Geotechnical Earthquake Engineering and Soil–Structure Interaction

Introduction and Context

Earthquake damage is governed not only by the intensity of shaking at the ground surface but also by the interaction between seismic waves, soil deposits, foundations, and structures. Two sites subjected to the same bedrock motion may experience substantially different surface motions because of differences in soil stiffness, layering, groundwater conditions, basin geometry, and nonlinear soil behavior. In addition, earthquake-induced loss of soil strength can cause liquefaction, lateral spreading, bearing-capacity failure, settlement, landslides, and excessive foundation deformation.

Geotechnical earthquake engineering therefore links engineering seismology, soil mechanics, structural dynamics, foundation engineering, and engineering geology. Its objective is not simply to calculate a force, but to predict the deformation and failure mechanisms that control performance.

Learning Objectives

By the end of this chapter, the reader should be able to:

1. Explain how local soil conditions modify earthquake ground motion.

2. Evaluate the mechanisms and consequences of cyclic soil behavior and liquefaction.

3. Identify and assess lateral spreading, seismic slope instability, and seismic earth-pressure problems.

4. Distinguish inertial and kinematic soil–structure interaction.

5. Select appropriate analytical, numerical, and field-investigation methods.

6. Interpret uncertainty and develop practical mitigation and design strategies.

1. Site Response and Cyclic Soil Behavior

1.1 Seismic wave propagation through soil

A soil deposit modifies incoming bedrock motion through reflection, refraction, impedance contrast, damping, and nonlinear hysteretic behavior. For a linear elastic soil layer, shear-wave velocity is

\[

V_s=\sqrt{G/\rho}

\]

where \(V_s\) is shear-wave velocity, \(G\) is small-strain shear modulus, and \(\rho\) is mass density. The relation applies to a homogeneous, isotropic, linear elastic material under small-strain conditions.

For a layer of thickness \(H\) over a relatively stiff base, the fundamental quarter-wavelength period is approximately

\[

T_0=\frac{4H}{V_s}

\]

For layered deposits, the equivalent travel-time approximation is

\[

T_0\approx4\sum_{i=1}^{n}\frac{H_i}{V_{s,i}}

\]

where \(H_i\) and \(V_{s,i}\) are the thickness and shear-wave velocity of layer \(i\). This is an approximate one-dimensional estimate; it does not represent basin-edge effects, irregular topography, or strong nonlinear response.

A site tends to amplify motion near its natural periods, particularly when the period of significant ground motion approaches the site period. Soft, deep deposits generally lengthen the predominant period, whereas shallow stiff sites tend to preserve shorter-period motion. Amplification is not uniform across all frequencies: material damping and radiation into the underlying medium reduce resonant amplification.

1.2 Small-strain and nonlinear stiffness

The maximum or small-strain shear modulus is commonly expressed as

\[

G_{max}=\rho V_s^2

\]

As cyclic shear strain increases, soil stiffness decreases and damping increases. A modulus-reduction curve, \(G/G_{max}\), and a damping curve are therefore needed for equivalent-linear or nonlinear site-response analysis. Equivalent-linear analysis represents soil properties by strain-compatible secant values and iterates until the effective strain is consistent with the selected properties. Nonlinear analysis updates stress and strain continuously and is preferable when strains are large, strong stiffness degradation is expected, or pore-pressure generation is important.

The stress–strain response is hysteretic. The area within a cyclic loop represents energy dissipated per unit volume. Real soils also exhibit permanent strain, fabric change, anisotropy, and pressure-dependent stiffness, effects that simple linear models cannot reproduce.

1.3 Site-response workflow

A defensible site-response study normally includes:

• geological and geomorphological interpretation;
• boreholes, cone penetration tests, laboratory tests, and groundwater observations;
• development of a representative shear-wave velocity profile;
• selection and scaling of bedrock motions consistent with the hazard level;
• one-dimensional analysis where horizontal layering is reasonable;
• two- or three-dimensional analysis for basins, ridges, valleys, abrupt lateral changes, or major infrastructure;
• sensitivity studies for stiffness, damping, groundwater, and input-motion uncertainty; and
• engineering interpretation in terms of acceleration, velocity, displacement, duration, and strain.

