Published and presented to you By M.H.Najafi www. pardisrasht.blogfa.com GEOTECHNICAL EARTHQUAKE ENGINEERING STEVEN L. KRAMERGeotechnical Earthquake EngineeringPRENTICE-HALL INTERNATIONAL SERIES IN CIVIL ENGINEERING AND ENGINEERING MECHANICS, William J. Hall, Editor ‘Au and Christiano, Fundamentals of Sructural Analysis ‘Au and Christiano, Structural Analysis Barson and Rolte, Fracture and Fatigue Control in Structures, 2/e Bathe, Finite Element Procedures in Engineering Analysis Berg, Elements of Siructural Dynamics Biggs, roduction o Structural Engineering Chajes, Structural Analysis, 2/ Chopra, Dynamics of Sructures: Theory and Applications to Earthquake Engineering Collins and Mitchel, Prestressed Concrete Structures Cooper and Chen, Designing Steel Structures Corting et a, The Art and Science of Geotechnical Engineering Gallagher, Finite Element Analysis Hendrickson and Au, Project Management for Construction Higdon etal, Engineering Mechanics, 2nd Vector Edition Hultz and Kovacs, Inoduction in Geotechnical Engineering Humar, Dynamics of Sructures Joliston, Lin, and Galambos, Basie Steel Design, 3/e Kelkar and Sewell, Fundamentals of the Analysis and Design of Shell Structures Kramer, Geotechnical Earthquake Engineering MacGregor, Reinforced Concrete: Mechanics and Design, 2 Mehta and Monteiro, Concrete: Stcture, Properties and Materials, 2/e Melosh, Structural Engineering Analysis by Finite Elements Meredith et al. Design and Planning of Engineering Systems, /e Mindess and Young, Concrete Nawy, Prestressed Concrete Nawy, Reinforced Concrete: A Fundamental Approach, 2e Prefer, Solid Waste Management Popov, Engineering Mechanics of Solids Popov, Iniroduction to the Mechanics of Solids Popov, Mechanics of Materials, 2/e Schneider and Dickey, Reinforced Masonry Design, 2/e ‘Wang and Salmon, introductory Structural Analysis Weaver and Johnson, Siructural Dynamics by Finite Elements Wolf, Dynami Soil-Siructure Interaction Wray, Measuring Engineering Properties of Soils Yang, Finite Element Structural AnalysisGeotechnical Earthquake Engineering Steven L. Kramer University of Washington Prentice-Hall International Series in Civil Engineering and Engineering Mechanics Prentice Hall Upper Saddle River, New Jersey 07458teary Congres Catan Pubeton ate ee opm heey de Acqison stor BULL STENQUIST Proguctonedioe IRWIN ZUCKER (Cover aiecoe JAYNE CONTE: Boyes JULIA MEEHAN ator assistant: MEG WEST £1956 Pence Hal In Sion & Schuster A Vico Company Sale River, NIGTASS All igs served No part ofthis book iy be reproduced in ny fr or by ary means ‘witout permission a wr the puishe ‘The author and publisher ofthis ok have wed thei bst efor in prepaing this book. These efor ince the developmen rer al stg ofthe thewies and pogo deerme tet efeeveness The sue nd blr make no wary of any kind expresed or pie, with regan these pogras oe he dcumetaton ‘tained this bok. The thr apis shal at be able in any even fore elo conse rages ‘mconneton wih or arnng oa of the ashing perfomance, ewe of hee Progr Printed in he Unite Stes of America ost esas ISBN 0-23-374943-6 ‘Preice Hal keratin (LK) Lite, Loon rence Hal of Austais ty. Lied. Shey rome Hal Cana Ine, Tera rence Hall Hipuaeamesicinn, SA. Meco Preice Hal of fa Private Limited, Now Delhi rece Halo Japan, le. Tokyo Sioa & Schuster Asa Pe Lu Singapore ators Prentice Hal do Bras, La Ro de JaneiroTo my parentstote el Published and presented to you by otohammad cftossein Najaf www.civil.10r.ir www.pardisrasht.blogfa.comContents Preface wv Introduction to Geotechnical Earthquake Engineering 1 L1 Introduction 1 12 Background 1 13° Seismic Hazards 2 13.1 Ground Shaking, 2 12.2 Siruetural Hazards, 3 133° Liquefaction, § 134° Landslides,9 13.5 Retaining Structure Failures, 11 136 Lifeline Hazards, 11 13.7 Townamt and Seiohe Hazards, 13 4 Mitigation of Seismic Hazards 14 LS Significant Historical Earthquakes 4Contents Seismology and Earthquakes 18 21 Introduction 18 22 Internal Structure of the Earth 18 22.1. Seismic Waves, 19 22.2 IntemalStrctur, 20 23 Continental Drift and Plate Tectonics 23 23,1 Plate Tectonics, 24 23.2 Plate Boundaries, 29 24 Faults 33 24.1 Fault Geometry. 33 24.2 Fault Movement, 34 25 Elastic Rebound Theory 36 25.1 Relationship 10 Earthquake Recurrence, 39. 252 Relationship 1o Tectonic Environment, 41 25.3 Seismic Moment, 42 26 Other Sources of Seismic Activity 42 27 Goometrie Notation 43 28 Location of Earthquakes 44 29 Size of Earthquakes 45 2.9.1 Earthquake Intensity, 45 2.92 Earthquake Magnitude, 46 29.3 Earthquake Energy. 50 210 Summary 51 ‘Strong Ground Motion 54 3.1 Introduction $4 32 Strong-Motion Measurement 56 321 Seismographs, 56 322 Data Acquisition and Digitization, 39 323° Surong-Motin Processing, 61 324 Strong-Motion Instrument Arrays, 62 325. Swrong-Motion Records. 66 33. Ground Motion Parameters 65 33. Ampliude Parameters, 66 332 Frequency Content Parameters, 70 33.3 Duration, 79 334 Other Measures of Ground Motion, 82 33S Discusion, 84Contents ix 344 Estimation of Ground Motion Parameters 84 34.1 Magnitude and Distance Egects, 85 342 Development of Predictive Relationships, 86 343° Esimation of Amplitude Parameters, 88 S44 Estimation of Frequency Content Parameters, 91 34.5 Esimationof Duration, 95 346. Estimation of Other Parameters, 95 35 Spatial Variability of Ground Motions 100 36 Summary 102 4 Seismic Hazard Analysis 106 4.1 Introduction 106 42 Identification and Evaluation of Earthquake Sources 107 42.1 Geologie Bvidence, 107 42.2 Tectonic Evidence, 13 423° Historical Seismicity, 113 424 Instrumental Seismicity, 14 43 Deterministic Seismic Hazard Analysis 114 44° Probabilistic Seismic Hazard Analysis 117 44.1 Earihquate Source Characterisation, 118 44.2 Predictive Relationships, 126 443 Temporal Uncertainty, 127 444 Probability Computations, 129 45 Summary 138 5 Wave Propagation 143 5.1 Introduction 143 52 Waves in Unbounded Media 14 52.1 One-Dimensional Wave Propagation, 144 52.2 Three-Dimensional Wave Propagation, 149 5.3. Waves in a Semi-infinite Body 156 S31 Raleigh Waves, 156 532 Love Waves, 162 52.3 Higher-Mode Surface Waves, 164 SSA Dispersion of Surface Waves, 166 SIS Phase and Group Velocities, 165 SA Waves in a Layered Body 165 341 One-Dimensional Case: Material Rowndary in an Infinite Rad, 165 542 Three-Dimensional Case: Inclined Waves, 170 55 Attenuation of Suess Waves 174 551 Material Damping, 175 5532 Radiation Damping, 179 36 Summary 180x Contents 6 Dynamic Soil Properties 184 6.1 Introduction 184 62 Representation of Stress Conditions by the Mohr Circle 185, 6.2.1 Principal Siresses, 187 622 Siress Paths, 188 63 Measurement of Dynamic Soil Properties 191 6.3.1 Field Tests 197 632 Laboratory Tests 215, 6.3.3 Inerpretaion of Observed Ground Response, 228 64 Suess-Struin Behavior of Cyelically Loaded Soils 228 64.1 Some Basic Aspects of Particulate Matter Behavior, 228 64.2 Equivalent Linear Model, 230 643 Cree Nonlinear Models, 240 644 Diseusion, 244 65 Strength of Cyeclically Loaded Soils 244 65.1 Definitions of Ftlur, 244 63.2 Crtie Stengih, 245 63.3 Monotonic Strength, 246 66 Summary 248 7 Ground Response Analysis 254 7.1 Introduction 254 712 One-Dimensional Ground Response Analysis 255 7.21 Linear Approach, 256 7.22 Nonlinear Approach, 275 7.23 Comparison of One-Dimensional Ground Response Analyses, 279 73 Two-Dimensional Ground Response Analysis 280 7.3.1 Dynamic Finte-Element Analsis,281 7.32 Equivalent Linear Approach, 284 7.33 Nonlinear Approach, 286 7.34 Other Approaches 10 Two-Dimensional Ground Response Problems, 286 7.35 Comparison of Two-Dimensional Ground Response Analyes, 201 74 Three-Dimensional Ground Response Analysis 291 74d Equivalent Linear Finite-Element Approach, 292 742 Nonlinear Finite-Element Approach, 292 743 Shear Beam Approach, 293 744 Comparison of Three-Dimensional Ground Response Analyse, 24 75 Soil-Stucture Imeraetion 294 75.1 Illustration of Soil-Structure Imeraction Feet, 295 752. Methods of Analysis, 300 7.6 Summary 303Contents 8 Local Site Effects and Design Ground Motions 8.1 Introduetion 308 8.2 Bifeets of Local Site Conditions on Ground Motion 309 82.1 Evidence from Theoretical Ground Response Analyses, 309 822 Evidence from Measured Amplification Functions, 310 823 Evidence from Measured Surface Motions, 372 824 Compilations of Data on Local Se Effects, 317 52.5. Effects of Surface Topography and Basin Geometry, 319 83 Design Parameters 323, 831 Design Earthquakes, 324 83.2 Design Spectra, 325 84 Development of Design Parameters 327 84.1 Site Specific Development, 327 842 Code-Based Development 328 85 Development of Ground Motion Time Histories 340 85.1 Modification of Actual Ground Motion Records, 340 852 Time-Domain Generation. 341 85.3 Frequency Domain Generation, 43 854 Green's Function Techniques, 343 855 Limitations of Arial Ground Motions, 345 86 Summary 345 9 Liquefaction 9.1 Introduction 348 9.2 Liquefaction-Related Phenomena 349 92.1 Flow Liquefaction, 349 922 Cyclic Mobily, 349 9.3 Evaluation of Liquefaction Hazards 350 94 Liquefaction Susceptibility 351 D4.) Historical Criteria, 352 942 Geologie Criteria, 353 943 Compositional Criteria, 354 944 State Criteria, 355 95 Initiation of Liquefaction 361 95.1 Flow Liquefaction Surface, 361 952 Influence of Excess Pre Pressure, 355 95.3 Evaluation of ition of Liquefaction, 268 9.6 Effects of Liquefaction 397 96.1 Alteration of Ground Motion, 398 962 Development of Sand Boils, 409 963° Selemont 402 964 Instatiley, 408 97 Summary 417Contents 10 Seismic Slope Stability 423 10.1 Introduction 423 10.2 ‘Types of Barthquake-Induced Landslides 424 10.3 Farhquake-Induced Landslide Activity 426 10.4 Evaluation of Stope Stability 429 10.5 Satie Slope Stability Analysis 430 10.3.1 Limit Equiiriam Analysis, 430 1052 Sivess-Deformation Analyses, $33 10.6 Seismic Slope Stability Analysis 433, 10.1 Anatsis of Inertial Instability. 433 105.2 Analysis of Weakening fntbiliy, 450 107 Summary 462 11. Seismic Design of Ret 466 us 116 118. 119 Introduction 466 Types of Retaining Walls 466 ‘Types of Retaining Wall Failures 467 Static Pressures on Retaining Walls 469 Hh Rankine Theory, 469 1142 Coulomb Theory, 72 143° Logariehmie Spiral Method, 474 LA. Stress Deformation Analisis, 76 Dynamic Response of Retaining Walls 477 Seismic Pressures on Retaining Walls 477 1146.1 Yielding Walls, 478 1.6.2 Nomielding Walls, 484 116.3 Effects af Water om Wall Pressures, $86 1164 Fioite-Element Analysis, 489 Seismic Displacements of Retaining Walls 489 H121- iehards-Bins Method, 389 11.22 Whiimas-Liao Method, 92 11.7.3. Finie-Blemene Analysis, 493 Seismic Design Considerations 494 HBS Gravity Walls, 498 1182 Canmitever Walls, 195 118.3 Braced Walls, 95 TSA Reinforced Sot Walls, 300 Summary 503Contents 12 Soil Improvement for Remediation of Seismic Hazards 506. 12.1 Inroduetion 506 122 Densification Techniques 507 122.1 Vibro Techniques, 508 12.22 Dynamic Compaction, $10 3 Blasting, 512 4 Compaction Growing, 313 1225. Aveal Extent of Densification $14 123 Reinforcement Techniques $15 12.3.1. Stone Columns, 515 1232 Compaction Piles, 516 1233 Drilled Inclusions, 516 124 Grouting and Mixing Techniques 517 P41 Growing, S18 1242 Mixing, 519 125 Druinage Techniques 521 12.6 Verification of Soil Improvement 5: 126.1 Laboratory Testing Techniques, 5 1262 InSitu Testing Techniques, 523 1263 Geophysical Testing Techniques, 523 127 Other Considerations 524 128 S24 A Vibratory Motion 527 Al Introduction 527 A2 —Typesof Vibratory Motion 527 21 Simple Hannonic Motion, 328 422 Trigonometrc Nowation for Simple Harmonie Motion, 329 A23 Other Measures of Moron. 533 AB FourierSeries 536 ABI Trigonometric Form, 536 A32 Exponential Form, 539 AS3 Diserete Fourier Transform. 541 ABA Fast Power Transform, S41 ABS Power Spectrum, S42 B Dynamics of Discrete Systems 543 But Introduction 543 B2 Vibrating Systems S44 B3 —Single-Degree-of-Freedom Systems 544Contents B4 Equation of Motion for SDOF System 545, B4.1 Equation of Moron: Exemal Loading, 545 BA2 Equation of Moron: Vibration of Support (Base Shaking), $47, BS Response of Linear SDOF Systems 547 BSI Undamped Free Vibrations, 548 BS2 Damped Free Vibrations, S51 BS3_ Response of DOF Systems to Harmonic Loading, $53 BSA Response of SDOF Systems to Periodic Loading, $61 BSS _ Response of SDOF Systems to General Loading, 564 B6 Damping 567 B61 Viscous Damping, 567 B62. Other Measures of Energy Dissipation, 569 B63 Complex Siffuess, 570 B7 Response Spectra $71 B8 Response of Nonlinear SDOF Systems to General Loading $72 BSI Incremental Equation of Motion, 572 BS2 Numerical Integration, 574 9 Multiple-Degree-of Freedom Systems $75, B.A Equations of Motion, 575 B92 Undanped Free Vibrations, 577 B93 Mode Superposition Method, 379 B94 Response Spectrum Analysis, $8] BIS Discussion, 582 Probability Concepts 583 C1 Introduction $83 C2 Sample Spaces and Events 583 C3 Axioms of Probability 584 C4 Probabilities of Events $85 C5 Random Variables 588 C6 Expected Values and Standard Deviations 589 7 Common Probability Distributions 590 C71 Uniform Distribution, 390 C72 Normal Disribuion, $91 C73 Lognormal Distribution, $94 References 596 Index 643Preface ‘Compared to most disciplines of civil engineering, geotechnical earthquake engineering is Quite young. While the damaging effects of earthquakes have been known for centuries, the strong contribution of soils to the magnitude and pattern of earthquake damage was not widely appreciated until relatively recently, Following damaging earthquakes in 1964 in Niigata, Japan and Alaska, and spurred by the growth of the nuclear power industry in the 1960s and 1970s, the field of geotechnical earthquake engineering has grown rapidly Although much remains to be learned, the field has matured tothe point where generally ‘accepted theories and analytical procedures now exist for many important problems. ‘The purpose of this book isto introduce the reader tothe concept, theories, and pro- cedures of geotechnical earthquake engineering. Its intended for use as a text in graduate courses on geotechnical earthquake engineering and as a reference book for practicing engic neers, Recognizing that geotechnical earthquake engineering is a broad, multidisciplinary field, the book draws from seismology, geology, structural engineering, risk analysis, and ‘other technical disciplines, ‘The book is writen a a level suitable fr students with knowledge equivalent to that of a senior (fourth-year) civil engineering student, The student should have had basic courses in soil mechanics, structural engineering, and hydraulics; introductory courses in geology and probabilty/statstcs would also be helpful. Many graduate students will havewi Preface ‘had courses in structural dynamics or soil dynamics by the time they begin study of geo- technical earthquake engineering. For those readers without prior exposure, introductions to the nomenclature and mathematics of dynamic systems, structural dynamics, and prob- ability ate presented in tree apper ORGANIZATION ‘The subject matter falls into two main categories. The appendices and the frst six chapters present fundamental principles of seismology, ground motion, dynamics, and sol behavior. Applications ofthese principles tothe practical problems most commonly encountered in eotechnical earthquake engineering practice are presented in the last six chapters, Caper 1 introduces the reader to the types of damage that can occur during earth- ‘quakes and tothe problems they present to geotechnical