The Gateway Upgrade Project in the State of Queensland, Australia, is the largest road and bridge infrastructure project in the state's history. A statistical analysis of intact rock strength properties of the sub-horizontally interbedded sandstone stratum underlying the main river span of the duplication bridge was carried out using pre-construction geotechnical data. These findings were then compared with an analysis based on additional borehole data captured during construction. However, this has been confined to the additional data obtained at the southern pier location of the main span only for this paper. Having a borehole at each socket location at this pier showed that signification variation of ground conditions could occur locally between piles, even for the same pier location.

Given the data variation, an appropriate statistical density function is required for statistical modeling to asses the reliability of the design. Using the probabilistic models identified, the study undertakes to rationalize the design rock strength input model adopted for socket design. The impact of rock strength anisotropy on the design UCS is also investigated. The implications of using the design UCS with various probability distribution models to satisfy limit state material characterization requirements are briefly discussed. Assumption of normality in the data distribution was shown to have the potential to significantly affect the design values.