A surface response spectrum is not a unique property of a soil type. It depends on the input motion, stratigraphy, nonlinear behavior, and analysis assumptions.

2. Liquefaction and Cyclic Mobility

2.1 Mechanism

Liquefaction is a severe reduction in soil stiffness and effective stress caused by cyclic loading under undrained or nearly undrained conditions. It is most commonly associated with loose, saturated, clean sands and silty sands, although susceptible behavior can occur in some low-plasticity silts and soils with unusual fabric.

The effective-stress relation is

\[

\sigma'=\sigma-u

\]

where \(\sigma'\) is effective stress, \(\sigma\) is total stress, and \(u\) is pore-water pressure. During cyclic shearing, contraction of a loose soil tends to generate excess pore pressure \(\Delta u\). If \(\Delta u\) approaches the initial vertical effective stress \(\sigma'_{v0}\), the effective stress and shear resistance may become very small.

A commonly used state indicator is the excess pore-pressure ratio

\[

r_u=\frac{\Delta u}{\sigma'_{v0}}

\]

The ratio is dimensionless and should be interpreted with the stress reference clearly stated. A value near unity indicates severe loss of vertical effective stress, but liquefaction-related deformation can begin before complete stress cancellation.

The phenomenon is not necessarily a fluid-like state. Soil may experience cyclic softening, large shear strains, and flow or spreading without satisfying an idealized condition of zero effective stress everywhere.

2.2 Triggering assessment

The simplified procedure compares cyclic stress ratio (CSR), representing the earthquake demand, with cyclic resistance ratio (CRR), representing soil resistance:

\[

FS_{liq}=\frac{CRR}{CSR}

\]

A simplified form of CSR is

\[

CSR=0.65\left(\frac{a_{max}}{g}\right)\left(\frac{\sigma_{v0}}{\sigma'_{v0}}\right)r_d

\]

where \(a_{max}\) is peak horizontal ground acceleration, \(g\) is gravitational acceleration, \(\sigma_{v0}\) and \(\sigma'_{v0}\) are total and effective vertical stresses, and \(r_d\) is a depth-reduction factor. The coefficient 0.65 converts a peak acceleration idealization into an equivalent cyclic demand. The equation is a screening approximation for level or gently sloping ground and must be applied with an earthquake magnitude or duration adjustment when comparing demand with CRR.

CRR is commonly estimated from corrected SPT blow counts, CPT resistance, or shear-wave velocity. Corrections account for equipment energy, overburden stress, fines content, rod length, borehole diameter, and other test conditions. Because empirical databases and correction procedures differ, the method must be consistent with the selected design standard.

A factor of safety greater than one does not guarantee acceptable performance. Post-liquefaction settlement, lateral spreading, bearing failure, and pile demands may remain unacceptable. Conversely, a triggering index is not a direct prediction of deformation.

2.3 Consequences and mitigation

Principal consequences include:

• total and differential settlement;
• bearing-capacity failure and foundation rotation;
• lateral spreading toward a free face or channel;
• buoyant uplift of buried tanks and pipelines;
• flotation or rupture of utilities; and
• loss of lateral support to piles.

Mitigation options include densification, stone columns, gravel drains, rigid inclusions, deep foundations, replacement of susceptible soil, drainage improvements, ground improvement by compaction grouting, and structural isolation from the moving ground. Selection depends on depth, groundwater, access, environmental constraints, construction vibration, and the deformation that can be tolerated.

3. Lateral Spreading and Seismic Slope Instability

3.1 Lateral spreading

Lateral spreading is the permanent horizontal displacement of gently sloping or nearly level ground toward a free face, such as a riverbank, or toward a zone of lower elevation. It commonly combines liquefaction, cyclic softening, and downslope movement. Damage is often controlled by displacement rather than by peak acceleration.