earthquake engineers. Basic con- ‘cept of earthquake seismology and the terminology used to describe earthquakes and their effects are described in Chapter 2. Chapter 3 describes ground motion measurement, the ‘parameters used to characterize strong ground motion, and methods for prediction of those ‘parameters. Deterministic and probabilistic seismic hazard analyses are presented in Chap- ter 4. Chapter 5 introduces the reader to wave propagation, beginning with simple one- «dimensional body waves in homogeneous materials and extending to surface waves and ‘multidimensional, layered systems. The properties of soil that control their wave propaga tion behavior ae described in Chapter 6, Field and laboratory techniques for measurement ofthese properties are also described, ‘Chapter 7 presents methods for analysis of ground response during earthquakes, beg ning with one-dimensional ground response analysis and moving through two- and three- «dimensional dynamic response analyses. Both frequency’ and time-domain approaches are described. Chapter 7 concludes with an introduction tothe basic concepts and elfects of soil- Structure interaction. The effects of local soil conditions on ground motions and earthquake damage are described in Chapter 8. Chapter 8 also introduces the concept of design ground ‘motions, and how they are obtained from site-specific analyses and from building codes. ‘Chapter 9 deals with iquefsetion—it begins with a conceptual framework for understanding various liquefaction-related phenomena and then presenis practical procedures for evalua tion of iquetaction hazards. Seismic stability of slopes is covered in Chapter 10, and seismic design of retaining structures in Chapter 11. Chapters 10 and 11 address thei respective topics initially from pseudo-staic and then from permanent displacement standpoints (Chapter 12 introduces commonly used soil improvement techniques for mitigation of seis- ‘mic hazard, PEDAGOGY ‘This book isthe first to deal explicitly withthe topic of geotechnical earthquake engi ing. During its preparation, a great deal of time and effort was devoted to decisions regan ing content and organization. The final form naturally reflects my own preference, but the texthas been reviewed by many engineers ftom both academis and professional practice. Preparation of the text also involved a great deal of interpretation of information from aPreface vil ‘wide variety of sources. While the text efleets my own interpretation of tis information, it is heavily referenced to allow readers o explore background or more detailed information ‘on various geotechnical earthquake engineering topics. ‘A couple features are noteworthy. Two ground motions from the Loma Prieta earth- «quake, one from a rock outcrop and one from the surface of a nearby deep soil deposit ae used to illustrate a number of concepts throughout the book. Differences in the amplitudes, frequency contents, and durations ofthe motions are emphasized in Chapter 3. The reasons for these differences later become apparent jn Chapters 7 and 8, Te book also emphasizes the use of ansfer functions, particulary in te solution of ground response problems. The transfer function approach helps students form a more complete understanding of ground response—in the frequency domain as well asthe time domain. With the advent of com- puter programs such as MATLAB, MathCad, and Mathematica, the Fourier analyses ‘required in the transfer function approach are quite simple; students use MATLAB exten- sively in my soil dynamics and geotechnical earthquake engineering courses. ‘The book contains worked examples and homework problems. The example prob- Jems are intended to illustrate the basic concepts of the problems they address; to allow the results to be checked, @ number involve calculations carried out to more significant Figures than the accuracy of the procedures (and typical input data) would justify. Many of the ‘important problems of geotechnical earthquake engineering, however, do not lend them= selves tothe type of short, well-lefined homework problem that is readily placed ina book. My preference isto assign longer, project-oriented assignments based on actual case histo- ries, and {recommend thatthe homework problems inthis book be supplemented by such assignments unrrs As in many other fields, the use of units in geotechnical eathquake engineering is neither uni- {orm nor consistent. The current state of knowledge in geotechnical earthquake engineering dnas resulted from advances in a variety of technical fields and a variety of counties, many of ‘which customarily use different unis. Fortunately, most conform to relatively standard met- ric or British systems. Rather than attempt to force the use of ane system or the other, this book uses dual unit, In recognition oftheir origins, the most common units for each quantity is listed first withthe alternative following in parentheses. The approach is intended to allow all readers to proceed through the material without stopping to convert (mentally or other- wise) from one set of units to another. To encourage familiarity with both sets of units, some ‘example and homework problems ate specified in metric units and some in British units. ACKNOWLEDGMENTS: ‘A number of people have helped directly and indirectly in the preparation of this book. The professional and academic portions of my career have benefited greatly from a number of people that I have worked closely with, particulary (in chronological order) Bill Houston, ‘Tom Tejima, H.B. Seed, Joe Mahoney, and Bob Holtz. Their assistance, advice, and encouragement has taken many forms, and Tam grateful . each,iit Preface As this book evolved from a collection of lecture notes and handouts, it was contin- ually improved by comments and suggestions from many students in my soil dynamics and geotechnical earthquake engineering courses. Their assistance is greatly appreciated, 1am also grateful to many colleagues who provided constructive ertical reviews of different parts ofthe book, including Dr. Donald G, Anderson, Pr. Juan Baez, Mr. David Baska, De. Gopal Biswas, Prof, Ross W. Bovlanger, Dr. C.B. Crouse, Prof, Emeritus William J. Hall, ‘Ms. Karen Henry, Prof. Carlton L. Ho, Prof, William D, Kovacs, Prof. Roberto 'T. Leon, Prof. Gregory R: MacRae, Dr. Lelio H. Mejia, Dr. Robert Pyke, Prof, Peter K, Robertson, Prof. Raj Siddharhan, Prof. Stewart Smith, Prof. Timothy D. Stark, and Prof. George M. Turkiyyah. Each made suggestions that improved the quality ofthe book. Prof, Geoflrey R. Martin and Prot.T, Leslie Youd reviewed substantial portions ofthe book; theirefforts are particularly appreciated, Finally. Lam most grateful to my wife, Diane, and to my daughters, Katie and Megan, Preparation of this book involved several years of long working hours; { could not have \done it without their cheerful understanding and encouragement,Introduction to Geotechnical Earthquake Engineering Leese | 1.1 INTRODUCTION Earthquake engineering deals withthe effects of earthquakes on people and their enviton- ment and with methods of reducing those effects tis very young discipline, many ofits ‘most important developments having occurred inthe past 30 040 years, Earthquake engi> neering isa very broad field, drawing on aspects of geology, seismology. geotechnical engineering, strdtural engineering, isk analysis, and other technical fields ls practice also requires consideration of social, economic, and politcal factors. Most earthquake engineers have entered the field from structural engineering or geotechnical engineering back. ‘rounds, a fact that is reflected in the practice of earthquake engineering. This book covers feotechnical aspects of earthquake engineering. Although its primary audience is geotech~ nical engineering students and practitioners, it contains a great deal of information that should be of interest to the structral engineer and the engineering seismologist 1.2 BACKGROUND ‘The study of earthquakes dates back many centuries. Written records of earthquakes in China date s far back as 3000 years. Japanese records and records from the eastern Med. iterranean region go back nearly 1600 years. In the United States the historical record of 12 Introduction to Geotechnical Earthquake Engineering Chap. 1 \ “0 Le i <> (ee jee) 2 ; aoa Figure 221. (4) Survey ines rss San Anes and Calaers fut California {byctange in chord enh extension postive) (Fm Eathguates by Bolt Cpyeght {© 1908 by WH. Freeman apd Compan Used wih pesmision) process ofthe successive buildup and release of strain energy in the rock adjacent faults, Is often ilustrated as shown in Figure 2.22, The nature of the buildup and release of stress is of interest. Faults are not uniform, either geometrically or in terms of material properties—both strong and weak zones can exist over the surface of a fault, The stronger zones, refered to as asperiies by some (Kan- mori and Stewart, 1978) and barriers by others (Aki, 1979), are particulsely important. The ‘asperity model of fault rupture assumes that the shear stresses prior to an earthquake are not ‘uniform across the fault because of stress release inthe weaker Zones by creep oF foreshocks Release ofthe remaining stresses held by the asperiies produces the main earthquake that leaves the rupture surface in a stale of uniform stress, In the barrier model, the pre-eath- quake stresses on the fault are assumed to be uniform. When the main earthquake occurs, stresses are released from all parts of the fault except forthe stronger barriers; afershoeks then occur as the rock adjusts tothe new uniform stress field. Since both Foreshocks and aftershocks are commonly observed, it appears that some strong zones behave as asperites and others as barriers (Aki, 1984). The engineering significance of asperites and barriers lies in their influence on ground. shaking characteristics close tothe fault. site located close to one of these strong zones may experience stronger shaking than a site equSec. 2.5 Elastic Rebound Theory 39 Detormaton of duct rock Detonation ofa ductie tek { oY | Lo « tI Fracture of bila rock Fracture ofa bit stick LI Figure 2.22 laste rebound bay feats. 1a (te slow deformation of rock inthe vicinity of plat boundary ress in uildup of tain energy in the ek Ine ‘sme way hat stn ener bls pin cit ck defied se show onthe ht, Ihe sweagh ofthe rock exceed wl aptre,reeaing tain nergy in he or ‘of vibrations, mich asthe energy nth ick wld be eleased who the sk bres, Alter the arthgake the rocks iplced rom ts xn poston, The al elaine thsplacemen ofthe lates she um ofthe sp dspace at the fait ant posse sisplasments doe o warping disirion of the edges ofthe Paes cate ful Air Foster, RJ. General Geology, Se © 1988. Adapted by pein of Pete Hal Upper Sade River, New ey. to the fault but farther froma strong zone, At larger distances from the fault the effects of fault nonuniformity deerease. Unfortunately, methods for locating these strong zones prior to rupture have not yet been developed ‘Rupture generally progresses across fault as a series of dislocations (some multiple «event earthquakes can be thought of as a series of small earthquakes that occur in close spa- tial and temporal proximity). Small earthquakes can be modeled as point processes since their rupture surfaces usually span only a few kilometers. Large earthquakes, however, can rupture over distances of tens, or even hundreds, of kilometers, and the nature of ground shaking can be influenced by the characteristics ofthe rupture process. For example, waves emanate from the fault with different strengths in different directions; such directivity effects can produce azimuthal differences in ground motion characteristies (Benioff, 1955; Ben-Menachem, 1961). Constructive interference of waves produced by successive dislo- cations can produce strong pulses of large displacement called fling (Figure 2.23) at nearby sites toward which the rupture is progressing (Benioff, 1955; Singh, 1985). 2.5.1 Relationship to Earthquake Recurrence ‘The theory of elastic rebound implies that the occurrence of earthquakes will relieve siresses along the portion ofa fault on wich rupture aceur, and that subsequent rupcure40 Seismology and Earthquakes Chap. 2 sitar in Docion Away, Reaulan in Diresion From Direction of Fault Rupture ‘of Fault Rupture Figure223 Schema israon of dieting et on gourd motions a ites ‘od and away rom rection alt rape. Overlppig of ples ca al to ‘toning pls aes wand which he fal rape (Aer Sigh, 18S; wsed by will not occur on that segment until the stresses have had time to builel up again. The chances ofan earthquake occurring on a particular fault segment should therefore be related in some way to the time that has elapsed since the last earthquake and, pethaps, to the amount of energy that was released. Ina probabilistic sense, then, individual earthquakes on ‘8 particular fault segment should not be considered as random, independent events, This characteristic is important in seismic hazaed analysis ‘Because earthquakes relieve te strain energy that builds up on faults, they should be more likely to oecurin areas where lite or no seismic activity has been observed for some time. By plotting fault movement and historical earthquake activity along a fault itis pos- sible to identify gaps in seismic activity at certain locations along faults. According to elas fic rebound theory. either the movement is occurring aseismically or strain energy is building in the vicinity of these seismic gaps. In areas where the later is known to be the case, seismic gaps should represent the most likely locations for future earthquakes. A num- ber of seismic gaps have been identified around the world and lage earthquakes have sub- sequently been observed on several of them. The 1989 Loma Prieta earthquake occurred on ‘asegment ofthe San Andreas fault that had previously been identified as a gap. as shown in Figure 2.24. The use of seismic gaps offers promise for improvement in earthquake predic- tion capabilities und seismic risk evaluation,Sec.2.5 Elastic Rebound Theory a 50 100 180 200 DISTANCE (mi) Figure224 Cross section ofthe San Ants al tom not of Sua Francie ‘oath of Pate) smi athe 20 years pie tote 1989 Lome Pets ‘acu is shown withthe Loma Pret ap hshlihicd: (rin shock ope cic) tnd aftershocks ofthe Loma Pica earthquake, Note the temaising gaps beeen San Francisco an Porta Valley and seth of Patel (Ae: Hse et 190), 2.5.2 Relationship to Tectonic Environment Blastic rebound also implies that tectonic environments capable of storing different ‘amounts of energy will produce earthquakes of Wifferent size, Consider, for example, the tectonic environment in the vicinity ofa spreading ridge plate boundary. Fist, the crust thins hence the volume of rock in which strain energy ean build ups small. Second, the hor- ‘zontal component ofthe relative plate movement is extensional; hence the normal stress on the fault plane, and with it the rupture strength, is low. Thiel, the rocks relatively warm and luctle, soit will not release strain energy suddenly. Taken together, these factors limit the {otal strain energy that can build up and be suddenly released ata spreading ridge boundacy. ‘These factors explain the observed absence of very large earthquakes at spreading ridge boundaries By the time the oceanic crust has moved from a spreading ridge toa subduction zone, ithhas cooled and become much thicker and stronger. Relative movement of the plates is toward each other, so high compressive normal stresses inerease the rupture strength on the fault plane, Because subduetion zone plate boundaries are inclined, the potential rupture area is large, All of these factors support the potential buildup of very large amounts of strain energy that, when suddenly released, can produce great earthquakes. Infact, the larg- est recorded earthquakes have been produced by subduction zones. At transform faults, the rock is generally cool and britle, but large compressive sresses donot usually develop because the Faults are often nearly vertical and movement is ‘ypically of a strike-slip nature. Because the depth of transform faulting i limited, the total2 Seismology and Earthquakes Chap, 2 amount of sain energy that can be stored is controled by the length of rupture. Very large ‘earthquakes involving rupture lengths of hundreds of kilometers have been observed on transform faults, but truly “great” earthquakes may not be possible. 