State Geophysics 103 Elastic Constants of Isotropic Solids 103 ‘Temperature and Pressure Derivatives of Elastic Moduli 106 Intrinsic and Extrinsic Temperature Effects 107 Seismic Constraints on Thermodynamics of the Lower Mantle 110 Correcting Elastic Properties for Temperature 120 Elastic Properties of Composite Materials 121 Liquids 122 References 126 7 Nonelastic and Transport Properties 129 ‘Thermal Conductivity 129 Diffusion and Viscosity 132 Homologous Temperature 135 Dislocations 135, Dislocation Creep 136 ‘The Lithosphere 138 Effective Elastic Thickness of the Lithosphere 140 Melting and Origin of Magmas 140 onic Radi 143, References 144 8 Chemical Composition of the Mantle 147 Methods of Estimating Mantle Chemistry 147 The Upper Mantle 153 ‘The Lower Mantle 155 Magmas 156 Midocean-Ridge Basalts (MORB) 157 Hotspots 158 Ocean-Island Basalt 161 Kimberlte 162 ‘The Kimberite-MORB Connection 164 ‘The Kimberite-KREEP Relation 165 Alkali Basalt 166 Continental Flood Basalts 168 Andesite 169 Marginal Basins 172 Komatiites 175 Summary 175 References 176 9 The Source Region 179 Background 179 Ultramafic Rocks 181 Pyrolite 184 Eclogites 186 Piclogite 188 Depths of Mantle Reservoirs 189 Mantle Metasomatism and the Enriched Reservoir 192 Rise of Deep Diapirs 192 Trace-Element Modeling 194 References 19610 Isotopes 197 Lead Isotopes 198 Strontium and Neodymium Isotopes 201 ‘The Lead Paradox 202 Isotopic Evolution of the Mantle 203 Oxygen Isotopes 205 Rare Gases 206 Isotopic Constraints on Magma Genesis 210 A Model for Magma Genesis and Evolution inthe Mantle 213 References 215 11 Evolution of the Mantle 217 ‘Models of the Mantle Petrogenesis 218 Petrological Evolution of the Mantle 221 ‘Chemical Stratification of the Mantle 225 Nature of the Enrichment/Depletion Process 228 ‘The Role of Magma Mixing 231 Melts from the Enriched Re Magma Ocean 235 Summary of Constraints on Mantle Evolution 237 References 238 voirs 233 12 The Shape of the Earth, Heat Flow and Convection 239 Topography 239 ‘The Geoid 241 Heat Flow 247 Rotation of the Earth and Polar Wander 249 Convection 253 Cooling of the Earth 257 References 257 contents Vii 13 Heterogeneity of the Mantle 259 Upper-Mantle Heterogeneity from Surface-Wave Velocities 259 Global Surface-wave Tomography 260 Regionalized Inversion Results 263 Spherical Harmonic Inversion Results 264 Correlation with Heat Flow and Geoid 267 Azimuthal Anisotropy 269 Lateral Heterogeneity from Body Waves 270 Body-wave Tomography of the Lower Mantle 272 References 277 14 Anelasticity 279 Seismic Wave Attenuation 279 Activated Processes 284 Dislocation Damping 287 Partial Melting 292 Bulk Attenuation 292 Relaxation in Systems Undergoing Phase ‘Transformations 294 Physics of Attenuation in Fluids 295 Absorption-Band Q Model for the Earth 297 References 301 15 Anisotropy 303 Introduction 303 Origin of Mantle Anisotropy 305 Anisotropy of Crystals 307 Theory of Anisotropy 309 ‘Transverse Isotropy of the Upper Mantle 315 ‘Transverse Isotropy of Layered Media 317 ‘The Effect of Oriented Cracks on. Seismic Velocities 319 Inversion Results for the Upper Mantle 320vili_ contents Global Maps of Transverse Isotropy as a Function of Depth 321 ‘Age-Dependent Transverse Isotropy 324 Azimuthal Anisotropy 326 Shear-wave Splitting and Slab Anisotropy 330 Partial Derivatives 331 References 334 16 Phase Changes and Mantle Mineralogy 337 Spherical Ions and Crystal Structure 337 Minerals and Phases of the Mantle 342. Phase Equilibria in Mantle Systems 343 Calculation of Phase Relations 347 Isobaric Phase Changes and Lateral Variations of Physical Properties 348 Slabs 349 References 353 Appendix 357 Earth Model PREM 357 Conversion Factors 361 Physical Constants 361 Earth Parameters 361 Determinants 361It was a long time before man came to understand that any true theory of the earth must rest upon evidence furnished by the globe itself, and that no such theory could properly be framed until a large body of evidence had been gathered together. Sik ARCHIBALD GEIKIE, 1905he maturing of the Earth sciences has led to @ frag- ‘mentation into subdisciplines which speak imperfectly to one another. Some of these subdisciplines are field ge- ology, petrology, mineralogy, geochemistry, geodesy and seismology, and these in turn are spit into even finer units. ‘The science has also expanded to include the planets and even the cosmos. The practitioners in each of these fields tend to view the Earth in a completely different way. Di ccoveries in one field