Assessment should identify the free-face geometry, thickness and continuity of susceptible layers, groundwater conditions, seismic demand, and consequences for bridges, pipelines, retaining walls, and buildings. Empirical displacement models can provide screening estimates, but they are uncertain outside the database from which they were developed. For critical facilities, site-specific deformation analysis and structural accommodation may be required.

3.2 Seismic slope stability

Earthquake loading can increase driving forces, reduce effective stress, and generate excess pore pressure. A pseudo-static analysis introduces a horizontal coefficient \(k_h\), and sometimes a vertical coefficient \(k_v\), into limit-equilibrium calculations. The pseudo-static factor of safety is conceptually

\[

FS=\frac{\text{available resisting force or moment}}{\text{seismic driving force or moment}}

\]

The method is useful for screening and for slopes expected to undergo limited deformation. It is not a rigorous prediction of displacement because earthquake acceleration reverses with time and may exceed the selected coefficient only briefly.

For slopes that may deform, Newmark sliding-block analysis estimates permanent displacement by integrating relative motion whenever the driving acceleration exceeds the yield acceleration. Its reliability depends on the selection of yield acceleration, input motion, block geometry, and treatment of pore-pressure changes. Coupled effective-stress finite-element or finite-difference analysis is more suitable where soil softening, complex geometry, or progressive failure is important.

4. Seismic Earth Pressures and Retaining Systems

Earthquake shaking changes the lateral stress state behind retaining walls. The classical Mononobe–Okabe method extends Coulomb wedge theory by using horizontal and vertical seismic coefficients. It assumes a planar failure surface, a rigid wall, homogeneous soil, defined wall friction, and a sufficiently long-duration condition in which the active wedge can mobilize.

The method can become unreliable for flexible walls, cohesive or partially saturated soils, sloping backfills, irregular geometry, high groundwater, strong wall–soil interaction, or walls that cannot move sufficiently to mobilize active conditions. It also does not directly predict transient hydrodynamic pressure, seismic settlement, or wall displacement.

Professional design should consider:

• active, at-rest, or passive pressure according to wall movement;
• dynamic earth pressure and its point of application;
• drainage and hydrostatic pressure;
• wall inertia and foundation flexibility;
• sliding, overturning, bearing, and global stability; and
• deformation compatibility with adjacent structures and utilities.

For critical walls, displacement-based analysis is often more informative than applying a single amplified static pressure.

5. Foundation Response and Soil–Structure Interaction

5.1 Nature of interaction

Soil–structure interaction (SSI) occurs because the foundation modifies the motion of the supporting soil, while the deformable soil modifies the motion and forces in the structure. Two effects are distinguished:

• Kinematic interaction: foundation motion differs from free-field ground motion because of embedment, foundation averaging, wave scattering, and foundation geometry.
• Inertial interaction: structural inertia produces foundation translation, rocking, and sometimes sliding, changing the forces and deformations in both soil and structure.

A fixed-base model may be conservative for some force quantities but unconservative for others. SSI generally lengthens the system period and adds radiation and hysteretic damping, but it can increase displacement, rocking, settlement, and demands on nonstructural components.

5.2 Foundation stiffness and impedance

For a foundation idealized as a rigid body, soil support can be represented by frequency-dependent impedance functions. In simplified form,

\[

\mathbf{F}(\omega)=\mathbf{K}(\omega)\mathbf{u}(\omega)

\]

where \(\mathbf{F}\) is the vector of foundation forces and moments, \(\mathbf{u}\) is the vector of translations and rotations, \(\omega\) is circular frequency, and \(\mathbf{K}(\omega)\) is a complex dynamic stiffness matrix. Its real part represents stiffness and its imaginary part represents damping-related effects. This formulation is appropriate for linear or equivalent-linear frequency-domain analysis; nonlinear contact, gapping, uplift, and soil yielding require more advanced treatment.