2.5.3 Selsmic Moment ‘The concept of elastic rebound theory can be used to develop a useful measure of the size ‘of an earthquake. The seismic moment of an earthquake is given by My = AD en where isthe rupture strength of the materia along the ful, A the rupture area, and B the average amount of sip. The seismic moments name for its units of fore times length however, it is more a measure ofthe work done by the earthquake. As such, the seismic ‘moment correlates well with the energy released during an earthquake. The seismic moment ‘ean be estimated from geologic records for historical eahquakes, or obtained from the long-period components of a seisnogram (Bullen and Bolt, 1985) 2.6 OTHER SOURCES OF SEISMIC ACTIVITY ‘The sudden release of strain enerey by rupture of the rock at plate boundaries isthe primary cause of seismic activity around the world. There are, however, other sources of seismic aetivity that produce smaller earthquakes that maybe important in localized regions Earthquakes have been associated with volcanic aetvity. Shallow volcanic earth- quakes may result from sudden shifting or movement of magma. In 1975, a magnitude 12 ‘earthquake on the big island of Hawai produced significant damage and was followed shortly by an eruption ofthe Kilauea voleano. The 1980 eruption of Mt. St Helens in south «em Washington was actually wiggered bya small (M, = 5.1), shllow, volcanic earthquake that triggered massive landside on the noth slope of the volcano. The unloading of the north lope allowed the main eruption to occur approximately 30 sc later. Volcanic erup- tions themselves can release tremendous amounts of energy essentially athe earth's sur- face and may produce significant ground motion Seismic waves may be produced by underground detonation of chemical explosives ‘or nuctear devices (Bot, 1975). Many significant developments in seismology during the ‘Cold War years stemmed from the nged to monitor nuclear weapons testing activites of ‘other countries. Collapse of mine or cavern zoofs, or mine bursts, can cause simal lcal ‘arthquakes, aan large landslides. A 1974 landslide involving 1.6% 10" m'2.1 10? ya) ‘of material along the Montaro River in Peru produced seismic waves equivalent those of ‘magnitude 4.5 earthquake (Bolt, 1989). Reservoit-nduced earthquakes have been the subject of considerable study and some controversy, Local seismicity increase significantly after te filling of Lake Mead behind Hoover Dim on the Nevade-Aizona bord in 1938. When the Koyna Dat (Indi) reser- ‘oir was filled, local shallow earthquakes became common in an area previously thought Ihave been virtually aseismic, In 1967, five years after filing ofthe Koyna reservoir hadSec. 2.7 Geometric Notation ry begun, a magnitude 6.5 earthquake killed 177 persons and injured more than 2000 more. Local seismicity has been observed to increase seasonally with seasonal increases in reser. voir level. In 1975, seven years after the filling of Oroville Dam in an area of low historical seismicity in northern California, a swarm of earthquakes culminated in a magnitude 5.7 tain shock. After construction of the High Dam, a magnitude 5.6 earthquake occurred in Aswan, Egypt where very litle significant seismic activity had been observed inthe 3000. Year history ofthe area. In these cases, seismic activity appears to have been triggered by the presence of the reservoir. While the effect of the weight ofthe impounded waters likely to bbe negligible at the depths of the induced seismic activity, an increase in porewater pressure that migrates asa “pulse” away from the eservoir after filling may have been sufficient to ‘reduce the strength ofthe rock tothe point where rupture could occur. 2.7 GEOMETRIC NOTATION ‘To describe the location of un earthquake, it is necessary to use accepted descriptive termi- nology. Farthquakes result from rupture ofthe rock along a fault, and even though the up ture may involve thousands of square kilometers of fault plane surface, it must begin somewhere. The point at which rupture begins andthe first seismic waves originates called the focus, or hypocenter. of the earthquake (Figure 2.25). From the focus, the rupture spreads across the fault at velocities of 2 to 3 kan/see (1.2 to 1.9 miles/sec) (Bolt, 1980), Although fault supture can extend tothe ground surface. the focus is located at some foced depth (or hypocentral depth) below the ground surface. The point on the ground surface dlitectly above the focus is called the epicenter. The distance on the ground surface between an observer or site and the epicenter is known as the epicentral distance, and the distance between the observer and the focus is called the focal distance or hypocentral distance, Ground surtace "Ce Epicorra stance Ste or observer Hypocontal distance Foes orpocenter SG ZIN Figure 225 Notion fo description of athqke heaton“ Seismology and Earthquakes Chap, 2 2.8 LOCATION OF EARTHQUAKES ‘The location of an earthquake i often initially specified in terms of the location oft epi- ‘center. Preliminary epicentral location isa simple and straightiorward matter, but efinement ‘ofthe final focation ean be considerably more complex, Preliminary location is based on the ‘relative avival times of p- and s-waves ata set of at least thee seismographs. Since p-waves travel faster than s-waves, they will arrive frst ta given seismograph, ‘The difference in artival times will depend on the difference between the p- and s-wave velocities, and om the distance between the seismograph and the focus of the earthquake, acconding as, -, e2 ‘where a, is the difference in time between the first p- and s-wave arrivals, und vp and y, are the p-and s-wave velocities, respectively. In bedrock, p-wave velocities are generally 3 08 km/sec (1,9 to 5 miles/sec) and s-wave velocities range from 2 to 5 kmisee (1.2 to 3.1 rilessee). Atany single seismograph i is posible to determine the epicentral distance but not the direction of the epicenter. This limited knowledge is expressed graphically by plot- ting a circle of radius equal tothe epicentral distance, When the epicental distance from a second seismograph is plotted as a circle about its location, the possible location of the epi- center is narrowed to the ewo points of intersection ofthe circles. Obviously, a third seis ‘mograph is necessary to identify the most likely location of the epicenter as illustrated in Figure 2.26. More refined estimates of the epicentral, or hypocentra, location ate made ‘sing multiple seismographs, a three-dimensional seismic velocity model ofthe earth, and ‘numerical optimization techniques. The accuracy of these techniques depends on the ‘number, quality, and geographic distribution ofthe seismographs and on the accuracy of the seismic velocity model (Dewey, 1979) P-Stime shows that eartnquake cecured ‘al ths distance fram tation A Epconter Figure2.26_ Preliminary locaton of eicates rom diferent! wave aval tine messtrmeats at sismographs A,B, an Most ely spent acto a the imseton of te te eee, (ter Foster. RL, General Geology. Se © 1988 Adapted by permission of Preaie Hal Upper Sale River, New ene.)Sec.2.9 Size of Earthquakes 4s 2.9 SIZE OF EARTHQUAKES The “size” of an earthquake is obviously a very important parameter, and it has been «described in different ways, Prior to the development of modern instrumentation, methods for charatericing te size of earthquakes were based on crude and qualitative descriptions ofthe effects of the earthquakes. More recently, modern seismographs have allowed the development of a number of quantitative measures of earthquake size. Since several of these measures are commonly used in both seismology and carthquake engineering, the d- ‘inguishing features of each should be understood. 1 Earthquake Intensity 24 ‘The oldest measure of earthquake size isthe earthquake intensity. The intensity is a quali tative deseription of the effects of the earthquake a a particular location, a evidenced by ‘observed damage and human reactions at chat location, Because qualitative descriptions of the effects of earthquakes are available throughout recorded history the concept intensity an be applied to historical accounts estimate the locations and sizes of earthquakes that ‘occurred prior to the development of modern seismic instruments (preinstrumental earth. quakes). This application has been very useful in characterizing the rates of recurrence of earthquakes of different sizes in various locations, aeritical step in evaluation ofthe like- lihood of seismic hazards. Intensities can also be used to estimate strong ground motion lev els (Section 3.3.11), for comparison of earthquake effects in different geopraphie regions, and for earthquake loss estimation, ‘The Rossi-Forel (RF) scale of intensity, describing intensities with values ranging from I to X, was developed in the 1880s and used for many years. It has largely been replaced in English-speaking countries by the modified Mercalli intensity (MMI) scale ‘originally developed by the Kalin seismologist Mercalli and modified in 1931 to better ep resent conditions in California (Richter, 1958). The MMI scale iy illustrated in Table 21 ‘The qualitative nature of the MMI scale is apparent from the descriptions of each intensity level ‘The Japanese Meteorological Agency (JMA) has its own intensity scale, and the Medvedev-Spoonheuer-Karnik (MSK) scale is used in central and eastern Europe. A com parison of the RF, MMI, JMA, and MSK scales is shown in Figure 2.27 ‘Earthquake intensities are usually obtained from interviews of observers after the {event The interviews are often done by mail, but in some seismically active areas, perma. nent observers are organized and trained to produce rational and unemotional accounts of| ‘round shaking. Since human observers and structures are seatered more widely than any scismological observatory could reasonably hope to scatter instrument, intensity observa. tions provide information that helps characterize the distribution of ground shaking in a region, A plot of reported intensities a different locations on a map allows contours of equal intensity, or ésoseisms, to be plotted, Such a map is ealled an isosetsmal map (Figure 2.28), ‘The intensity is generally greatest inthe vicinity of the epicenter ofthe earthquake, and the term epicentral intensity is often used as a crude description of earthquake size. Isoscismal maps show how the intensity decreases, of attenuates, with increasing epicentral distance,6 Seismology and Earthquakes Chap, 2 Table 241 Modified Marcall Intensity Sle of 1931 1 Novfettexcept by avery few nde especialy favorable ckeunstances Feley only ae persons ates especially on upp floes of bling deticuely spend objets may sing Fettguitenocesbly indoors, especialy on upper lors of bilings, but many people dot teotgnize it as ancarngunke: Sanding moter cars my ook lghy vibration ike passing ‘of mck: dation ested ry Dering he dy indoors by many, cudoos by few a night some awakened shes ‘windows, dors disturbed: walls mke cracking oan sensation ike Rev ek siking icing: andig motorcar rocked noterably ¥ Petty nly everyone, any akene some dss, windows, te, ken: eins ‘of ack plaster unstable objets verted dintarance of ec, pls aber objets sometimes nocd: penal lols may stop V1 ettby al nay highend dru tors; some any faite ved & few instants of fallen plaster or damaged choineys: damage sight Everybody nus outdoors; damsge nope in uid of god design td omtucon, slg wo moderate in webu odinaey series, considera in ply ul al) "signed strstr: some chimneys broken: nid by persons diving motr cats Damage sph in peily designed sauces, considerable in xdnaty substan ‘ulings, with paral ols. retin poly ull sauces: panel walls throwa out of fame Sucre: al of himneys. actry sacks, columns ponumens, wal; bar Fonte ove tare; saad ad md jet in ml oun changes in we water: persons ing me ‘ar dir Damage considerable in specially designed strctres; wel dvied ame rates uf punts eea in substantial lings, with paral collapse buns sited of ound x_n; ground racked conspieusly unearoun pipes woken Some wel. but Wooden sts desoyed: mot masonry ad rape suuctues deste with foudations: round bly rack als ben ansios considerable fom ive ats Xt anstep slopes; shifted sand and mad: wate splased ove banks Fei ay (masonry) stoctres remain standing: idges dese: broad Fissure in ron: underground pipelines completly cut vce car slumps andl sips nsf round XM sallsbem greatly ‘Damage otal practically all wrks of contro re dumages greatly or destroyed waves son ‘on ground surface ines of ight and level are ino jess cen int te sit w vu 2.9.2 Earthquake Magnitude ‘The possibility of obtaining a more objective, quantitative measure ofthe size of an earth- {quake came about with the development of modern instrumentation for measuring ground ‘motion during earthquakes. In the past 60 years, the development of seismic instruments, and our understanding ofthe quantities they measure, have increased dramatically. Seismic instruments allow an objective, quantitative measurement of earthquake size called earth quake magnitude tobe made. Most measurements of earthquake magnitude are instrumen- tal ({.., based on some measured characteristic of ground shaking).Sec.2.9 Size of Earthquakes a wm} fa fm fw! vy bv] viv} x} x | xt | xn Re fu} om fi} v |v fo) vn | ox ma 1 ufamfa] oy wl vu mx for fal om fav} vy |v | vir{vml ax x fxs | xn Figure 227 Comparison of nteasty vues fem modified Meal (AMID, Rossi Forel (RF), Japanese Meteorological Apecy UMA, and Medvedev Spronheaes Karnik (SK) seals. (Afr Rice (1958) and My and "Ben 37 a = | ) Figure 228, Isoxesmal maps fom the 1968 Inngaon canta a New Zealand (Ales Eby, 1980) an (bth 189 Loma Poa cru it thers Cahora (Modied Mec tenses) (Ae Hae 1590)a Selsmology and Earthquakes Chap. 2 2.9.2.1 Richter Local Magnitude In 1935, Charles Richter used a Wood~Anderson seismometer to define a magnitude scale for shallow, local (epiceatral distances less than about 600 km (375 miles) earth {quakes in southern California (Richter, 1935). Richter defined what is now known asthe local magnitude asthe logarithm (base 10) of the maximum trace amplitude (in micro meters) recorded on a Wood~Anderson seismometer located 100 km (62 miles) from the ‘epicenter of the earthquake. The Richter local magnitude (M)is the best known magnitude scale, but it is not always the most appropriate scale for description of earthquake size. 2.9.2.2 Surface Wave Magnitude ‘The Richter local magnitude does not distinguish between different types of waves, Other magnitude scales that base the magnitude on the amplitude of « particular wave have been developed. At large epicentral distances, body waves have usually been attenuated and scattered sufficiently that the resulting motion is dominated by surface waves. The surface ‘wave magnitude (Gutenberg and Richter, 1936) is a worldwide magnitude scale based on ‘the amplitude of Ray is obtained from tn waves with a period of about 20 sec.