diffuse only slowly into the conscious ress of a specialist in another. In spite of the fact that there is only one Earth, there are probably more Theories of the Earth than there are of astronomy, particle physics or cell biology where there are uncountable samples of each ob- ject. Even where there is cross-talk among disciplines, itis usually as noisy as static. Too often, one discipline’s un- proven assumptions or dogmas are treated as firm boundary conditions for a theoretician in a slightly overlapping area. ‘The data of each subdiseipline are usually consistent with a range of hypotheses. The possibilities can be narrowed con- siderably as more and more diverse data are brought to bear (on a particular problem. The questions of origin, composi- tion and evolution of the Earth require input from astron- ‘omy, cosmochemistry, meteoritics, planetology, geology, petrology, mineralogy, crystallography, materials science and seismology, at a minimum. To a student of the Earth, these are artificial divisions, however necessary they are to make progress on a given front. ‘Examples abound. A seismologist struggling with the meaning of velocity anomalies beneath various tectonic provinces, or in the vicinity of a deeply subducting slab. is, apt to interpret seismic results in terms of temperature var tions in a homogeneous, isotropic half-space or series of layers. However, the petrological aspects—variations in Preface ‘mineralogy, crystal orientation or partial melt content—are ‘much more important. These, in turn, require knowledge of phase equilibria and material properties. ‘An isotope geochemist, upon finding evidence for sev- eral ancient isolated reservoirs in the rocks and being gen- erally aware of the geophysical evidence for a crust and a 650-km discontinuity, will tend to interpret the chemical data in terms of crustal contamination or recycling, a “normal” mantle source and a lower mantle source. This “standard” petrological model is « homogeneous peridotite ‘mantle which contains about 20 percent basalt, available as needed, to fuel the midocean ridges with uniform magmas. Exotic basalts are assumed to be from the core-mantle ‘boundary or to have interacted withthe crust. The crust and shallow mantle may be inhomogeneous, but the rest of the ‘mantle is viewed as well homogenized by convection. ‘The convection theoretician treats the mantle as a ho- mogencous fluid or as @ two-layered system, with con- stant physical properties, driven by temperature-induced bouyancy, ignoring melting and phase changes and even pressure. In Theory of the Earth | attempt to assemble the ‘and pieces from a variety of disciplines which are relevant (o an understanding of the Earth. Rocks and magmas are ‘our most direct source of information about the interior, but they are biased toward the properties ofthe crust and shal- ow mantle. Seismology is our best source of information about the deep interior; however, the interpretation of seis- mic data for purposes other than purely structural requires input from solid-state physies and experimental petrology. Although this is not a book about seismology, it uses seis- ‘mology in a variety of ways ‘The “Theory of the Earth” developed here differs inxii PREFACE ‘many respects from conventional views. Petrologist’s mod- els forthe Earth’s interior usually focus on the composition ‘of mantle samples contained in basalts and kimberltes. The “simplest” hypothesis based on these samples is that the ‘observed basalts and peridotites bear a complementary re lation to one another, that peridotites are the source of ba- salts or the residue after their removal, and that the whole ‘mantle is identical in composition to the inferred chemistry of the upper mantle and the basalt source region. The man- tle is therefore homogeneous in composition, and thus alll parts of the mantle eventually rise to the surface to provide basalts. Subdueted slabs experience no barrier in falling through the mantle to the core-mantle boundary. ‘Geochemists recognize a varity of distinct reservoirs, or source regions, usually taken as the upper mantle for midocean-ridge basalts and the lower mantle for hotspot, plume or ocean-island basalts. In some models the mantle is still grossly homogencous but contains blobs of isoto- pically distinct materials so that it resembles a marble cake. Seismologists recognize large lateral heterogencity in the upper mantle and several major seismic discontinuities. The discontinuities, in the