For preliminary judgment, foundation flexibility is influenced by soil shear modulus, foundation dimensions, embedment, contact condition, and frequency. Static spring models can be useful for serviceability or preliminary seismic analysis, but they should not be treated as universal constants.

5.3 Foundation failure modes

Seismic foundation performance should be checked for:

• liquefaction-induced bearing failure;
• cyclic degradation of bearing resistance;
• sliding and overturning;
• uplift and loss of contact;
• differential settlement and rocking;
• pile bending, buckling, downdrag, and kinematic loading; and
• interaction with adjacent foundations, basements, and retaining systems.

Pile foundations do not simply transfer inertial load to deep soil. They may be loaded by structural inertia at the pile head and by lateral soil movement along the embedded shaft. A liquefied layer can lose lateral support, while nonliquefied crusts and sloping ground can impose large curvature demands.

6. Analytical Methods and Design Strategy

A rational analysis level should match hazard, consequence, and uncertainty.

Level 1: screening

Use geological mapping, groundwater data, index tests, published hazard maps, simplified liquefaction checks, and preliminary slope or retaining-wall calculations. This level identifies obvious hazards but should not justify critical design by itself.

Level 2: site-specific engineering analysis

Use CPT, SPT, shear-wave velocity measurements, laboratory cyclic tests, one-dimensional site response, liquefaction triggering and settlement estimates, and deformation-compatible foundation design. Sensitivity studies are essential because small changes in groundwater level or soil state can significantly affect results.

Level 3: advanced analysis

Use nonlinear effective-stress finite-element or finite-difference models, three-dimensional basin analysis, dynamic contact, coupled pore-fluid flow, and soil constitutive models calibrated against laboratory and field evidence. Advanced analysis does not eliminate uncertainty; it can expose model-form uncertainty and requires strong verification and validation.

Best practice is to maintain a clear chain from hazard definition to ground model, constitutive assumptions, input motions, computed response, performance objective, and design decision. Independent review is appropriate for major infrastructure.

7. Illustrative Applied Case: Riverfront Hospital on Liquefiable Alluvium

Background

Consider a proposed five-storey hospital near a river. The site consists of 2 m of fill, 8 m of loose saturated sand, and dense sand over weathered rock. The groundwater table is 1.5 m below grade. A design earthquake produces a peak horizontal acceleration of 0.30g. The hospital must remain operational after the earthquake.

Problem

The loose sand may liquefy. Lateral spreading toward the river could impose large ground displacements on foundations and buried services. A conventional shallow-foundation design based only on bearing capacity would not address these hazards.

Analysis

The investigation combines CPT soundings, borings, shear-wave velocity measurements, laboratory classification, groundwater monitoring, and riverbank geometry. Corrected penetration resistances indicate low cyclic resistance in portions of the sand. Simplified triggering calculations show that the factor of safety is below the project target in the most susceptible layers. Deformation analysis indicates that settlement and riverward movement, rather than peak structural acceleration, govern performance.

Approach and solution

The design team adopts a performance-based strategy:

1. Remove and replace shallow uncontrolled fill.

2. Improve the liquefiable sand beneath the building using a grid of densification columns, verified by post-treatment CPTs.

3. Extend ground improvement toward the riverbank to reduce differential movement.

4. Use a stiff piled foundation tied by a reinforced mat, with pile design considering both inertial and kinematic loads.

5. Install flexible, accessible utility connections.

6. Design the riverbank and retaining elements for seismic deformation, drainage, and global stability.

7. Perform nonlinear site-response analysis and structural SSI analysis using multiple ground motions.

8. Instrument the site with piezometers and settlement points for construction verification and post-earthquake assessment.

Outcome

The selected system does not assume that ground movement can be completely prevented. Instead, it reduces liquefaction susceptibility, limits differential movement, provides foundation redundancy, and allows inspection and repair of utilities. The hospital’s continuity objective is addressed through both geotechnical measures and operational planning.