‘The surface wave magnitude M, = log +1.66 loz +20 ea) \where is the maximum ground displacement in micrometers and A is the epicentra ds tance ofthe seismometer measured in degrees (360° corresponding to the circumference of ‘the arth). Note that the surface wave magnitude is based on the maximum ground displace ‘ment amplitude (rather than the maximum trace amplitude of a particular seismograph); therefore, itcan he determined from any type of scismograph. The surface wave magnitude is most commonly used to describe the size of shallow (less than about 70 km (44 miles) focal depth), distant farther than about 1000 km (622 miles)) moderate to large earthquakes, 2.9.2.3 Body Wave Magnitude For deep-focus earthquakes, surface waves are often too smal to permit reliable eval- uation of the surface wave magnitude. The Dody wave magnitude (Gutenberg, 1945) is a Worldwide magnitude scale based on the amplitude ofthe first Few cycles of p-waves which ‘are not strongly influenced by the focal depth (Bolt, 1989), "The body wave magnitude can be expressed as my = logA—logT #001459 ey where isthe p-wave amplitude in micrometers and Tis the peti of the p-wave (usually bout one see), Body wave magnitude can also be estimated from the amplitude of one-se: ‘ond-period, bigher-mode Rayleigh waves (Nuttli, 1973); the resulting magnitude, my. ‘commonly used to describe intraplate earthquakes. 2.9.2.4 Other Instrumental Magnitude Scales Magnitude scales using different parts of the instrumental record have also been pro- ‘posed. The coda of an earthquake motion are the backscattered waves (Aki, 1969) that fol- low passage ofthe primary (unreflected) body and surface waves. Aki (1969), showing that ‘certain characteristics of the coda are independent ofthe travel path, developed a coda ‘magnitude, M,that could be obtained from those characteristics. The duration magnitude,Sec.2.9 Size of Earthquakes « 2.9.2.5 Moment Magnitude 4138 important realize that the previously described magnitude scales are empirical ciara ave and Richter local magnitudes saturate t magnitude of Goo as surface wave magnitude saturates To describe the size of very large earth. sround-shaking levels, and consequently secs nt saturate, would be desirable. The only magnitude sale thats an subject fo satu based onthe tent magnitude (Kanamori, 1977: Haaks and Kaname, 19? sine aay Hise on the seismic moment, which isa direct measure ofthe fcton ct produce rupture along the fault. The moment magnitude i given by = ABH 99 as here My is the seismic moment (Section 2.5.3) in dyne-cm. Jin elationship between the various magnitude seales can be seen in Figure 2.29, Saturation ofthe instrumental scales is indated by thei latening higher magnitude a er ee) Moment Magnitude trea 2 Seton vaio zine cl: noms magne ty sao aed) M, (sats wane magi) m shaper ay oe azn) mong rod iy wave manta) dig Metoologial Agency magiuds) (Aferin is50 Seismotogy and Earthquakes Chap, 2 values. AS an example ofthe effects of magnitude saturation, both the 1906 San Francisco and 1960 Chile earthquakes produced ground shaking that led to surface wave magnitudes ‘of 8.3, eventhough the sizes oftheir rupture surfaces, illustrated by the shaded areas in Fig. ‘re 2.30, were vastly different. The great disparity inenergy release was, however, reflected {in the moment magnitudes of the earthquakes: 7.9 for San Francisco and 9.5 for Chile ~ Chie, 1960 Figure 280° Comparison of relive areas of fuk pire shad) a maids or 1906 San Francisco and 1960 Chi earhqakes,Ahoagh he saking of Bath cases produce surface wave mapnitade of 3, the mount of eneryeeleased ete ery ifn selon tot moment magus (After Boe 197. The ‘men of ground during euauaks, cient American, Vol. 237. No.6, used sh pemsion Bolt (1989) suggests that M, oF m,be used for shallow earthquakes of magnitude 3 to 7.M, for magnitudes 5 to 7.5, and M, for magnitudes greater than 7.5. 2.9.3 Earthquake Energy ‘The total seismic energy released during an earthquake is often estimated from the rcation- ship (Gutenberg and Richter, 1956) log = 118+ 15M, 6) where E's expressed in ers. This relationship was later shown (Kanamori 1983) o be applicable to moment magnitude aswel. implies that unit change in magtude core sponds to a 10'* or 32-fold increase in seismic energy. A magnitude 5 earthquake therefore wold ease only about 0.01 times the enersy of a magitade earthquake, herby ills- trating the ineffectiveness of small earthquakes in eievng the buildup of stra energy that causes very large earthquakes. Combining equations 2.8) and (2.6) using M) shows that the amount of energy released during an earthquake is proportional tothe seine moment The amount of energy released by earthquakes is often difficult to comprehend: although single eg is smal (1 erg = 7.5% 10" feb), the energy’ released ina tome bomb ofthe size wed at Hiroshima (20,000-t0n TNT equivalnt) would comespond 4 magnitude 6.0 earthquake. On that ass, the 1960 Chile earthquake (M, = 9.5) released as much energy as 178,000 such atomic bombs (Figure 2.31).Motori impact (10m diameter, 20 ssc velo) Ears emu haat ow, U.S. Annual energy consign. ‘Average nuricane (10 day tine) steno Pai. ee ait awe aren Now Moc 101-12 Hrshina som bomb 2's ges Etectial encigy of 10" ‘plea thunderstorm ‘Average tomas 10 thine oer) Lining Bo 20 30 40 $0 60 70 80 G0 100110 120180 Equivalent Moment Magnitude (i) 51 50 60 70 80 90 190 Moment magnituse (i) igure 2.31 Relative energy of aus naar end hean-made phenomena. (ARE ino, 1990. Reprinted by persion of USGS) 2.10 SUMMARY 1. The earth has a layered structure—the surficial crust is te, the outer core, and the inner core, The temperature depth, The temperature gradient i the mantle causes slowly by eonveetion, ‘underlain in turn by the maa- of each layer increases with the semimoltea rock to move 2. The crusts broken into a numberof large plates and smaller platelets, Shear stresses ‘on the bottoms ofthe plates, caused by lateral movemé and gravitational forces cause the plates to move with 3. Relative movement ofthe plates causes stresses to bul ent ofthe convecting mantle, respect to cach other. id up on their boundaries, As ‘movement occurs, strain energy accumulates in the vicinity of the boundaries. This2 22 Seismology and Earthquakes Chap. 2 ‘energy is eventually dissipated: either smoothly and continuously or in a stck-slip manner that produces earthquakes. The size ofan earthquake depends on the amount of energy released. ‘There are three different types of plate boundries and their characteristics influence the amount of strain energy that can build up near them. As a result the different types ‘of boundaries have different earthquake capabilities: subduction zone boundaries can produce the largest earthquakes, followed by transform fault boundaries and then spreading ridge boundaries. The surfaces on which relative movements occur are called faults. At a particular locaton, a fault is assumed tobe planar with an orientation described by it strike nd «lip. Fault movement is divided into dip-slip components (normal and reverse fault- ing) and strike-slip components (left lateral and righ lateral faulting). The energy-releasing function of earthquakes suggests that a period of time for strain eneray accumulation should be expected between large earthquakes a the same loea- tion. Italso suggests that earthquakes should be mos Tikely to occur along portions of| fault for wich litle seismic activity has been observed —unless the plate movement thas occurred aseismically . Earthquake intensity isa qualitative measure ofthe effects of an earthquake at a par- tioular location, 1s related to the size ofthe earthquake but is also influenced by ‘other factors, Isoseismal maps can be use to describe the spatial variation of intensity fora given earthquake. Because no instrumental measurements are requited, his «al accounts can be used to estimate intensity values for preinstrumental earthquakes, Earthquake magnitude is a quantitative measure of the size of an earthquake. Most magnitude scales are based on measured ground motion characteristics, The local ‘magnitude is based on the trace amplitude of a particular seismometer, the surface ‘wave magnitude on the amplitude of Rayleigh waves, and the body wave magnitude ‘on the amplitude of p-waves, Because these amplitudes tend co reach limiting values, these magnitude scales may not accurately reflect the size of very large earthquakes, ‘The moment magnitude, which is not obtained from ground motion characteristics, is, able to describe the siz of any earthquake, . Earthquake magnitude seals are logarithmic, hence a unit change in magnitude cor responds toa 10-fold change in the magnitude parameter (ground motion character- istic or seismic moment). The energy released by an earthquake is related to ‘magnitude in such a way that a unit change in magnitude corresponds to 32-fold change in energy. HOMEWORK PROBLEMS (Convection caused by thermal gradients in tbe upper mantles thought to be primary ease of ‘ontinentl drift. Estimate the average thermal gradient inthe upper mantle ‘The coeicient of thermal expansion of the upper mantle fs about 25> 10K, Estimate the ratio of the density a the top ofthe upper mantle otha atthe botom onthe upper maneChap. 2 Homework Problems 53 2.3, Using the data from Figure 2.21, determine whether the San Ances and Calaveras faults are ‘undergoing right lateral or lef lateral strike slip faulting. 24 Using the data from Figure 221, estinaethe average rate of relative movement along the San ‘Andreas and Calaveras faults dering the period from 1959 to 1910, 25. Assuming p-and s-waves traveled through the erst at 6 knvscc and 3 kee, respective estimate the epicentral location (latitude and longitude) ofthe hypothetical earthquake whose characteristics ar given below: Seismogeh atiude Vonpinde ——_povave arial tine —_ swave ail ine Sra" REST" osanaNeS 6:112690 BraS00" PROF Dats gs -187 rns ease getttr26 O612353 26 Using 2 map of California, determine which fal the hypothetical eathquake of Problem 2.5 would mos likely have occured on? 27 Anearthquake causes an average of 2S m stke slip displacement over an 80 kt long, 23 km ‘dep portion ofa transform fault, Assuming that the rock along the fault had an average rupture ‘ength of 175 kPa, estimate the seismic moment and moment magnitude ofthe eatauake.Strong Ground Motion 3.1 INTRODUCTION ‘The earths farfrom quiet—it vibrates almost continuously at periods ranging from milliseconds {days and amplitudes ranging from nanometers to meters. The great majority ofthese vibra- tions are so weak that they cannot be felt or even detected without specialized measurement ‘equipment. Such microseismic activity sof greater importance to seismologists than engineers, [Eanhquake engineers are interested primarily in strong ground motion emotion of sufficient strength to affet people and their environment). Evaluation ofthe effects of earthquakes at 2 particular site requires objective, quantitative way’ of describing strong ground mation. The ground motions produced by earthquakes can be quite complicated, Ata given point, they can be completely described by three components of translation and thee com. Ponents of rotation. In practice, the rotational components are usually neglected; three orthozonal components of translational motion are most commonly measured. Typical _ground motion records, such as the aceeleration-ime histories shown in Figure 3.1, contain a tremendous amount of information. To express all of this information precisely (ie, to reproduce each of the tree time histories exactly), every twist and turn in each plot must be described. The motions shown in Figure 3.1, for example, were determined from 2000 acceleration values measured at time inerements of 0.02 sec. This large amount of informa- tion makes precise description ofa ground motion rather cumbersome.Sec.3.1 Introduction 55 goss oreytes 3 Ee ot i. | Lifton . 7 + ae 9 eee 7 ao o 2 vwonccon] 2 oreo Bo ao ba L,, a = ie a a wo ene a a 0 a «0 ‘Timo (66c) Tame (600) Figure 34 Accleration te histories rece att its in Gio. Cairn ring the 1989 Loma rea eathquake, Te Gitoy No. {lasrmcn a laced on nutcrop of Fesncsan sandstone wile he Gilroy No. inrument was ura by 365m (5400 of i alluvial ss, The Giloy No 1 Gok) and ly No. a ses wee fcated at eicental distncs of 2 km (138 miles) and 22h (Led ies. respectively Fortunately, itis not necessary to reproduce each time history exactly to describe the ground motion adequately for engineering purposes. It is necessary, however, to beable to describe the characteristics of the ground motion that are of engineering significance and t0 identity a number of ground motion parameters that reflect those characteristics. For engi- neering purposes, three characteristics of earthquake motion are of primary significance: (1) the amplitude, (2) frequency content, and (3) duration of the motion. A number of different ‘ground motion parameters have been proposed, each of which provides information about fone or more ofthese characteristics, In practice, its usually necessary t6 use mote than one ‘of these parameters to characterize a particular ground motion adequately, ‘This chapter describes the instruments and techniques used to measure strong ground ‘motion, and the procedures by which measured motions are conrected. It then presents v= ety of parameters that can be used to characterize the amplitude, frequency content, and dia tion of strong ground motions. Relationships that can be used 0 predict these paramcters are also presented. The chapter concludes with a brief description of the spatial variability of56 Strong Ground Motion chap. 3 round motions Before proceeding further, the reader should review the topics discussed in Appendices A and B—familiarity with the concepts presented in those appendices assumed in this chapter and the remainder af the book: 3.2 STRONG-MOTION MEASUREMENT ‘The identification and evaluation of ground motion parameters requires access to measurements ‘of strong ground mexions in actual earthquakes. Accurate, quantitative measurement of song. _ground motion is critical for both seismological and earthquake engineering applications. As stated by the National Research Council Committee on Earthquake Engineering Research (Housner.1982):“The recording of strong ground mexioa provides the basi data for earthquake engineering, Without a knowledge ofthe ground shaking generated by earthquakes, its not pos sible to assess hazards rationally orto develop appropriate methods of seismic design.” 