Paper No. 6
Presentation Time: 10:15 AM

SUGGESTED APPROACH FOR QUANTIFYING UNCERTAINTY AND VARIABILITY IN GEOLOGY FOR RELIABILITY-BASED DESIGN PROJECTS


KEATON, Jeffrey R., AMEC Environment & Infrastructure, Inc, 6001 Rickenbacker Rd, Los Angeles, CA 90040, Jeff.Keaton@amec.com

Geology must be expressed explicitly in terms that describe variability to have value in reliability-based design projects. Geologists use solid, dashed, dotted, and queried lines to convey confidence in interpretation of locations and nature of formation boundaries and faults. The lines have location error related to map scale, terrain, vegetation, and field methods, and uncertainty related to the competency of the geologist and the allotted field mapping time. A suggested nine-component complexity rating scheme for quantifying geologic conditions has four components related to regional-scale geologic complexity (1. genetic deposition or emplacement, 2. epigenetic deformation, 3. epigenetic alteration, and 4. epigenetic weathering and erosion). The remaining components pertain to (5) site-scale geologic complexity, (6) terrain, (7) information quality, (8) geologist’s competency, and (9) time used to prepare the model.

The suggested four-level rating criteria and scores are similar to rock fall hazard rating systems. Simple and uniform conditions can be predicted with confidence (score = 3), whereas complex and nonuniform conditions cannot be predicted (score = 81); intermediate conditions have scores = 9 or 27. Possible nine-category cumulative scores range from 27 to 729. Cumulative complexity scores could be expressed in a form suitable for reliability-based design projects by accepting the highest possible cumulative score as the mean value and defining the coefficient of variation (COV = standard deviation/mean) of the highest possible score as 1.000. The standard deviation would be accepted as the actual cumulative score. Thus, the COV of the lowest possible score would be 0.037 (=27/729). The COV of the interpreted geologic model could be applied to the geotechnical parameters that are used to convert the geologic model into a ground model to support the geotechnical model.