homogeneous-mantle scenario, represent equilibrium phase changes. They could also be, in part, due to changes in chemistry. The ocean and cont- nental lithospheres, or high-velocity layers, may represent material different from the underlying mantle. It is difficult to take data which refer to the present Earth and to extrapolate conditions very far back into the past. The “present is the key to the past” training of Earth scientists leads them naturally to the view that the Earth was always pretty much the same and, if it assumed that the ‘mantle is homogeneous now, it always was. A modification of this view is thatthe crust formed, in early Earth history, from the upper mantle, and the lower mantle is still “pri- ‘mordial” or undifferentiated, and always has been If one considers the circumstances of the Earth’s birth, however, one comes up with a different “simplest” hy- pothesis. The energy of accretion is so great, and the melt- ing temperatures and densities of the products are so dif- ferent, that chemical stratification is the logical outcome. Basalts and the incompatible elements are expected to be ‘concentrated toward the surface, and dense refractory erys- tals are expected to settle toward the interior. ‘These various hypotheses can be tested with modern data. It is only by considering a wide-variety of data that ‘we can narrow the possibilities. The title of this book was not picked casually. This year is the two-hundredth anniversary of the publication of “Theory of the Earth; or an Investigation of the Laws Ob- servable in the Composition, Dissolution, and Restoration ‘of Land Upon the Globe” by James Hutton, the founder of modern geology. It was not until much progress had been made in all the physical and natural sciences that geology could possess any solid foundations or real scientific status. Hutton's knowledge of chemistry and mineralogy was con- siderable, and his powers of observation and generalization were remarkable, but the infancy of the other basic sciences made his “Theory of the Earth” understandably incom- plete. In the present century the incorporation of physics, chemistry and biology into geology and the application of new tools of geophysics and geochemistry has made ge- logy a science that would be unrecognizable to the Found- cr, although the goals are the same. Hutton’s uniformitarian principle demanded an enormous time period for the pro- cesses he described to shape the surface of the Earth, and Hutton could see thatthe differend kinds of rocks had been formed by diverse processes. These are still valid concepts, although we now recognize catastrophic events as well Hutton’s views prevailed over the then current precipitation theory which held that all rocks were formed by mineral deposits from the oceans. Ironically, a currently emerging view is that crystallization from a gigantic magma ocean ‘was an important process in times that predate the visible geological record. Uniformitarianism apparently cannot be carried too fa. The word “theory” is used in two ways. A theory is the collection of facts, principles and assumptions which guide the workers in a given field. Well-established theories from physics, chemistry, biology and astrophysics, as well a from geology, are woven into the Earth sciences. Stu- dents of the Earth must understand solid-state physics, crystallography, thermodynamics, Hooke's Law, optics and so on. Yet these collections of theories do not provide “Theory of the Earth.” They provide the tools for unravel- ing the secrets of the Earth and for providing the basic facts Which in turn are only clues to how the Earth operates. By assembling these clues we hope to gain a better understand- ing of the origin, structure, composition and evolution of cour planet. This better understanding is all that we can hope for in developing a “Theory ofthe Earth.” ACKNOWLEDGMENTS 1 wish to express my sincere appreciation to many col- Jeagues, including students, who have collaborated with me over the years, particularly David Harkrider, Nafi Toks6z, Charles Archambeau, Hiroo Kanamori, Toshiro Tanimoto, Hartmuth Spetzler, Robert Hart, Tom Jordan, Jeff Given, ‘Adam Dziewonski, Charles Sammis, Ichiro Nakanishi, Heenri-Claude Nataf, Janice Regan, Jay Bass and Bemard Minster. ‘The chapters on anisotropy and heterogeneity include ‘material based on joint work with Adam Dziewonski, Janice Regan, Toshiro Tanimoto and Bernard Minster. I thank them and the American Geophysical Union, Physics of the Earth and Planetary Interiors and the Geophysical Journal for permission to use figures and other material‘The chapter on anclasticity is based partly on joint studies with Hiro Kanamori and Jeff Given. 