3.2.1 Selsmographs Although written descriptions of earthquakes date back as far as 780 8.C., the first accurate measurements of destructive ground motions were made during the 1933 Loag Beach, Cal- ifornia earthquake (Hudson, 1984). Measutement of ground motion has advanced consid- erably since then, most rapidly inthe past 20 years or so. ‘Various instruments are available for ground motion measurement. Seismographs ae used to measure relatively weak ground motion; the records they produce are called seis- ‘mograms. Strong ground motions are usually measured by aecelerographs and expressedin the form of accelerograms. The simplest type of seismograph can be illustrated by a mass spring-dampersingle-degree-of-freedom (SDOF) system, as shown in Figure 3.2. A rotat ing drum is connected to the seismograph housing with a stylus attached to a mass, The ‘mas is connected tothe housing by a spring and dashpot arranged in parallel, and the hous ing is connected tothe ground. Since the spring and dashpot are not rigid, the motion ofthe ‘mass will not be identical to the motion ofthe ground during an earthquake, The relative ‘movement of the mass and the ground will he indicated by the mace made by the stylus on Housing ~~ a sy _e— Figure 32 Simple mas-sring-daspot ype of siaogrph The hosing i fly ones the ground. When he ground shakes, he sys aks ase o the ‘oi dru hat shows. she relate dplaseent twos the sand he gon “Mos moder instrumeat re more epithe one shown ereSec.3.2 Strong-Motion Measurement 7 the rotating drum. A typical seismograph station may have three seismographs orieated to ‘ecord motion inthe vertical and wo perpendicular horizontal directions ‘Seismographs can be designed to measure various ground motion characteristics, To tunderstand how this can be done it is necessary to consider the dynamic response ofa sim ple seismograph such asthe one shown in Figure 3.2. This seismograph isa SDOF oscillator ‘hose response to shaking is given by the equation of motion (Appendix B) mice ku = —mit, an Where «is the scismograph trace displacement (the relative displacement between the seis- ‘mograph and the ground) and u, isthe ground displacement. If the ground displacement i simple harmonic ata circular frequency a, the dis Placement response ratio (the rato of tace displacement amplitude to ground displacement amplitude) will be fw a2 Ml TO 2gpy whet B (= e/a) i the tuning ratio, (= 7m) is the undamped natural circular fre dene, and & (= €/2.fkm) isthe damping ratio, Figure 3a shows how the displacersent response ratio varies with frequency and damping, For ground motion Fequencics well above ‘he natal frequency ofthe seismograph ie, large values of, the trace amplitude sexe tothe ground motion amplitude, The lowest frequency for which this equality hols iin a given range of accuracy) depends onthe damping rato. Because the fequcicy mesons es fa and phase anles are preserved at damping ratios of 60%, SDOF dxplacement econo sraphs ar usally designed with damping ratios in that range (Richr et, 1070), Similarly, the acceleration response ratio he ratio trace displacement amplitude o ground acceleration amplitude) is piven by a 1 Md” ail? CeBE The variation of acceleration response ratio wit frequency and damping is shown in Figure 3.3b. The wace amplitude is proportional to the ground acceleration amplitade for frequencies wel below te natural frequency ofthe seismosraph (ie. low values of B) seismograph with 60% damping will accurately measure accelerations at fequencios vp bout 55% ofits natura requeney. Most seismographs ofthis type have natural requenc ios of about 25 Hz with damping ratios near 60%, with desirable fat response (constant accel, eration respons rato) t frequencies up to about 13 He The preceding paragraphs sow how the same physical sytem can et ts bth a dis placement seismograph and an accelerograph It measures displacement at frequencies ‘ell above and accelerations at fequencies well below ts natural frequency. The Wood Anderson seismograph, used by Richter to develop the frst earthquake magnitude cele ‘sed small mass suspended eccentrallyonathin tungsten torsion wire, Amro ateched to the wire allowed optical recording with aground motion magnification of about 3000, Damping was provided electromgnetically a 80% of eritial: the damped nara period ‘was about 08 sec. G3)58 Strong Ground Motion Chap. 3 x ° o gure 3 (a) Displacement esponse aio und aceleton response ta ‘alse or SDOF yom sujet to simple aro bus eon, ‘In most modem seismographs, an electronic transducer often referred to asa seismo- ‘mezer senses the motion and produces an analog (continuous) electrical signal that recorded for subsequent processing. Most accelerographs currently in use ate aecelerome- ters, electronic transducers that produce an output voltage proportional to acceleration, A number of different types of accelerometers are available. Servo (or force balance) accel- crometers use a suspended mass to which a displacement transducer is attached, When the housing is accelerated, the signal produced by the relative displacement between theSec.3.2 StrongMotlon Measurement 59 ‘housing and mass js used to generate a restoring force that pushes the mass back toward its ‘equilibrium position. The restoring force is proportional to the acceleration and can be mea. sured electronically. Servo accelerometers can provide very good accuracy over the range ‘of frequencies of greatest interest in earthquake engineering. Piezoelectric accelerometers use amass attached toa piezoelectric material (usually quartz, tourmaline, ora ferroelectric ceramic) to sense accelerations. The piezoelectric material acts as the spring in the diagram of Figure 3.2; damping is generally negligible. When accelerated, the inertial force ofthe ‘mass strains the piezoelectric material, which develops an electrical charge om its surfaces. ‘The resulting voltage is it the dielectric constant dacs not vary with charge) proportional 0 acceleration, Because piezoelectric materials are quite stiff, their natural frequencies are very high, so they are particularly useful for high-frequency measurements. Their response at ow frequencies, however, can be strongly influenced by signal conditioning system char- acteristics. Traxial accelerometers, in which three onhogonal components of acceleration ‘are measured with a common time base, are commonly used. From the three components, the acceleration in any direction can be computed. Some seismographs use velocity trans. ddacers, or geophones. in addition or as an alternative to accelerometers Seismographs. accelerographs, and ancillary equipment are protected by an instru iment shelter (Figure 3.4). An important component ofa ssismograph or accelerogeaph isan accurate clock, particularly when more than one component of motion is measured or when the ground motion at one location is being compared with that at another. Most modern instruments maintain time accuracy by synchronizing on a daily basis with radio time sig- nals transmitted by a standard time service or by recording such signals along with the _ground-motion data. Universal Coordinated Time (the scientific equivalent of Greenwich ‘Mean Time) is used as a common worldwide time basis. A seismoscope (Hudson, 1958) isa relatively inexpensive ground motion instrument. Scismoscopes are conical pendulums (Figure 35a) in which a metal stylusattached tus. pended mass inserbes a record of ground motion on a smoked glass plate, producing a two «imenstonal record ofthe type shown in Figure 3.5b, Scott (1972) found that small oscile lations ofthe trace were related tothe instrument rather than the earthquake and that they ‘ould be used to provide atime scale tothe seismoscope tre, The time scale allows accel- ‘erograms to be computed from the sismoscope trace 3.2.2 Data Acquisition and Digitization Early ground motion instruments transformed the motion of the ground tothe motion of a Physical mechanism. The mechanism, perhaps in the form of a pen or stylus or reflective Inirtor, caused the motion to be recorded in analog form on paper or photographie film attached to a rotating drum, Later-generation instruments recorded motions electronically in analog form on magnetic tape. Rather than record continuously, instruments of these types lay dormant until triggered by the exceedance of a small threshold acceleration at the beginning of the earthquake motion. Asa result, any vibrations that may have preceded tig- _Beting were not recorded, thereby introducing a baseline error ino the acceleration recor. To use the recorded ground motions for engineering computations, the analog ground ‘motion records must be digitized. Originally, digitization was performed manually with paper, Pencil, and an engincering scale. Semiautomatic digitizers, with which a user moved a lens with ‘crosshairs aerossan aecelerogram mounted on a digitizing table, were commonly used in the Late© Figure 34 (0) Mode digital stong maton instant (sear powered, 168 resolution, 250 samples, GPS timing nd celia oder) mounted in ance ‘enforced concrete an, and () completed insallation wth insted over and solar nel (Couey of Tera Teshloay Corp, Redmond, Washington 60Sec.3.2 Stong-Motion Measurement or Figure 35 (a) Typical Wilnottype sssmoscope in which ied stylus series a record of lative motion ons stoked pie pte mounted on te sae seismoscope: b pia sesoscoe ser (Aer Newmark tnd Roenblath, 1971) 1970s. By pressing a foot-operated switch, the coordinates of the crosshairs were recorded ‘These forms of digitizing involved exacting and tring work: operator accuracy an fatigue were ‘important considerations (Hudson, 1979). Fully automatic computer-based digitization, typic cally at sampling rates of 200 or more samples per second, is now commonplace In recent years, digital seismographs have become muich more commonly used. Although they use analog transducers, digital instruments convert the analog signal t digital form in the field. They record ground motions continuously at rates of 200 1008) sam- ples/see with 12-to 16-bit resolution, saving the recorded data only if tiggering acceleration is exceeded, Their on-board memories can typically save 4106 Mb wort of data, rom before an earthquake begins until afteritends, thereby preserving the inital portion othe record that is lost with triggered analog systems, Because digital systems are more complex, more expen- sive, and more difficult to maintsin in the fied, they have aot yet replaced analog systerns. 3.2.3 Strong-Motion Processing ‘The raw data obianed from a strong-motion instrument may include errors from several pos- sible sourees, each of which must be carefully evaluated and corrected to produce an accurate record of the actual ground motion. Raw data often include background noise irom different Sources. Microscisms from ocean waves canbe detected by sensitive instruments. Other noise ‘ay be caused by taffic, onstruction activity, wind (ranstited tothe ground by vibration of twee buildings etc), and even atmospheric pressure changes. Obviously, this range of sources «can produce nonseismic noise at both low and high frequencies. To isolate the motion actually produced by the earthquake, background noise must he removed or atleast suppresseda Strong Ground Motion Chap. 3 All accelerographis have their own dynamic response characteristics, or instrament ‘response, that can influence the motions they measure. Consequently. instrument response ‘must be corrected for in strong-motion processing. Instrument response corrections are ust ally performed by modeling the instrument iiself as a SDOF system and using the SDOF mode! to decouple the response ofthe instrument from the actual ground motion. For most ‘modem acceletographs with Na frequency response up to about 12 to 13 Hz, the instrument correction is only important for frequencies above the usual range of engineering interest. However. some aceelerographs are located in buildings (usually on the ground floor or inthe basement) or near the abutments of dams or bridges. The motions they record canbe affected al frequencies of interest by the response of the structure in or near which they are located, Even the motions recorded by strong motion instruments located inthe fre field (away fom, the influence of large structures) may be influenced by the response oftheir instrument shel- ter (Bycrof, 1978; Crouse etal, 1984), although these effects are usually important only st relatively high frequencies (Crouse and Hushmand, 1989) for typical instrument shelters, Another correction is required to reduce the effects of errors in ground motion mea surement, such as those associated with the triggering of analog seismographs. [Fa seismo- _raph does nt star until some triggering level of motion is reached, the entire accelerogram is in eror by the level of motion at the time of triggering. Integration of an uncorrected acceleration time history, for example, will produce a linear error in velocity and a quadratic cor in displacement. An acceleration error as small as 0.0012 a the beginning of a 30-see- Jong accelerogram would erroneously predict a permanent displacement of 441 em atthe tend of the motion, Correction of such errors, termed baseline correction, was originally accomplished by subtracting a best-fit parabola from the aecelerogram before integrating to velocity and displacement but is now performed using high-pass filters and modern data processing techniques (Joyner and Boore, 1988), The motions shown in Figure 3.1, for example, were bandpass filtered to remove frequencies below 0.08 Hz and frequencies above 23 Hz. Computer software for processing strong-ground-motion records (Converse, 1992) is available from the U.S. Geological Survey (USGS). 3.2.4 Strong-Motion Instrument Arrays Large earthquakes produce ground motions with different characteristics at diferent points fon the ground surface. ‘The spatial variation of ground motion, whether on worldwide, regional, o local scales, is important in both seismology and earthquake engineering Arrays and networks of Strong motion instruments have proven useful in determining the spatial variation of strong ground motion. 