Learning points

The case demonstrates that liquefaction assessment is only the first step. A satisfactory design must also address lateral spreading, foundation compatibility, utility performance, uncertainty, monitoring, and the facility’s required post-earthquake function.

8. Limitations, Governance, and Contemporary Practice

Geotechnical earthquake predictions are uncertain because subsurface conditions are spatially variable, earthquake duration and frequency content vary, empirical correlations have limited databases, and constitutive models simplify fabric and drainage behavior. Results should therefore be reported as ranges or scenarios where appropriate, not as unjustifiably precise values.

Regulatory requirements differ by jurisdiction, but professional practice generally requires investigation, hazard classification, design ground motions, liquefaction screening, slope and retaining-wall checks, foundation design, construction quality control, and documentation of assumptions. Engineers have an ethical duty to communicate uncertainty, avoid concealing adverse data, and distinguish code compliance from acceptable risk.

Modern tools include distributed sensing, high-resolution CPT interpretation, downhole arrays, remote sensing, probabilistic seismic hazard analysis, and machine-learning-assisted pattern recognition. AI can help classify data, detect anomalies, and accelerate parametric studies; however, it cannot replace geological judgment, physical testing, model verification, or professional responsibility. Training-data bias and opaque predictions are particularly problematic for rare failure mechanisms.

Best practice includes conservative but realistic ground models, redundant investigation methods, calibrated constitutive models, multiple records, explicit performance objectives, deformation-based evaluation, independent checking, and observational monitoring during construction and after major earthquakes.

Chapter Summary

Local soil conditions can amplify, lengthen, or otherwise modify earthquake motion. Site response depends on stiffness, damping, layering, geometry, and nonlinear behavior. Cyclic loading in saturated susceptible soils can generate excess pore pressure, reduce effective stress, and produce liquefaction, settlement, lateral spreading, and foundation failure. Seismic slope and retaining-wall assessment must distinguish screening methods from deformation-compatible analysis. Soil–structure interaction comprises kinematic and inertial effects and can alter period, damping, force, displacement, rocking, and foundation demand. Reliable practice integrates field investigation, laboratory and in situ testing, appropriate numerical models, uncertainty assessment, performance objectives, and construction verification. Advanced technology supports but does not replace engineering judgment.

Key Terms

Site response: Modification of earthquake motion by local soil and geological conditions.

Shear-wave velocity: Propagation velocity of shear waves through a material.

Small-strain shear modulus: Initial soil shear stiffness, commonly calculated as \(G_{max}=\rho V_s^2\).

Equivalent-linear analysis: Frequency-domain analysis using strain-compatible stiffness and damping.

Nonlinear site response: Time-domain analysis in which stress–strain properties evolve during shaking.

Effective stress: Total stress minus pore-water pressure.

Excess pore pressure: Earthquake-induced increase in pore-water pressure.

Liquefaction: Severe cyclic loss of stiffness and effective stress in susceptible saturated soil.

Cyclic stress ratio: Normalized seismic shear-stress demand.

Cyclic resistance ratio: Normalized soil resistance to cyclic loading.

Lateral spreading: Permanent ground movement toward a free face or lower elevation.

Pseudo-static analysis: Stability analysis using constant equivalent seismic coefficients.

Newmark sliding block: Deformation method integrating motion beyond a yield acceleration.

Seismic earth pressure: Lateral soil pressure generated or modified by earthquake loading.

Kinematic interaction: Modification of foundation motion by soil and foundation geometry.

Inertial interaction: Soil response caused by structural inertia transmitted through the foundation.

Impedance function: Frequency-dependent relation between foundation force and displacement.

Kinematic pile loading: Pile demand caused by soil movement along the embedded pile.

Performance-based design: Design organized around explicit damage and functionality objectives.

Review Questions

Short Answer Questions

1. Why can two sites exposed to the same bedrock motion experience different surface motions?

2. Define effective stress and explain its significance in liquefaction.

3. Distinguish kinematic interaction from inertial interaction.

4. Why is pseudo-static slope analysis not a direct prediction of permanent displacement?

5. State two consequences of lateral spreading.

Long Answer Questions

1. Explain the mechanisms controlling one-dimensional nonlinear site response and compare equivalent-linear and nonlinear analysis.

