Despite major improvements over the past several decades in computer methods for analyzing and synthesizing exploration and site investigation geological data for engineering purposes, which should have decreased project risk, unexpected failures and construction problems still occur. Often, this unexpectedness can be tied back to inadequate structural geological understanding of actual site-specific conditions. Part of this diminished appreciation of structural geological controls on rockmass behaviour can likely be attributed to inappropriate use of computational geostatistics, block modelling and geomechanics computer codes, without significant structural geological input. Part may be because of inadequate domaining in the raw data analysis phase, such that no differentiation was ever made between uncertainties due to true natural variability versus those due to data inadequacy. Modelled data error issues cannot easily be resolved once mixing and smearing of parameters have occurred as a result of earlier inappropriate synthesis of an uncontrolled database. The increased use, industry-wide, of inadequately calibrated, often totally unverified, geomechanics computer modelling for design of civil or mining projects is a worrying trend that needs reversing. Much more attention needs to be paid to properly sub-dividing project sites into realistic geological entities through use of rigorous geological structural domaining techniques, such that much better understanding is gained of project-specific geological risk. This in turn will lead to improved reliability and representativeness in project geomechanical parameter selection for use in design. This paper attempts to set out some guidelines to help practitioners undertake more rigorous domaining, as an aid to parameter selection for design.