3.2.4.1 Worldwide and Regional Arrays Understanding of earthquake and tectonic processes improved dramatically withthe establishment of the Worldwide Standard Seismograph Nenwork (WWSSN) in 1961. The WWSSN was originally developed. in arge pat, to monitor compliance with nuclear Weap- ‘ons testing bans. Before that time, worldwide seismicity data were obtained from a wide variety of very different types of instruments operated by many different organizations. Dif- ferences in instruments and operating procedures made comparison of results difficult WWSSN stations use standardized instruments; each station has atleast two three-compo- ‘nent analog seismographs to monitor both short- and long-period motions. The capabilities of the WWSSN instruments, however, are limited by modetn standards (Aki and Richards,Sec.3.2 StrongMotion Measurement a 1980), and they ae being replaced by digital instruments such a those ofthe Global Digital ‘Seismometer Network (GDSN) and the Giobal Seismographic Network (GSN). The Incor- porated Research Institutions for Seismology (IRIS), a consortium of U.S. and foreign ‘esearch institutions, oversees operation of the GSN and a set of portable instruments that can be deployed to monitor aftershocks following large earthquakes. Regional arrays of seismographs are now operating in most seismically active coun- tie. In the United States, for example, the USGS operates regional arrays in different parts ‘of the country. In the California Strong Motion Instrumentation Programa (CSMIP) the Cal- ‘fornia Division of Mines and Geology operates an extensive acray of free-field seismo- graphs (Figure 3.6) as well as seismographs in buildings and bridges. Figure 36. Locuions of strong maton isrment pest ty te U.S, Geological ‘Surry in cooperation wit ter geaies 3 of Api, 190, Bons ther southern Califor indicate aes wit igh instrament desi, Af ayer und oor, (Gotcha News, Meh 1991 p28 Use by persion of Bech Publishes tid) 3.2.4.2 Local and Dense Arrays ‘While widely spaced regional and worldwide arrays are useful for studying earthquake ‘mechanisms and the spatial distribution of many important earthquake parameters, eotech- nical earthquake engineering often requires spatial distribution information on a smaller areal Scale and below the ground surface. In recent years 2 numberof local and dense arays, some ‘with downhole instrumentation, have been installed at various locations around the world. Japan has been very active in the installation of loalstrong-motion instrument arrays ‘The thtee-dimensional dense accelerometer aay at Chiba (Katayama and Sato, 1982), for example, includes 44 three-component aceelerometers, 15 of which are a the ground Sur- face and the remainder at depths of up to 40 m (130 f.).In Taiwan, the SMART-I dense accelerometer array near Lotung (Figute 37) consists ofa central accelerometer surrounded64 Strong Ground Motion Chap. 3 + Surace accelerometer ns30W 0 Brawn ssn Figure 37. Orginal confguion of Figure 38 The El Cet Arty in suthern California SMART aray in Lotng Tuan, “Tne EA Co Ditecenl Ay Is std ner Staton 9 by three rings of 12 accelerometers each at rai of 200 m (650 ft), 1 km (0.6 mi), and km (1.2 mi). 4 few years after the SMART-1 array was installed, | and +, -scale models of ‘nuclear containment structures were constructed in its midst (Tang, 1987). The structures were instrumented to record their response during earthquakes, and additional surface and «downhole (to depths of 47m (154 fh) ground motion instruments were installed adjacent to the !-scale model and in the free-field, In the United States, one ofthe most important local arrays has been the El Centro Array, a 45-kin-long (28 mi) array of 13 stations that crosses the Imperial and Brawley faults in southern California (Figure 3.8), Ialso contains the El Centr Differential Array, 1a dense array consisting of six dhe-component accelerometers along a 305-m (1000 ft) Tine. Shortly after installation, the arrays recorded the 1979 Imperial Valley earthquake (M, = 6), whieh occurted only 5.6 km away and produced very useful information on ‘near-field ground motions. Near Anza, California, an array often three-component stations along a 30-km (19 mi) stretch of the San Jacinto fault was installed to study various eath- ‘quake characteristics (Berger eta, 1984). Data are telemetered by digital VHF radio oa ‘nearby mountain peak station and then on to another station in La Jolla, California, ‘These are but a few of the many strong-motion arrays that have been installed in seis ‘ically active countries around the world. The rapid proliferation of local, regional, and worldwide seisimograph arrays in recent years has come hand in hand with technological ‘advances in data acquisition, storage, and communication, The ability to acquire and stone large quantities of digital seismic data at high speeds, and to retieve the data from remote locations by telemetry, has and will continue to make such data more plentiful 3.2.5 Strong-Motion Records ‘Strong-motion records can now be easily obtained from a nomber of sources. The U.S. Geolog- ical Survey fr example, published a compact disk (Seekins eta, 1992) that contained uncor- ‘ected strong motion records from North American earthquakes between 1933 and 1986; more than 4000 records were include. A numberof strong-motion databases can be accessed overthe Internet, with individual records downloaded by anonymous fip (file wansfer protocol)Sec.3.3 Ground Motion Parameters 65 procedures. The Gilroy records shown in Figure 3.1 were blaine from the database maintained by the Lamont-Doherty Earth Observatory at Columbia University in conjunction withthe "National Cente for Earthquake Engincering Research at SUNY Buffalo. An example ofthe information provided with such records is shown in Figure 39, A useful Word Wide Web site \with links to many sources of ground motion and earthquake information is maintained a the University of Washington (up. /tvwv.geophys.washington.eduseismosurfing him). vatiety of geotechnical earthquake engineering information can be found ata web site maintained atthe University of Southem California (hap:/ rece0.usc.edweqdatafhome him, TNCEER ASCIT STROWG-WOFTON DATA FORCE year=1989 monthe10 day=18 hour|minote (24hr)=0004 second 2,200 tine codesuTe LOCATION: Latitudes 37.03700 longitudes -121.80300 depth (Xm NAME: SANTA CROZ MOMS. (LOMA PRIRTA) EARTHQUAKE LOCATION: IatStuden 36.9730 longitudes -121,57200 elevation (m)= 0.0 SEE xD: 47378 - GrLROY #1 - GAvrzan cotzEon, WATER TAN START DATE: year=1989 month=10 day-18 START TIE: our|mimite (24hr)=0004 eeconde23.900 time code-uTC EFICENTRAL DISTANCE: distance (Im)= 21.8 azimuth (dog)=8209-1 SPECS: sampling rate (se0)=0.020 unktencu/szc/a5c CCourome® OF MOTION: az{mutheNO (HORIZONTAL) -18.68 3.95 5.46 38 981-048 3.67 740 sso -15.90 278 wela3 82 -20.93 15.56 0.33 662.16 33.36 67.5¢ 3.26 -32.89 Figure 39° ent sie and recon nfomaton proce the dgtizedwceceration afr the Gilroy No (oe) stang-mation eon 3.3 GROUND MOTION PARAMETERS Ground motion parameters are essential for describing the important characteristics of song ground motion in compact, quantitative form, Many parameters have been proposed66 Strong Ground Motion Chap. 3 to characterize the amplitude, frequency content, and duration of strong ground motions; some describe only one of these characteristics, while others may reflect two or thee Because ofthe complexity of earthquake ground motions, identification of a single param. ter that accurately describes all important ground motion characteristics is regarded as ‘impossible Gennings, 1985; Joyner and Boore, 1988). 3.3.1 Amplitude Parameters ‘The most common way of describing a ground motion is with atime history. The motion parameter may be acceleration, velocity, or displacement, or all three may be displayed as show in Figure 3.10. Typically. only one of these quantities is measured directly withthe ‘others computed from it by integration and/or differentiation, Note the different predomi= ‘nant frequencies in the acceleration, velocity, and displacement time histories. The acceler- ation time history shows a significant proportion of relatively high frequencies. Integration produces a smoothing or filtering effect [in the frequency domain, ¥(«9) = (oo) and to Nos East-West Coy No.2 East-west oe 35 Bo - 0 tit i oe wo Tne 86) 7° : Poly i Bo iyi dE Lenin Fm 3 26 5 a arr ea Tine 8) Tee) = & & Q ; A 0 a 0 0 Ga 0 Tine ee) Tine ee) gure 210 Accelero, velocity, and displacement me hist for tbe EW irpnents ofthe Gi No.1 oe) and Gy No, 2 (ai sttong motion ears The velo and dpacement mere obtained by inte ping the ascelerton records showin Figue 3.1 usin the apse ale Note tht he Coy No (ack te ‘xpetenced higher aeration, bathe Giro No, 2 (ste expenenoa ihe ‘leis and displacements,Sec.3.3 Ground Motion Parameters o (@) = 1(€)/, where i, 7, and & are the transformed displacement, velocity, and aceel- y ‘The quantity 27ny/dyu) fs equal othe predominant period of a imple harmonic maton, To See how wellitcoresponds tothe predominant period ofthe Giroy No. I (rock) and Gio) ‘No.2 (sol) ground metions, Gilroy No.t (rock): aR! = 0.49 see 03950 Gilroy No2 (oil: 2, sec Ty = O53 Though the ratio vnu/@aay certinly indicates that the Gilroy No.1 (rock) motion has a higher feequeney content than the ily No, 2 soi) motion, it overestimates the predominant pried ‘of both the Gilroy No, | (roe) motion and Gilroy No.2 (sol) motions, Due othe approximate ‘ature ofthe predominant period and the stochastic nature ofboth tng ai dye close agree. ‘ment between Yau/ena nd predominant period should not be expected 3.3.3 Duration ‘The duration of strong ground motion can have a strong influence on earthquake damage, Many physical processes, such asthe degradation of stiffness and srength of eertin types of structures and the buildup of porewater pressures in loose, saturated sands, ae sensitive to the number of load or stress reversals that occur during an earthquake. A motion of short duration may not produce enough load reversals for damaging response to build upin a ste ture, even ifthe amplitude of the motion is high. On the other hand, a motion with moderate ‘amplitude but long duration can produce enough load reversals to cause substantial damage,20 ‘trong Ground Motion Chap, 3 The duration of a strong ground motion is related 10 the time requi accumulated strain energy by rupture along the fault, As the length, or area, of fault rupture ‘ncreases the time required for rupture increases. As.a result, the duration of strong motion increases with increasing earthquake magnitude, While this relationship has ben supported bby empirical evidence for many years, advances in source mechanism modeling (Hanks and MeGuire, 1981) have provided theoretical support, indicating that the duration should be proportional to the cube root ofthe seismic moment. When bilateral rupture [i.e rupture that propagates in opposite directions from the focus (as inthe ease ofthe 1989 Loma Pricta caithquake)] occurs, the strong motion duration may be considerably lower, An earthquake accelerogram generally contains all accelerations from the time the certhquake begins until the time the motion has returned tothe level of background noise For engineering purposes, only te strong-motion portion ofthe accelerogram is of imterest. Different approaches have been taken to the problem of evaluating the duration of strong ‘motion in an acvelerogram. The bracketed duration (Bolt, 1969) is defined as the time ‘between the firs and last exceedances of «threshold acceleration (usually 0.0Se). Another {efinition of duration (rifunac and Brady, 1975p) is based on the time interval between the Points at which 5% and 95% of the total energy’has been recorded, Boore (1983) has taken the duration to be equal tothe corner period (ie. the inverse of the corner frequency). The rate of change of cumulative root-mean-square (rms) acceleration has also been used asthe basis for evaluation of strong-motion duration (McCann and Shah, 1979). Power spectral density concepts can aso be used to define a strong-motion duration (Vanmarcke and Lai, 1977), Other definitions of strong-motion duration have been proposed (Perez, 1974; Ti funac and Westermo, 1977). Because it implicitly reflects the strength of shaking, che bracketed duration is most commonly used for earthquake engineering purposes The duration of strong motion has been investigated by interpretation of accelero- grams from earthquakes of different magnitudes. Using a 0.05¢ threshold acceleration, (Chang and Krinitszky (1977) estimated the bracketed durations for soil and rock sites at short (less than 10 km) epicental distances shown in Table 3-2 Duration has also been expressed in terms of equivalent cyeles of ground motion, One such approach was developed in conjunction with an early procedure for evaluation of tig uefaction potential (Seed etal, 1975), To represent an iregular time history of shear stress ‘Table 32. Typical Earthquake Durations at Epicentral Distances Less Than 10 km eration 5) Magnitde Rock Sis Soi Siey 50 4 : 55 6 2 60 8 6 6s u a 70 Io 2 75 2 45 Soure Chang and Keinizahy (197),Sec. 3.3. Ground Motion Parameters ar by a uniforin series of harmonie stress cycles, the concept of an equivalent number of sige nificant stress cycles was developed. The equivalent number of uniform stress cycles, shown in Table 3-3, was selected to cause pore pressure buildup equivalent to that of an ‘etal shear stress-time history ata harmonic stress amplitude of 65% of the maximum actual shear tess. Table 3.3. Equivalent Number of Uniform Stress Cycios arta Number of Significant Magna ‘Sues Cyeles st 2 6 5 6 0 1 Is q i Example 35 Determine the bracketed durations ofthe E-W components of the Gilkoy No.1 (rok) and Gile ry No, 2 (soil) ground mations. Solution Based on threshold uceleration of 0.05, the bracketed durations can be obtained graphically from the accelerograms shown in Figure E33 os eo ‘Accoeration (g) os ° - 098 a = © = 4 05 s Aeceleraton (a) b ° a a a a) Tino (sec) Figure 3SGikoy Not (rock): Gilroy No.2 (iy 3.3.4 Other Ground Motion Parameters ‘The preceding parameters related primaily tthe amplitude, requency content or duration of round motion. Since al ofthese characteristics ae important, proud melon pone eters that reflest more than oe are very useful. The fellowing paragraph resent eter ‘f parameters tha reflect wo oF ce important ground motion chareteroten ‘single parameter that includes the effects th amplitae an esuency content of a stong motion records the ms acceleration define as [Te ims = fp flora fi where Tis the duration ofthe song motion and 2y isthe average intensity (or mean- sare acceleration), Because the integral in equation (16) is ot stony lence large. high oquency accelerations (which occu only over avery shot prea tine oe because ts infueneed by the duration ofthe motion, lhe rms aeceeration canbe very ee fal for engineering purposes. Is vale, however, canbe sensitive tothe method ued a 74Sec. 34 Estimation of Ground Motion Parameters 89 where Ris the closest distance (5 60 km) to seismic rupture in kilometers (with minimum values of 7.3, 5.8, 3.5, and 3.0 km for magnitudes of $.0,5.5, 6.0, and 6.5, respectively); the source term, F, takes on values of 0 for strike-slip and normal faulting, and | for reverse, roverse-oblique, and thrust faulting; sy = | for soft-rock sites (sedimentary deposits of Ter. tary age), Sux = 1 for hard-rock sites (primarily older sedimentary deposits, metamorphic rock, and crystalline rock), and Sq = Sig = 0 for allavium sites. The 1994 relationship, hich is based on more data is clearly more specific (and more complicated) than the 1981 ‘lationship. The incorporation of additional terms reflecting source and site characteristics ‘are typical ofthe refinement of predictive relationships that has taken place in recent years. Boore etal. (1993) used data from western North America earthquakes of magnitude 5.0107.7 at distances within 100 km (62 mi) ofthe surface projection ofthe fault to develop the predictive relationship log PHA(2) = by +b2(My~6) +bs(My 6)? + bgR + Ds lOBR + blip + byGe (3.26) where R = {d+ 7°, dis the closest distance tothe surface projection ofthe fault in kilo- meters, and 0 forsite class A Jo for site cass A 1 forsitectissB —Ga= {0 for site class B 0 for site cass © 1 Forsiteclass © Note that che Boore et al. (1993) attenuation relationship is expressed in terms of the com> ‘mon (base 10) logarithm rather than the natural logarithm, The site classes are defined on the basis of the average shear wave velocity in the upper 30 m (100 ft) (Table 3-5), Coef- ficients forthe Boore ct al. (1993) attenuation relationship were developed for two mea- sures of peak acceleration: the randomly oriented component and the larger horizontal ‘component (the former considers