1 thank them and the AGU and Nature for permission to use material from these publications. 1 also thank Earth and Planetary Science Letters and Science for permission to use material from articles I have published in these journals in Chapter 12. prerack — Xiii ‘Thanks are also due to Roberta Eager and Ann Freeman, for working through numerous drafts of the manuscript. Finally I acknowledge my debts to Nancy. While I was doing this she was doing everything else.The Terrestrial Planets Twant 10 know how God created this world. I am not interested in this or that phenomenon, in the spectrum of this or that element. I want to know his thoughts, che rest are details. arth is part ofthe solar system. Although its the most studied planet, it cannot be completely understood in isolation. The chemistry of meteorites and the Sun provide constraints on the composition of the bulk of the Barth. The properties of other planets provide ideas for and tests of theories of planetary formation and evolution. In trying to understand the origin and structure of the Earth, one can take the geocentric approach or the ab initio approach. In the former, one describes the Earth and altempts to work ‘backward in time. For the latter, one attempts to track the evolution of the solar nebula through collapse, cooling, condensation and accretion, hoping that one ends up with something resembling the Earth and other planets. In Chap- ter 1 I develop the external evidence that might be useful in understanding the Earth. In Chapter 2 I describe the Earth and Moon THEORIES OF PLANETARY FORMATION ‘The nature and evolution of the solar nebula and the for- ‘mation of the planets are complex and difficult subjects. ‘The fact that terrestrial planets did in fact form is a sulli- cient motivation to keep a few widely dispersed scientists ‘working on these problems. There are several possible mechanisms of planetary growth. Either the planets were assembled from smaller bodies (planetesimals), a piece at a me, or diffuse collections of these bodies, clouds, became gravitationally unstable and collapsed to form planetary sized objects. The planets, or protoplanetary nuclei, could ‘have formed in a gas-free environment or in the presence of ‘a large amount of gas that was subsequently dissipated. "The planets are now generally thought to have origi- nated in a slowly rotating disk-shaped “solar nebula” of gas ‘and dust with solar composition. The temperature and pres sure in the hydrogen-tich disk decrease both radially from its center and outward from its plane. The disk cools by radiation, mostly in the direction normal to the plane, and ‘part of the incandescent gas condenses to solid “dust” par- ticles. As the particles grow, they settle to the median plane by processes involving collisions with particles in other o- bits, by viscous gas drag and gravitational attraction by the disk. The total pressure in the vicinity of Earth's orbit may hhave been of the order of 10~? to 10 bar. The particles in the plane probably formed rings and gaps. The sedimenta- tion time is fairly rapid, but the processes and time scales involved in the collection of small objects into planetary sized objects are not clear. The common thread of all cos- ‘mogonic theories is that the planets formed from dispersed ‘material, that is, from a protoplanetary nebula. Comets, ‘some meteorites and some small satellites may be left over from these early stages of accretion The following observations are the main constraints on theories of planetary origin: 1, Planetary orbits are nearly circular, ie virtually in a single plane, and orbit in the same sense as the Sun's rotation. The Sur's equatorial plane is elose to the orbital plane. The planets exhibit a preferred sense of rotation. 2. The distribation of planetary distances is regular (Bode’s Law). 3. The planets group into compositional classes related to distance from the Sun, The inner, or terrestrial planets (Mercury, Venus, Earth and Mars), are small, have high density, slow rotation rates and few satellites. The Moon is often classified asa terrestrial planet. The giant plan- ets Gupiter, Saturn, Uranus and Neptune) ae large, have ow density, rotate rapidly and have numerous satellites. Although the Sun contains more than 99 percent of the 1