A Method of Visualising Uncertainty in Three Dimensional Digital Ground Models

It is now commonplace to represent subsurface conditions using digital three dimensional representations, or digital ground models. This facilitates the integration of ground attributes with built elements in a digital environment; a technique which brings many benefits. However, doing this usually involves marrying built elements the locations and form of which are known with certainty with a ground model; the elements of which usually have a much lower certainty. In the authors’ experience, this can create problems whereby users of the ground model, for example civil designers, perceive and assume the elements provided in the digital ground model to be represented with a certainty as high as that of the built elements. This is almost never the case. It is essential that the level of certainty or ‘reliability’ of three dimensional ground models be communicated, to anyone who might use the model, and is ideally an intrinsic part of any digital ground model. This paper describes a systematic way of visually representing the reliability of three dimensional digital ground models by producing a digital ‘twin’ of the ground model, which zones the ground based on its uncertainty or the degree of reliance that should be placed upon it. This model, termed here the ‘reliability model’, is a three dimensional digital zoning of the ground based on a systematic assessment of the quality of the information which informs the ground model, a necessarily subjective assessment of the geological complexity, and the susceptibility of the project to ground uncertainty. A clear, three dimensional visual representation of ground model uncertainty can be married with existing or proposed built elements to allow assessment of where the interaction of the ground with structure is uncertain and where prediction of ground response is likely unreliable. This informs the assessment of geotechnical risk and the optimization of ground investigation. The methods and algorithms used to develop the digital reliability model are set out in this paper along with examples of digital reliability ground models.