Department of Civil and Environmental Engineering, UCLA

Global sea levels are expected to increase up to five meters by the year 2150 (Fox-Kemper

et al., 2021). Ocean water level fluctuations will cause an increase in coastal

groundwater in addition to extreme water levels from storm surge and spring tides. These

fluctuations pose multiple interrelated coastal hazards, including beach erosion and infrastructure

damage. Currently, there has been minimal attention paid to the effects of sea level rise on groundwater,

specifically how rising groundwater levels interact with other hazards, such as earthquake-induced

soil liquefaction. In collaboration with the California Geologic Survey, the United States

Geologic Survey, and the California Seismic Safety Commission, we attempt to

quantify the effects of sea level rise on earthquake-induced liquefaction.

We propose to use a probabilistic liquefaction hazard assessment

(PLHA) (Kramer and Mayfield, 2007) to obtain annual rates of nonexceedance of factor of safety of liquefaction.

This methodology will be used to analyze the effects of sea level rise on earthquake-induced soil liquefaction in

coastal California, which is an area of land which houses over 25 million people and has important transportation and economic implications.

PLHA is an extension of probabilistic seismic hazard analysis (PSHA). It conceptually accounts for the rate of occurrence of all earthquake events that might shake a site of interest. This includes measures of ground motion intensity resulting from each earthquake, the groundwater conditions at the time of the

earthquake, and the likelihood of liquefaction given a particular shaking intensity and groundwater

level. We propose to integrate the liquefaction uncertainty into the hazard calculation, which is a more direct end-to-end approach. The new method integrates the PLHA directly with the PSHA without requiring a deaggregation. The liquefaction hazard is assessed probabilistically for all the rupture scenarios within any fault source model being used. With currently available computing power, running the probabilistic liquefaction hazard inside the hazard integral can be done efficiently (seconds), requiring only slightly more computation time than a PSHA. There is an added benefit that we can deaggregate the liquefaction hazard, which is not possible when the calculations are performed outside the hazard integral.

Using traditional stress-based liquefaction evaluation methods, practitioners are unable to determine the

return period associated with the computed factor of safety. This new methodology allows us to choose the factor of safety against liquefaction for a desired return period. The Uniform California Earthquake Rupture Forecast (Field and Cornell, 2003) Fault Source model is used for both the PSHA and PLHA. Preliminary results

 have been calculated and calibrated for Cardiff State Beach in Southern California. Specifically, preliminary results of a 1-meter of sea level rise scenario using a calibrated groundwater model shows over a 100-year reduction in return period. Initial results suggest that for sites where more events contribute to the hazard the deterministic factor of safety and integrated PLHA factor of safety will differ.  Additional research includes applications of additional ground motion models (2D and 3D over longer time periods) and use of more ground motion and PLHA triggering models.

Global sea levels are expected to increase up to five meters by the year 2150. Due to elevation differences, ocean dynamics, and the uneven redistribution of water from global ice melting, California will experience 1.25:1 of relative amounts of sea level rise (SLR). Ocean fluctuations such as storm surge and spring tides, as well as higher marine water levels can periodically elevate coastal groundwater. These fluctuations pose multiple interrelated coastal hazards, including groundwater salinization, contaminant mobilization and infrastructure damage. Specifically in California, over 2.5M people live along the coast making it the most populated coastline in the continental US. This densely populated area is also a critical infrastructure hub that serves not only California but other US states. The state has vital energy and water facilities, transportation lifelines, shipping ports, and countless residences – all under threat from rising sea levels. One meter of SLR could inflict economic damage of up to $100 billon US dollars to the state of California. However, there has been minimal attention paid to the effects of sea level rise on groundwater, specifically how rising groundwater levels interact with hazards such as earthquake-induced soil liquefaction. Therefore, we attempt to quantify the effects of sea level rise on earthquake-induced liquefaction in coastal California.

We propose to use a probabilistic liquefaction hazard assessment (PLHA) to obtain annual rates of nonexceedance of factor of safety of liquefaction. PLHA is an extension of probabilistic seismic hazard analysis (PSHA). It conceptually accounts for the rate of occurrence of all earthquake events that might shake a site of interest, including ground motion intensities, groundwater conditions at the time of the earthquake, and consequently the likelihood of liquefaction. We propose to integrate the liquefaction uncertainty into the hazard calculation. Existing codes for performing probabilistic liquefaction hazard analysis (PLHA) are decoupled as first a probabilistic seismic hazard analysis (PSHA) is required to obtain a hazard curve. Drawbacks of this method are it only accounts for the influence of magnitude on liquefaction because it uses a small number of magnitudes and selection of the magnitudes requires judgment. The new method integrates the PLHA directly with the PSHA without requiring a deaggregation. The liquefaction hazard is assessed probabilistically for all the rupture scenarios within any fault source model being used. Using traditional stress-based liquefaction evaluation methods, practitioners are unable to determine the return period associated with the computed factor of safety. This new methodology allows us to choose the factor of safety against liquefaction for a desired return period. With currently available computing power, running the probabilistic liquefaction hazard inside the hazard integral can be done efficiently (seconds). There is an added benefit that we can deaggregate the liquefaction hazard, which is not possible when the calculations are performed outside the hazard integral.