two orthogonal horizontal records at a particular site as ‘separate events and the latter considers only the larger ofthe two). The coefficients are given in Table 3-6. Table 25. Definitions of Site Classes for Bore eta. 1983) Atenustion olationship Site Cass Fy in Upper 40m (1008) ‘A §150 mec (2500 ne) B 340-750 mse (1200-2800 ses) 180-360 mee 00-1200 3 ‘Table 346 Cootcionts for Boore et al (1999) Attonuation Relationship Component ee ee ee Random 005 0229 00 00-078 162 0251 557 0280 Lamer__-0088 0216 00 09-0777 158 0284 54820890 Strong Ground Motion Chap. 3 Since the continental crust in eastern North America is stronger and more intact than the crust in western North America, peak accelerations tend to be higher. For the mid- continental portion of eastern Noxth America, Toro etal, (1994) developed an attenuation relationship for peak horizontal rock acceleration: In PHA(g) = 2.20+081(M, 6) ~ 1271, k, i where Ry = «/R*+ 9.3", Ris the closest horizontal distance tothe earthquake rupeure in km), 6 = 0.36 + 0.07(iMy ~ 6), and — om ost for R 20 km ‘Subduction zone earthquakes generally occur at greater hypocentral depths than earth ‘quakes that ccuron transform faults. Consequently the seismic waves thatemanate from sub- ‘duction zone earthquakes follow different paths from those of transform faults. Youngs et (1988) used strong-motion measurements obtained on rock from 60 earthquakes and numeri cal simulations of Mf,» 8 earthquakes to develop a subduction zone attenuation relationship In PHA(g) = 19.16 + L085, — 4.738 In [+ 205 Sexp(0.0968M,)} $0542, Gra = 155-0.125M, here Ris the closest distance tothe zone of rupture in kilometers and Z, is 0 for interface events and I for intraslab events ‘The four preceding attenuation relationships are shown graphically for earthquake ‘magnitudes 55, 6.5, nd 7.5 in Figure 3.22, The shapes ofthe attenuation relationships are similar, despite the fact that they represent different geographic regions and different source ‘mechanisms and use different measures of distance. Peak Velocity. Rezression analysis of PHV data has provided a number of ‘useful relationships describing the attenuation of that parameter. Joyner and Bore (1988), {or example, used strong-motion records from earthquakes of moment magnitude between 5.0 and 7.7 o develop the attenuation relationship r + J(M—6) + (M6)? + jylogR + j,R+ jg B29) where PHV can be selected as the randomly oriented or larger horizontal component, R = .Jr3+} and ris the shortest distance (in kilometers) from the site to the Vertical projection of the earthquake Faull rupture on the surface ofthe earth. The coetficiens forthe Joyner and Boore (1988) attenuation relationship ae given in Table 3-7 (328) PHY(emisec) Table3.7_Coetfecients for Joyner and Boore (1988) Feak Horizontal Velocity ‘Attenuation lationship, Componeat Boh ok hk A tw Random 200 Lower 207 oo 10 oom 0740s 0-10 oom 017 40stSee. 34 Estimation of Ground Motion Parameters a to Campbeland ozone) | __—Saoata 89 © ° fre #os i 00 =! 00 1 to ion 000s 10100 1000 Distance ton Distance (en Too ot a. (1994 eres 20 yc oa 80) to Yours ta (998) © o s 5 m woes 00 oo 1 o> wm "Ff % mimo Datanee a) Distance em) Figure 322 Vaition of ek barron aleraton with disace fo M = 3.5, 65, and M75 carhguakes according wo various aemston ries: (a) Campbell and Borers (1988), soft rack tes and sre fing fb) Bowe eal (193) te tas 6) Tool (1994) and) Yours al (1988) nasa even. 4 Estimation of Frequency Content Parameters Large earthquakes produce larger and longer-period ground motions than do smaller earthe quakes; consequently the frequency content of a ground motion i related tothe earthquake ‘magnitude. As seismic waves travel away froma faul, their higher frequency components are scattered und absorbed more rapidly than are their lower-frequeney components. As & result, the frequency content also changes with distance. 3.4.4.1 Predominant Period (One aspect ofthe change in frequeney content with distance involves the shifting of the peak ofthe Fourier amplitude spectrum to lower frequencies (or hizher periods). As ‘result, the predominant period increases with increasing distance, as illustrated in Figure 3.23.Predominant pero (eee) 0 40 80 120 60 200 240 280 00360 Distance trom causatve tat - km Figure 323 Variation fpedoainan period at ok alcops with magitde and dss (After See et, 1969) Example 37 Determine the predominant period that would have been expected forthe reconded motion at the Gikoy No, 1 (cock site roy No. I (cock) motion was recorded at site located a an epicental dis: tance of 21.8 km from the Loma Prieta (M = 7.1) eataguake, From Figure 3.23, the expected ‘redominant period would be 0.33 sec. AS determined in Example 3.3, the actual predominant Derod of the Gilroy No, 1 (rock) motion was 0.39 sec 3. Ordinates of the Fourier amplitude spectrum can be estimated empirically by regres- sion on the Fourier spectral ordinates of actual strong-motion data (e.g., Trfunae, 1976: McGuire etal, 1984: Trifunac and Lee, 1987; Castro etal, 1990). Alternatively, physically based mode! of source, travel path, and site behavior may be ealibrated to prediet Fourier amplitude spectra, Based on Brunc’s (1970, 1971) solution for instantaneous slip of a circular rupture surface, the Fourier amplitudes for a far-field event at distance R can be expressed (McGuire and Hanks, 1980; Boore, 1983) as £ jeans "GS Si Fou) ® were fis the comer frequency (see Figure 3.14) fy the cutoff frequency (Figure 3.14), (Q(/) isthe frequency: dependent quality factor (inversely proportional to the damping ratio of the rock; see Section 5.5.1), and C isa constant given by pl = [ews 830) Saft ean SapoSec. 3.4 Estimation of Ground Motion Parameters 93 where Ry (0.55) accounts forthe radiation pattern, F (=2) accounts forthe free-surface effect, V(= «/2/2) accounts for paitioning the energy into two horizontal components, isthe density ofthe rock along the rupture surface, and v is the shear wave velocity of the rock. fai assumed constant fora given geographic region (15 Hz.and 40 Hz are typical values for western and eastern North America, respectively), the spectra for different eat- ‘quakes are functions of the seismic moment, M,, and f, which ean be related (Brune, 1970, 1971) by f 49% 10% $2) x) where v is in km/sec, M, isin dyne-cm, and Ac is referred to asthe stress parameter or stress drop in bars. Stress parameters of SO bars and 100 bars are commonly used for Sources in western and castem North America, respectively. Figure 3.24a shows how the Fourier amplitude spectra predicted by equation (3.30) vary with magnitude. Note the ‘strong influence of magnitude on both the amplitude and frequency content ofthe motion. As the magnitude increases, the bandwidth increases and the corner frequency decreases, {implying that more low-frequeney (long-period) motion will occur. Figure 3.24b shows time histories of acceleration generated from the spectra of equation (3.30) for magnitude ‘4and magnitude 7 events.‘The stress parameter and seismic moment are commonly’ used 10 specify the source spectrum, given by the expression in brackets in equation (3.30). The final expression is the rravel path operator, which describes attenuation of the Fourier ‘amplitudes as the energy travels away from the site. An expression that describes the effects of soil response (a site operator) can be added to equation (3.31), if necessary, 10 account for near-surface effects. The response of soil deposits during earthquakes is dis- ceussed in detail jn Chapter 7, ‘Since it s based on the mechanics of source rupture and wave propagation, equation (3.30) offers significant advantages over purely empirical methods for magnitudes and dis- tances for which few or no data are available. Meao M=70 , 089 og | Frenne @ cy) Figure 324 () Vastion of Fourier ampli pects at & = 10 kn for diferent moment mages (0 = 100 bar) agslerograms gener fromthe M wt and M=7 spect (Aix Boor, 1983 Used by permission ofthe Seismolgsal Society of Ameren)94 Strong Ground Motion chap. 3 3.4.4.3 Ratlo Vmax/Amax ‘AS ameasure of the frequency content ofa ground motion, the ratio Vallgy is also ‘elated to earthquake magnitude and distance. This dependence has been studied by several investigators, with a summary of their results provided by MeGuire (1978). who proposed the magnitude and distance dependencies shown in Table 3-8. The table indicates that expected, the vinn/dnas ratio increases with increasing earthquake magnitude and increas- ing source-tosite distance. Table 38 Magnitude and Distance Dependence of Viw/tgne” Conon, Migntde Depenence stn Dependene Rock sites aa ee Sailsites one ee ‘eure: After MeGue (197 The to Sga/dgg Proportion totes depenence eeloshis Example 38 Estimate the values of vl that would e observed at rock and sil ies 40 km from the soure of aM = 6 earthquake located near the 1989 Loma Prieta eathyuake, Solution Using the values from the Gilroy No. | (rock) and Gilroy No. 2 (sil sites, and recalling that thse sites were located 21.8 km and 22.8 ki, respectively. from the Af = rma cole Rock site: $a = 9.078500 40” 3.4.4.4 Response Spectrum Ordinates ‘The importance of response spectra in earthquake engineering has led to the devel= ‘opment of methods for predicting them directly. For many years the shapes of all response spectra were fora given class of soil conditions, assumed to be identical. Design spectra ‘were developed by scaling standard spectral shapes by some ground motion parameter, ust ally the PHA, As more recorded motions became availabe, the magnitude dependence of spectral shapes was recognized increasingly. For example, Figure 3.25 shows the response spectra computed from the accelerograms shown in Figure 3.21, The difference in spectral shapes at different magnitudes, particularly in the long-period range, are apparent. This shape dependence was later accounted fo, at least approximately, by using PHA, PHY. and peak displacement to scale design spectra indifferent frequency ranges (Newmark and Hall, 1978, 1982), as discussed in Section 8.3.2. More recently, regression analyses have been 'used to develop predictive relationships for spectral ordinates at various oscillator periods (©. Joyner and Boote, 1982, 1988; Crouse, 1991; Boote etal, 1993), For example, Boore etal. (1993) determined values of coefficients that when used ith equation (3.26), predict pseudospectral velocities for oscillators of different natural periods, These attenuation relationships, the coefficients of which are presented inSec. 3.4 Estimation of Ground Motion Parameters 95 1000 “ey, PSV (emisee) Figure 328 Rospense specea computed som the aecerograns of Figure 320. Epicental distances rm cach aceleroram were ary equal Nte that the shapes all athe ales, ofthe spect vary with eathquake magia (ter Anderson, 191. Geotecnial News, Yo. 9, No. p35. As by prison of Pero (6) Bifech Publisher Lid) ‘Tables 3-9 and 3-10, ae applicable tothe same conditions as the attenuation relationship of| equation (3.26), These coefficients produce smooth response spectra, as illustrated in Figure 3.26. Although the smooth response spectra ae limited to periods of 2.0 see, theit general shapes are similar to those of actual spectra 3.4.5 Estimation of Duration ‘The duration of strong ground motion increases with inereasing earthquake magnitude, However, the mannerin which strong motion duration varies with distance depends on how itis defined. Since acceleration amplitudes decrease with distance, durations based on abso- lute acceleration levels, such as the bracketed duration, would be expected to decrease with distance; at some distance all accelerations will drop below the threshold aeeeleration and the bracketed duration willbe zero. Page etal, (1972) and Chang and Krinitasky (1977) eon- firmed this, as shown in Figure 3.27, Durations based on relative acceleration levels (€.2 ‘Trifunac and Brady, 1975b; Dobry etal, 1978) increase with increasing distance and may have long durations even when the acceleration amplitudes are very low. For engineering purposes the bracketed duration appears to provide the most reasonable indication of the influence of duration on potential damage, 3.4.6 Estimation of Other Parameters Parameters that reflect more than oh important characteristic of strong ground motion are likely to see increasing use. For most, however, only limited data forthe development of predictive relaionships are currently available,96 Strong Ground Motion Chap. 3 Table39 Smoothed Coetfocionts of Predictive Relationships forthe Larger Horizontal Component of 8 Damped PSV Toh hob koe t 10 1300 o321 “0080-090 0m) aa Ie 1 17 O39 0110 oo “9929 00S ISDST la O12 7 0430 0119 eo “ages? Oe 2D ee a ora 1938 032 ono “993903008709 alen O1S 195 0323 a7 ap a9 0a nat7 7138 O16 1982 os 017 00-09 1402 TIA Bloe DIT 202 0306 011700 ag 01) 022 TIO ISS Die 20190328 01S “a6 o> 9254 THs 98 019 20 03309 “oe on age OT aaeh 688 ow om 200 038-012 0d og OS wa 6g 022 205 0.6 01 a “09280180281 BS 024 dost oat “ons a0 026 2087 038 “oil 80 “neta 028 2066 039 “05600 “ats 930 200381 “oo 00 “ase 32 20st 0358 “ooss 0 oaoT 038 20 0362 “90800 “aa 046 2085 366 “a9 “BN O38 2m8 300 “ome an “asl 4020 0377 “oom 80 os o42 2001 0377 “oot 00 “pat 048 2033 om “006s 00 “nant 046 20 0383 as! 08 “aes 4K 19960386 © “0058 OD 08S 030 198 0350 “oss 80 OSS 035 1568 0397 “oom 00 “aa48 60 199 40s “oon 00 “aga 6 19m a310 “ag? 09 “agg 070 1917 a8 0330) “aS 078 19m) ps2 “ammo “0 om 1891 0427 ons 00 “gaz 04am 289 024 as iaBl aaa “nor 80 “a6 o4si 2850397 on Is2 a4% “oo an “asa 04328) 1240 9s 1i6t 9.40 018 ad -0Ras os 281 000 10) 188 pg “0168 00 “04s M0188 0482 “oo 00 “oa la 18M 0488 “001300 “aan 130 182 o46i “ald 99 os7 140 14s 046 “0013-00 “as 150 19 0am “ors 00 Coast let 1857 oar “01600 “859 Lm 1886 0482 “aais oa "asa 05486 adr ta 1s 048s “9002 noe 0387529 naw 19 1991 nase “00050 “se Ossi 574 oags ass 6210387 on 19050491 a2 00 “oe we Alter Boor eal C93),Sec. 34 Estimation of Ground Motion Parameters ” Table 3-10 Smoothed Coetficiens of Proditve Relationships forthe Ranciom Horizontal Component of 8% Damped PSV ee 01016530327 “a0 00-09 aoe Os az 02mm O11 1725 0318 -o100 00-0937 oom Ose gas Ome 012 1782 0312-0101 00-0939 003 ots gor Oat 013 1828 307 010109 “939 oN OII 708 0209 BIg 1866 0307-0100 00 “93827 me 718 0am bss 1492 0305-0099 a9 ass? OM) oT 723 att 06 1915 0305-00800 “935 SS 724 2H O17 1933 OS 10500-0935} ON 721 at 918 1948 0306-0084 00 “9s ous 0288 T6023 019 1939 0308-002 00 “9270269 710s 020 1967 0309-00900 “sz 17027972 022 1978 0313-008 00 “asts 02080276 ORS 024 1920318 “nom oo “9912 O24 Osea? 020) 026 1982033 “no! on “9s nak «G22 02 197 om 007 op “ow nas OBIT 235 030 194 056 00 00 “aps. 020s SOF 206 032 1967 OMO 0006 00 “ass 0248367 S12 O28 D3) 1959 OMS “0082 00 “ask? 02s! «5500280 0361950 03580-0057 09 “oT 03560387 5300232 0381940 0356-0085 on 72260 0396 S10 0238 040 19303610082 oo “oaeT 026! 