Evolution of geotechnical risk analysis in North American practice - Risk and Reliability in Geotechnical Engineering - page 469

to risk-informed planning, creating its Risk Management Center (RMC) as a platform for
the transformation in the programs for dam and levee safety. The RMC in recent years has
championed risk-based analysis of internal erosion in dams, dam and levee-screening crite-
ria, and international standards for tolerable risk.
A number of important developments have come out of this work on dam safety risk,
especially that done by USBR and private sector hydropower operators. In recent years,
the International Committee on Large Dams (ICOLD) has begun promulgating guidance
on dam safety risk analysis methods (ICOLD 2005). Scott (2011) has presented a detailed
overview of the major contributions of this body of work. Among other things, he cites the
following contributions as important, and it is hard to disagree with his assessment:
• Potential failure mode analysis
• Hazard analysis of annual exceedance probabilities for seismic, lood, and reservoir
loading
• Event tree analysis
• Subjective probability and expert elicitation methods
• Fragility curves
• Consequence (loss of life) evaluation
12.8 SYSteMS rISk aSSeSSMent (2005-ongoIng)
A major evolution in the practice of geotechnical risk analysis arrived in 2005 with the
landfall of Hurricane Katrina in New Orleans, Louisiana, and the Delta Risk Management
Program in California. In each case, the cognizant authority elected to employ systematic
and large-scale risk assessments to understand the respective hazards, vulnerabilities, and
consequences of failure of an extensive levee network (350 miles in New Orleans and 1100
miles in California) and to develop risk management plans.
12.8.1 new orleans
A primary contributor to the flooding of New Orleans during Hurricane Katrina was the
failure of levees and floodwalls that make up the hurricane protection system and in some
parts of the system their overtopping by storm surge (the levee system and its various compo-
nents is now called the Hurricane and Storm Damage Risk Reduction System , HSDRRS).
The HSDRRS had been designed to provide protection from storm-induced surges and
waves in an attempt to control naturally occurring conditions. The HSDRRS was designed to
perform this function without imposing unacceptable risks to public safety, property, and wel-
fare; however, some level of risk always remains. Even with the reconstruction and strengthen-
ing of the hurricane protection system (HPS) in the future, still some risk will remain.
The term risk was used to define hazards, losses, and potential outcomes, defined by
Risk = Hazard probability × Vulnerability × Consequences of failure , in which hazard
probability is the rate or uncertainty of occurrence of the causal event; vulnerability is the
reliability with which the constructed system withstands the loads or other demands caused
by the hazard; and consequences of failure are the costs in lives and dollars accruing in the
event of a failure (IPET8 2008). This is sometimes called the hazard-vulnerability-conse-
quence (HVC) model (or, in security risk, TVC, for “threat”).
The application of the probabilistic risk analysis to the HSDRRS of New Orleans was chal-
lenging, because the system is a complex of levees, floodwalls, pumping stations, and other
components that serve a large geographical region and our capability to model hurricanes