Risk Analysis for Dam Design in Karst

Probabilistic risk analysis is a useful technique in assessment of dam safety issues, and also has important applications in the early phases of dam design when design options are being compared and exploration strategies planned. At this stage, few data are usually available and judgment‐based decisions are most important. An application of probabilistic risk analysis to design of a dike in karst terrain is described. Assessments based on geologic information, observations, and subsurface investigations are expressed using a simple probability model. Options considered to reduce the likelihood or consequences of sinkhole‐induced failure include a secondary containment structure, a warning system, and additional exploration. These options are evaluated using an expected monetary value criterion. It is shown that simple probabilistic formulations consistent with preliminary levels of information reduced major sinkhole uncertainties into more readily evaluated components. These techniques identified the most cost‐effective design strategy for reducing the risk of sinkhole‐induced dike failure and provided an effective vehicle for communicating geotechnical judgments, risks, and alternatives to the project owner.