2. Describe a complete liquefaction assessment, including investigation, triggering, deformation, and mitigation.

3. Discuss the assumptions and limitations of the Mononobe–Okabe approach.

4. Explain how foundation flexibility, rocking, uplift, and pile–soil interaction influence seismic structural response.

5. Develop a geotechnical earthquake-engineering workflow for a critical facility on deep alluvial deposits.

Analytical / Application Questions

1. A 20 m uniform soil layer has a shear-wave velocity of 200 m/s. Estimate its quarter-wavelength fundamental period and discuss the limitations of the estimate.

2. At a depth in saturated sand, \(a_{max}=0.25g\), \(\sigma_{v0}=180\) kPa, \(\sigma'_{v0}=100\) kPa, and \(r_d=0.9\). Calculate the simplified CSR using the equation in this chapter and explain what additional information is needed before judging liquefaction triggering.

3. A bridge abutment is adjacent to a liquefiable riverbank. Propose an investigation and design strategy that addresses liquefaction, lateral spreading, pile demand, abutment movement, and utility continuity.

Multiple Choice Questions with Answers

Question 1

Which expression relates small-strain shear modulus to shear-wave velocity and mass density?

A. Gmax = ρ/Vs²

B. Gmax = ρVs²

C. Gmax = Vs/ρ

D. Gmax = ρ + Vs²

Correct Answer: B. Gmax = ρVs²

Question 2

For a homogeneous soil layer, the approximate quarter-wavelength fundamental period is:

A. H/Vs

B. 2H/Vs

C. 4H/Vs

D. Vs/(4H)

Correct Answer: C. 4H/Vs

Question 3

Liquefaction is primarily associated with:

A. Dry, dense gravel under drained loading

B. Saturated susceptible soil undergoing cyclic loading

C. Competent rock under static loading

D. Unsaturated stiff clay during slow loading

Correct Answer: B. Saturated susceptible soil undergoing cyclic loading

Question 4

If total stress is 200 kPa and pore-water pressure is 75 kPa, the effective stress is:

A. 75 kPa

B. 125 kPa

C. 200 kPa

D. 275 kPa

Correct Answer: B. 125 kPa

Question 5

In the simplified CSR equation, rd is a:

A. Depth-reduction factor

B. Damping ratio

C. Density correction

D. Rocking coefficient

Correct Answer: A. Depth-reduction factor

Question 6

Lateral spreading generally involves ground movement:

A. Vertically upward only

B. Toward a free face or lower elevation

C. Only toward the earthquake epicentre

D. Only within intact bedrock

Correct Answer: B. Toward a free face or lower elevation

Question 7

Which method estimates permanent sliding-block displacement by integrating motion beyond yield acceleration?

A. Mononobe–Okabe method

B. Newmark sliding-block analysis

C. Quarter-wavelength method

D. Coulomb compression test

Correct Answer: B. Newmark sliding-block analysis

Question 8

Kinematic soil–structure interaction refers primarily to:

A. Structural inertia increasing foundation force

B. Modification of foundation motion by soil and foundation geometry

C. Chemical degradation of concrete

D. Static bearing-capacity increase

Correct Answer: B. Modification of foundation motion by soil and foundation geometry

Question 9

A major limitation of pseudo-static slope analysis is that it:

A. Cannot use soil strength parameters

B. Assumes earthquake acceleration is constant in time

C. Requires no slope geometry

D. Directly predicts all permanent displacement

Correct Answer: B. Assumes earthquake acceleration is constant in time

Question 10

For the case with amax = 0.25g, σv0 = 180 kPa, σ′v0 = 100 kPa, and rd = 0.9, the simplified CSR is approximately:

A. 0.146

B. 0.263

C. 0.405

D. 0.900

Correct Answer: B. 0.263