The Uniform California Earthquake Rupture Forecast (Field and Cornell, 2003) Fault Source model is used for both the PSHA and PLHA. Preliminary results have been calculated and calibrated for Cardiff State Beach in Southern California. Specifically, preliminary results of a 2-meter of sea level rise scenario using a calibrated groundwater model shows over a 400-year reduction in return period. Initial results suggest that for sites where more events contribute to the hazard the deterministic factor of safety and integrated PLHA factor of safety will differ.  This methodology will be used to analyze the effects of sea level rise on earthquake-induced soil liquefaction in coastal California for multiple sites along the coast. Input data from this analysis will come from the Coastal Geotechnical Database, housed at University of California, Los Angeles (https://www.uclageo.com/coastal_database/). This data was made available by the California Geological Survey. The database contains digitized geotechnical data (standard penetration tests and cone penetration tests) from hospital and school construction locations along the coast of California. 

Global sea levels are expected to increase up to five meters by the year 2150. Higher marine water levels as well as ocean fluctuations will also raise unconfined coastal groundwater levels. These fluctuations pose multiple interrelated coastal hazards. Currently, there has been minimal attention paid to the effects of sea level rise on groundwater, specifically how rising groundwater levels interact with hazards such as earthquake-induced soil liquefaction. We propose to use a probabilistic liquefaction hazard assessment (PLHA) to obtain annual rates of nonexceedance of factor of safety of liquefaction. Our approach integrates the liquefaction uncertainty into the hazard calculation by performing a PLHA for every event in the PSHA. Existing codes for performing probabilistic liquefaction hazard analysis (PLHA) are decoupled as first a probabilistic seismic hazard analysis (PSHA) is required to obtain a hazard curve. Drawbacks of this method are it only accounts for the influence of magnitude on liquefaction because it uses a small number of magnitudes and selection of the magnitudes requires judgment. We adopt the UCERF3 source model, which uses over a million events. For each event, ground motion at the site is a random variable that is used to compute cyclic stress ratio (CSR). Cyclic resistance ratio (CRR) is also a random variable obtained from a liquefaction triggering model. Because CSR and CRR are log normally distributed, we can use a closed-form solution to obtain a distribution for factor of safety (FoS). The rate of the event is then multiplied by the cumulative distribution function (CDF) of FoS and the rate-weighted CDF’s are summed over all events to compute the liquefaction hazard curve. Using efficient vectorized operations, the PLHA runs within seconds for a single site despite the large number of calculations. The code is used to compute the PLHA for an example soil profile at Cardiff State Beach in Southern California for current groundwater conditions and for a 1-meter sea level rise scenario.

Existing codes for performing probabilistic liquefaction hazard analysis (PLHA) are decoupled as first a probabilistic seismic hazard analysis (PSHA) is required to obtain a hazard curve.

Drawbacks of this method are it only accounts for the influence of magnitude on liquefaction because it uses a small number of magnitudes and selection of the magnitudes requires judgment. Our approach integrates the liquefaction uncertainty into the hazard calculation by performing a PLHA for every event in the PSHA. We adopt the UCERF3 source model, which uses over a million events. For each event, ground motion at the site is a random variable that is used to compute cyclic stress ratio (CSR). Cyclic resistance ratio (CRR) is also a random variable obtained from a liquefaction triggering model. Because CSR and CRR are log normally distributed, we can use a closed-form solution to obtain a distribution for factor of safety (FoS). The rate of the event is then multiplied by the cumulative distribution function (CDF) of FoS and the rate-weighted CDF’s are summed over all events to compute the liquefaction hazard curve. Using efficient vectorized operations, the PLHA runs within seconds for a single site despite the large number of calculations. Streamlining the code required developing computationally efficient data storage structures for the UCERF3 model, and a vectorized source-to-site distance calculation. The code is used to compute the PLHA for an example soil profile at Cardiff State Beach in Southern California for current groundwater conditions and for a 1-meter sea level rise scenario.

When the direct shear test is performed in accordance with ASTM guidelines, the measured shear stresses at failure estimate drained strength parameters.  We investigate the possibility of estimating undrained strength using direct shear testing at variable shear displacement rates on specimens composed of various combinations of kaolinite and bentonite. Even at fast displacement rates, constant volume conditions are not achieved in the direct shear device because of changes in specimen height that are large relative to allowable ASTM thresholds for constant volume simple shear testing. However, undrained strengths established by constant volume simple shear testing at slow strain rates are well approximated by direct shear tests conducted at fast shear displacement rates (time to failure < t₅₀/8, where t₅₀=time to 50% consolidation in a conventional oedometer test). Because of the simplicity of direct shear testing, such estimates of undrained strength may be useful in engineering practice when access to a simple shear device is limited. Nevertheless, fast direct shear tests have shortcomings, including lack of control of rate effects, and constant volume testing is recommended for critical projects.