40s at O28 42 192 0365-008 00 “a2 267 aI 4TH 28 044 1310 037 007 00 “ass ozT! 0204570288 54 1900 0378-004 00 “oass_ 027} aT dal OMT 048 1390 037 © 00200 asso 0280433 2H O28 050 18st 0388-0039 © 00 “nis 027909413 2as 03518570384 too “os 028s 0s Oa o@ 1835 0403-0050 00 oss) 0289 OAs 357 O25t 06s sls oll 0008 © 0-423 oa OAM 3360284 07 17970418 002300 “Osis 0383200357 07s 1781 0425 “00% 00 “gai 0350 3070259 08 1768 0431 “ook ao “a0 Der 298 0361 8 1758 ats? ante 90 “oss 0503 292026 om 132 ase © “aois 00 002 0951732 mats 004-800 100 1m 450 “oo a0 08 10 1710 0487 “0013 an 0785 120 17010462 “001800 a88 130 1686 046 = “aos a9 “a08 ase 3301280 1401685 0469 “01700 “0704 O83 3620283 1301686 047-0019 00-79 ORR 0837— 382 O2RS 160 1700 0472-00200 “a7 0320538 4260386 130 1705 U7) OMS 00 axa 05394620289 180 1715 0472-07900 “oad 3510539 S01 0200 19 1738 osm ams: op ome 0356 0538 $42 a2 200 asi oat! -0087 00-082 asad 0537 Sas 0508289 0266, 0513288 0268 0317 2% om os; 29 aa asm he 0377 Source: Alter Boor eta (193),S,temsee} 98 Strong Ground Motion Chap. 3 100 [ oa} o7 06 os oa 03. | we L__ — 01 + 10 Petod (soe) Prod (20) Figure 3126 Response pec fo theron component of grou motion tate clas 8 ite at km sean athe redieteelaomip of Bore ta 15) a reuse vebelty wn ety fom eation (327 and Tae 310, an) gscdoptalacrcrationcompued from the psewdpcta vel 3.4.6.1 RMS Acceleration Hanks and McGuire (1981) used a database of California earthquakes of local mag- nitude 4.010 7.010 develop an altenuation relationship for ms acceleration for hypocentral distances between 10 and 100 km (6,2 and 62 mi: Gigs = O11. 33) ® where fis the comer frequency, fg isthe eutolT frequency, and R isin kilometers, Kavacanjian eta. (1985) used the definition of duration proposed by’ Vanmarcke and Lai (1980) with a database of 83 strong motion records from I8 different earthquakes to obiain fom = 0472 + 0.268M, +0.129 logl 9908 + 2235 )_o.1167R 34) eR)Bracketed curation (00) Bracketes duration (ec) gag a8 » eC, 3.4 Estimation of Ground Motion Parameters 99 3 0 10 20 20 40 50 60 70 80 90 100110 Epicentalestance (km) Figure 327 Vaiton of sted aration (008g teen) with agit and epcenalestace: (9) rock site; () o 9 10° 20 80 40 50 60 70 8) 90 100710 120 soll ses (Afer Chang ai Kenta. Epcentral listance (km) my where R is the distance to the closest point of rupture on the fault. The database was restricted to M,> 5, R < 110 km (68 mi), rupture depths less than 30 km (19 mi), and soil thicknesses greater than 10 m (33 f. 3.4.6.2 Arias Intensity Campbell and Duke (1974) used data from California earthquakes to predict the vai- ation of Arias intensity within 15 to 110 km (9 t0 68 mi) of magnitude 4.5 to 8.5 events eM0300, 149) me 1, (aisee 35) O57R™% for basement rock 1O02R"S' for sedimentary rock 037K" for alluvium < 60ft thick 0.658% for alluvinm > 608 thick ‘and R is the distance from the center of enetgy release in kilometers where §100 Strong Ground Motion Chap, 3 Wilson (1993) analyzed strong motion records from California to develop an attenu- ation relationship which, using the Arias intensity definition of equation (3.17), can be expressed as logl,(imisee) = M,~2logR— kR ~ 3,990 + 0,365 (1 ~ P) 36) where R = JDP, Dis the minimum horizontal distance to the ventcal projection ofthe fat plane his a correction factor (with a default value of 7.5 km (4.7 mi), kis a coeficient of anelastic absorption (with a default value of zea), and Ps the execedance probability Acceleration and Velocity Spectrum Intensities. Von Thun et al (1988) used 30 strong motion records, primarily from rock outcrops in the western United States and Italy, to develop the attenuation relationships for acceleration specteum intensity ‘and velocity spectrum intensity shown graphically in Figure 3.28, Large earthquakes in these areas are generally accompanied by surface faulting. The use ofthese attenuation rela- tionships is recommended only for areas with similar tectonie coneltions, 600, 250 a g= z = 3 : z 2 50 5 200 § E £ 00 a = i 4 dw s 0 ° 0 8 i018 20 25 00 510 18 20 25 m0 (©) Causative tautcsarce im) (®)-—Caunave fit distance) Figure 328. Arcraton of) aceleaion pcium init ad) velity ses intensity. (ARee Von Thun etal, 1988 Eathskesround ons or design so analysis das, Earthquake Erglering and Si! Dynms I. Rein by Persson of ASCE) 3.5 SPATIAL VARIABILITY OF GROUND MOTIONS, ‘The preceding sections considered the spatial variation of ground motions on a regional scale, Ground motions also vary spatially on local seals, and this local variation can be important for certain types of structures. The longest dimension of mos structures usually ‘small enough that the ground motion atone end is vitally the same as that a the other end For structures such as bridges and pipelines that extend over considerable distances, differ- ent ground motions may occur beneath different parts of the structure, In such cases theSee. 3.5 Spatial Variability of Ground Motions 101 local spatial variation (or incoherence) of the ground motion may exert an important influ= ence on the response of the structure, ‘Spatial incoherence can be caused by a number of factors. One is the traveling-wave ‘oF wave-passage effect, in which nonvertical waves reach differeat points on the ground surface at different times, producing atime shift between the mations a those points (Figure 3.29a). A cause of incoherence in the nearfield is the extended source effec, in which dif ferences in the relative geometry ofthe source and sites produce different time shift, and consequently different motions, atte sites Figure 3.296). Finally, ray-path effects caused by scattering (reflection, refraction, etc.) of waves by inhomogeneities along the travel path (Figure 3.30e) ean cause incoherence. (a) (o) te) 123 Las 1 ey ewepeeiy ony sia Fault Epicenter i M came sorce Pan vew Figure 329. Sources of ncohete grou mons) wavepasage effect cases ‘ncn wavefront 0 eh oan 2, a a fect tmes (extended source fest canes waves due orp at A an Bt each points an? iferen ess {e) searing of waves by herent caesifeteanwaves to ave a ifleent Ieton at diferentes (As Asamson 1981) ‘The similarity between ground motions at different locations can be described inthe time domain or the frequency domain. Consider two points jand kat which accelerograms (0 and a(t) are recorded. The similarity of the motions can be described in the time domain by the cross covariance Cut) = Yatipateeeo 639 where +s atime increment and is the numberof time samples. The autocovariance, Cj (ot Cx) is obtained by analyzing the covariance of an accelerogram against itself. The max imum value of the autocovariance will, obviously, correspond toa value of t= 0, The sim- ilarity ofthe motions inthe frequency domain can be described by the coherency ya(o) = Sule) 638) SOS) Where the smoothed cross-spectrum,5(), is the Fourier transform of the eross-covariance and the autospeetra, (@) and $,(@), are the Fourier transforms of the autocovariances, Gi) and Cu(0). The coherency describes the degree of positive or negative correlation between the amplitudes and phase angles of two time histories at each of their component frequencies, A value of | indicates fll coherence (or perfect correlation), while a value of102 Strong Ground Motion Chap. 3 zero indicates full incoherence (or no correlation). The modulus of the coherency (the ‘square root ofthe sum of the squares of the real and imaginary pars) is called the lagged coherency. Because the wave passage effect from a point source simply introduces a phase shift at each frequency. it does not influence the lagged coherency. Ground motions recorded by dense arrays show that coherency decreases with increas: ing distance between measuring points and with inereasing frequency, as shown in Figure 3.30. Measured coherency functions from dense arrays in California nd Fapan ae sna- ilar to those from the SMART-| array suggesting that they may’be applicable to other areas as, Well although research in this area is continuing (Abrahamson, 1991). Smooth analytical coherency functions (coherency aa function of separation distance an frequency) that match ‘the most significant trends in measured coherency functions have been proposed (Haricharan and Vanmarcke, 1986; Luco and Wong, 1986, Hao e al, 1989; Abrahamson eta, 1991) 3.6 SUMMARY 1. Complete description ofa strong ground motion involves three components of trans lation and three components of rotation. In practice, only the translational components are usually measured, and they are usually measured in oahogonal directions. 2. A numberof different instruments can be used for strong-motion measurement. Each has its own dynamic response characteristics that determine the conditions For which itisbest suited. Older strong-motion instruments are likely to aequite data in analog form, while newer instruments often acquire data digitally 3 Raw strong-motion data may include errors from several sources that require comee tion to produce accurate strong motion records, Strong motion processing is often quired to minimize background noise, corret forthe dynamic response ofthe traus- ducer, and vo correct for measurement errors, 4, Strong ground motions can be quite complicated, and their complete description involves a large amount of data. Fr engineering purposes, the essential characteris- tios of «strong ground motion ean be described in much more compact form using ground motion parameters 8. From an earthquake engincering standpoint, the most important characteristics of a ‘trong ground motion are the amplitude, frequency content, and duration, AI of these ‘characteristics ean significantly influence earthquake damage. Consequently, knowl edge ofthe amplitude, frequency content, or duration alone may not be suf describe accurately the damage potential of a ground motion 6. A variety of parameters are available for description of strong ground motions. Some ofthese parameters describe ground motion amplitude, some describe frequency con- ‘ent, and others deseribe duration. Other parameters reflect two or more ofthese ‘important characteristics. More than one parameter is generally required to chatac- terize a strong ground motion, 7. Commonly used amplitude parameters include peak acceleration, peak velocity and peak slsplacement. The peak acceleration provides a good indication ofthe high-frequency component of a ground motion. The peak velocity and peak displacement deseribe the amplitudes of the intermediate and low-frequency’ components, respectivelySec. 3.6 Summary 10 09 08 o7 06 os. value of coherency 04 jos 02 o4 09 Absolute vale of coherency © 400 200 900 00 500 800 700 600900” 1000 ‘Separation distance (rn) © Figure 3.0 Messi decay of coherency wit increasing frequency a septon stance for M=69 event at hypoceml depth of 3. kn an epee distance of $166 km from SMART: dese aay at Long, Tawar (After Haichaan and ‘anmarce, 196, Soeasic vaiaon af eathakeground motion in spe and tins oural of Braincering Mechanics, Vl. 112, No, 2 Reps bypemsion of ASCE) 103108 10. ML. 2. Be Mu 18, 16. Strong Ground Motion Chap. 3 ‘The frequeney content ofa strong ground motion is generally described through the use of different types of spectra, Fourier spectra and power spectra directly illustrate the frequency content of the motion itself. Response spectra reflect the inluence of the ground motion on structures of different natural periods. A variety of spectral parameters are available to describe the frequency content of a strong ground motion, 1. Strong-motion durations can be described in absolute or telative terms, The bracketed ‘duration, defined as the time between the frst and last exceedances of a threshold acecleration, is based on an absolute measure of acceleration (the threshold acceler ation). Measures of duration based on relative ground motion levels can define very Jong durations for weak ground motions. For engineering purposes, the bracketed ‘duration is most commonly used Sonic parameters reflect the amplitude, frequency content, and duration of a strong ound motion. Although these parameters, uch as mis acceleration, Arias intensity, and response speetrum intensity, may be more difficult to calculate than parameters ‘more commonly used, they often reflect the potential of the motion to produce dam- age more accurately. ‘The characteristics ofa ground motion ata panicular site depend on earthquake mag: nitude and on the distance between the source of the earthquake and the site, Conse- quently, ground motion parameters also vary with earthquake magnitude and source- torsite distance, Measured ground motion data have heen used to develop relationships that predict values of ground motion parameters funotions of earthquake magnitude and source: to-site distance. Predictive relationships are generally empirical; each is obtained by regression on a specific et of data. Consequently, each predictive relationship is appropriate for conditions that are consistent with the conditions of the database, Predictive relationships are not precise; they typically express the mean value of a _ground motion parameter and include a measure of the distribution of values about the ‘mean. The standard deviation ofthe parameter (or the natural logarithm ofthe param. ter) is usually estimated in the development of the predictive relationship. Predictive relationships for variables that decrease with increasing source-to-sit dis- tance are frequently referred to as attenuation relationships. Many attenuation relation ships have been reporced in te lterarare, and the most commonly used relationships, are updated every few years, Ground motions vary on local as well as regional scales. Local variations may cause differential movements ofthe suppors of long structures such as bridges and pipelines. Design and analysis of such structures may requite consideration of local variations, ‘The local variability of ground motions is usually expressed in terms of cohereney. ‘The coherency of two ground motions can be computed—it is @ measure of the correlation ofthe amplitudes and phase angles ofthe motions at diferent frequencies. ‘The coherency of two closely spaced ground motions is higher than that of wo distant ‘ground motions. Also, the coherency of the low-frequency (long-wavelength) com- Ponents ofa pair of motions is higher than that of the high-frequency (short-wave- length) componentsChap. 3 Homework Problems 105 HOMEWORK PROBLEMS Strong motion records canbe obtained fom a variety of sources over the Interne, ften by anonymous {tp Download the strong motion recon indicated by your instructor and use ito sce Problems 1 3.6. The use of = mathematical analysis program such as MATLAB is highly recommended ll all simplify the required computations, 361 Pot te sme history of acceleration an determine: (a) The peak acceleration, (©) The sustained maximum acceleration Crd eyele and Sth cycle), {) The bracketed duration 32 Integrate the time history of acceleration to produce tine histories of velocity and displace- ‘ment Plot the time histories of velocity and displacement and determine the peak velit and peak displacement. 33 Compute and plot the Fourier amplitude spectrum of the stron motion record 3A Determine the predominant period ofthe stong motion record 35. Compute the ms acceleration fr the song motion record 3.6 Compute the Arias intensity forthe strong motion recor, 347 Determine and plot the variations of peak horizontal acceleration with distance fora M, = 6.5 ‘arthquake using the attenuation relitionship of Campbell (1981), 38 Determine and plot the variations of peak horizontal acceleration with distance fora My = 6.5 ‘arthquake at soft rock, hard ock, and alluvium sits using the atenuation relationship of Campbell and Bocorgnis (19%). Which ofthese conditions agrees best with the attenuation relationship of Campbell (1981)? 39 Using the attenuation relationship of Tor etal (1994), determine the probability that aM, =7