Levees and Earthquakes: Averting an Impending Disaster

(This project was done as part of my doctoral work at UCLA)

This description of the project is taken from the website of the ucla geotechnical group. For the original page please click here.


The Sacramento / San Joaquin Delta is the hub of California’s water delivery system and serves over 20 million people in Southern California. It is composed of “islands” circumscribed by approximately 1100 miles of levees that constantly impound water, preventing flooding of the islands. Peaty organic soils that comprise much of the Delta have oxidized and blown away over the past 150 years, resulting in rapid subsidence. Many islands are now 3 to 5 meters below sea level.

Subsidence Process (Mount and Twiss 2005)

The levees are composed of uncompacted loose dredged soils that are saturated, and are therefore susceptible to liquefaction. The Delta lies on the eastern margin of the San Andreas fault system in a zone with moderate seismic hazard.

Seismic Hazard in the Delta (DWR: Delta Risk Management Study 2009)

Recent projections indicate that a Delta earthquake could result in widespread levee failures and the simultaneous inundation of multiple islands. Such an event would locally reverse the flow direction in the Delta, causing the intrusion of saline water from the west that would contaminate the water supply and halt water delivery for a period of years. The Delta earthquake scenario is one of the most important hazards facing the United States because the consequences of such an event would be unimaginably catastrophic. In 2004 a single levee breach in the Upper Jones Tract cost over $200 million in total losses. It took three weeks to repair the breach and five months to de-water the island. In 2009 a study from the Department of Water Resource estimated that a moderate earthquake in the Delta could disrupt the fresh water delivery system for 20 to 30 months in result in around $40 billion in economic losses.

Failure of the Upper Jones Tract (2004)

The research objective of this project is to investigate the deformation potential of liquefiable levees supported on peaty organic soils through centrifuge testing. Two large scale 9 m radius centrifuge tests (named “RCK01” and “RCK02”) were conducted at the NEES@UCDavis experimental facility on heavily instrumented models. A series of 11 small scale 1 m radius Schaevitz centrifuge tests were preliminarily conducted in order to estimate the response of the models during testing in the large scale tests.

Large centrifuge (9 m radius) at the NEES@UCDavis experimental facility

Each 9 m radius test consisted of two phases. First, a levee built with modeling clay was placed on top of a peat layer. Several ground motions were applied in order to study the interaction between soft peat and stiffer levee fills, and gain insight into the volume change behavior of organic soils under cyclic loading.

Schematic view of the first phase with a clayey levee

Sideview of the model before going on the arm

Secondly, the clayey levee was removed and replaced with a saturated sandy levee. The levee was then subjected to a target ground motion representative of the seismic risk in the Delta to study deformation potential of saturated sandy levees resting on peat due to liquefaction of the levee fill. A complex pumping system was installed to simulate the levee retaining on water on side.

Schematic view of the second phase with a sandy levee and standing water on one side

 The peat thickness in the second large scale investigation was varied in order to study its influence on system response. For more information about centrifuge testing, please visit the website of the UC Davis Center for Geotechnical Modeling.


To download data from our centrifuge tests, please visit the page dedicated to our project on NEEShub.


The following videos are from our second experiment (RCK02). The first video is a side view of our clayey levee atop peat during an earthquake with a peak ground acceleration (PGA) of 0.6g. During our tests the centrifugal acceleration reached 57g meaning that an earthquake that would last 57s in “real life” would last 1s in our centrifuge tests.

In the previous videos the ground motion is barely noticeable because of the speed at which it occurs. The next video shows the same video 50 times slower.

In the second part of our experiment we substituted the clayey levee with a sandy levee, and create a reservoir on side of the levee to model the real field conditions. The following video shows a view of the levee crest during an earthquake at model scale with a PGA of 0.4g. You can see the levee fill liquefying, losing its shape and settling, allowing water from the reservoir to rush into the island.

The last video shows the same earthquake from the side, slowed down 10 times. Once again you can clearly see the levee fill liquefying, resulting in a loss of freeboard of the levee, resulting in water rushing in the island.

Publications from this Research Project:

Cappa R., Yniesta S., Brandenberg S.J., Lemnitzer A., Stewart J.P. (2014). “Averting an Impending Disaster. Data Report for Centrifuge Experiments 12L and 13L.” Data report for NEES.

Cappa R., Yniesta S., Brandenberg S.J., Lemnitzer A., Stewart J.P. (2014). “Averting an Impending Disaster. Data Report for Centrifuge Experiments 14L and 15L.” Data report for NEES.

Cappa, R., Yniesta, S., Lemnitzer, A., Brandenberg, S.J., and Stewart, J.P. (2014). “Centrifuge Experiments to Evaluate the Seismic Performance of Levees on Peaty Soils in the Sacramento-San Joaquin Delta“ Proceedings, Dam Safety Conference, San Diego, CA, September 21-25, 2014

Cappa, R., Yniesta, S., Lemnitzer, A., Brandenberg, S. and Shafiee, A. (2015). “Settlement Estimations of Peat during Centrifuge ExperimentsProceedings, International Foundations Congress and Equipment Exposition (IFCEE), San Antonio, Texas, March 17- 21, 2015

Yniesta, S., Cappa, R., Lemnitzer, A. and Brandenberg, S. (2015). “Centrifuge Testing of Levees: Saturation Techniques during Model ConstructionProceedings, International Foundations Congress and Equipment Exposition (IFCEE), San Antonio, Texas, March 17- 21, 2015

Yniesta, S., Lemnitzer, A., Cappa, R., and Brandenberg, S.J. (2015) “Vacuum Pluviation Device for Achieving Saturated SandGeotechnical Testing Journal, 38 (3), 355-360

Cappa, R., Yniesta, S., Brandenberg, S.J and Lemnitzer, A. “Settlements and excess pore pressure generation in peaty soils under embankments during cyclic loadingProceedings, 6th International Conference on Earthquake Geotechnical Engineering (6ICEGE), Christchurch, New Zealand, November 1-4, 2015

Lemnitzer, A., Cappa, R., Yniesta, S. and Brandenberg, S.J. “Centrifuge Testing of Model Levees atop peaty soils: experimental data”. Earthquake Spectra, August 2016, Vol. 32, No. 3, pp. 1903-1924.

Cappa, R., Yniesta, S., Lemnitzer, A. and Brandenberg, S.J “Cyclic and Post-Cyclic Behavior of Levees atop Peaty Organic Soils during Centrifuge Testing” Journal of Geotechnical and Geoenvironmental Engineering (In Preparation)


DRMS, URS Corporation and Jack Benjamin and Associates Inc. (2009). “Delta Risk Management Strategy. Phase 1 Final Report.” California Department of Water Resources.

Mount, Jeffrey; & Twiss, Robert. (2005). “Subsidence, Sea Level Rise, and Seismicity in the Sacramento–San Joaquin Delta.” San Francisco Estuary and Watershed Science, 3(1). jmie_sfews_10966. Retrieved from: http://escholarship.org/uc/item/4k44725p


Funding was provided by the George E. Brown Network for Earthquake Engineering Simulation through contract numbers 1208170. This award is part of the National Earthquake Hazards Reduction Program (NEHRP).

Principal Investigator:

Scott J. Brandenberg, UCLA, Ph.D., P.E.

Jonathan P. Stewart, UCLA, Ph.D., P.E.

Anne Lemnitzer, UCI, Ph.D., P.E.

George Mylonakis, University of Bristol, Ph.D., P.E.