We investigate through laboratory testing the volume change characteristics of peaty organic soil from Sherman Island, California under static conditions (consolidation, secondary compression) and post-cyclic conditions. Incremental consolidation tests indicate the material to be highly compressible (C_c = 3.9, C_r = 0.4) and prone to substantial ageing from secondary compression (C_a/C_c = 0.05 following virgin compression). Strain-controlled cyclic triaxial testing of the peat finds the generation of cyclic pore pressures for cyclic shear strain levels beyond approximately 0.5-1.0%, with the largest residual pore pressure ratios r_ur (cyclic residual pore pressure normalized by pre-cyclic consolidation stress) being approximately 0.2-0.4. Post cyclic volume change occurs from pore pressure dissipation and secondary compression. The level of post-cyclic secondary compression increases with r_ur. Many of these phenomena have not been documented previously and suggest the potential for seismic freeboard loss in levees due to mechanisms other than shear failure.

This research involved analysis and field testing of several foundation support components for highway bridges. Two classes of components were tested - cast-in-drilled-hole (CIDH) reinforced concrete piles (drilled shafts) and an abutment backwall. The emphasis of this document (Part I of the full report) is CIDH shafts.

CIDH shafts are among the most common support structures in highway construction. Typically, drilled shafts have simple, prismatic geometries; yet, they display a complex, inelastic response under applied loading. The two major factors that affect their behavior are the interaction between the shaft and surrounding soil media, and the material inelasticity of the shaft itself. In this report we document the results of two single shaft tests and one shaft group test. All specimens are two-feet diameter reinforced concrete drilled shafts that extend approximately 24ft below ground line. The single shaft specimens include one in a flagpole configuration extending 13.3ft above ground line and the other capped at the surface in a fixed-head configuration. The group test specimen had 9 individual shafts in a 3 by 3 configuration anchored at the ground surface (with a moment connection) in a reinforced concrete cap. The test site consists primarily of low plasticity alluvial clay that is expected to exhibit an undrained response to the cyclic lateral loading. The quasi static loading was applied with a hydraulic control system in displacement-control mode, with the full suite of loading taking several days to complete for each test. The test data have been reduced to provide complete load-deflection backbone curves for loading in both directions, curvature profiles at pre-yield deflection levels, hysteresis curves documenting the cyclic behavior of the shaft soil system at pre-yield displacements, p-y curves for the single shaft specimens, and group interaction factors for the group specimen.

Pre-test response predictions of the CIDH specimens were obtained via (1) a three dimensional finite element model, (2) a macro-element model, developed at UCLA, and (3) the so-called strain wedge model adopted from the literature. Simulation results were compared with each other and with field measurements. It was observed that all of the three numerical approaches yielded reasonably accurate predictions for these small diameter shafts. We provide p-y curves in the API format calibrated to the test data and show that those curves improve the accuracy of predictions relative to generic p-y curves in commonly used design guidelines published by the American Petroleum Institute (API).

The p-y curves obtained from the experiments are shown to differ from what would be predicted using standard API models, with the data indicating a stronger and stiffer response at shallow depths where the shaft-soil interaction is most pronounced. We also compare results of various tests to evaluate head fixity effects on p-y curves and the adequacy of the diameter effect built into API p-y guidelines.

This research involved analysis and field testing of numerous foundation support components for highway bridges. Two classes of components were tested - cast-in-drilled-hole (CIDH) reinforced concrete piles (drilled shafts) and an abutment backwall. The emphasis of this document (Part II of the full report) is abutment backwall elements.

The backwall test specimen was backfilled to a height of 5.5 up from the base of the wall with well-compacted silty sand backfill material (SE 30). The wall is displaced perpendicular to its longitudinal axis. Wing walls are constructed with low-friction interfaces to simulate 2D conditions. The backfill extends below the base of the wall to ensure that the failure surface occurs entirely within the sand backfill soil, which was confirmed following testing. The specimen was constructed and tested under boundary conditions in which the wall was displaced laterally into the backfill and not allowed to displace vertically.

A maximum passive capacity of 497 kips was attained at a wall displacement of about 2.0 in, which corresponds to a passive earth pressure coefficient Kp of 16.3. Strain softening occurs following the peak resistance, and a residual resistance of approximately 460 kips is achieved for displacements > 3.0 inch. The equivalent residual passive earth pressure coefficient is Kp = 15.1 and the equivalent uniform passive pressure at residual is approximately 5.1 ksf, which nearly matches the value in the 2004 Seismic Design Criteria of 5.0 ksf. The average abutment stiffness K50 was defined as a secant stiffness through the origin and the point of 50% of the ultimate passive force. For an abutment wall with a backfill height H of 5.5 ft, this stiffness was found to be K50 = 50 kip/in per foot of wall. The load-deflection behavior of the wall-backfill system is reasonably well described by a hyperbolic curve.

The passive pressure resultant is under predicted using classical Rankine or Coulomb earth pressure theories. Good estimates of capacity are obtained using the log-spiral formulation and the method-of-slices. The method-of-slices approach is implemented with a log-spiral hyperbolic method of evaluating backbone curves that provides a good match to the data.