The UK Low Level Waste Repository Ltd (LLWR) submitted an Environmental Safety Case (ESC) for the disposal of low-level waste (LLW) to the UK Environment Agency on the 1 st of May 2011. As part of the ESC, the LLWR have to demonstrate that a programme of site investigation and site characterisation has been carried out to provide the requisite information for the ESC and support facility design and construction. This paper explains the development of the site investigation programme and how the understanding of the geology of the site has developed. The geological environment in the region of the LLWR consists of Quaternary age deposits overlying older bedrock. The site has been subjected to a series of site investigation programmes from 1939 to the present day. The development of 3-D geological models was necessary to integrate data from boreholes, trial pits, geophysical investigations and beach exposures and data gained from site operations. The understanding of the geology has developed with each new set of data. Early simple interpretations from a few boreholes have been superseded by a series of more complex interpretations each incorporating the increasingly detailed observations. Initial attempts to develop a lithostratigraphic representation of the geology proved difficult. It was also difficult provide a clear link between the geology and the hydrogeology using a lithostratigraphic approach as required for the development of hydrogeological models. A lithofacies approach to characterising the geology was adopted in 2007, which has allowed the grouping of geological units with similar hydraulic properties and the development of a regional 3-D geological model. The 3-D geological model has been used as the framework for the development of a hydrogeological model for the site. The development of the 3-D geological models has been iterative. It was observed that there are differences between models developed using solely mathematical interpolation and those controlled by geological interpretation. The different representations of the geological information have been used to consider the effects of uncertainty in the geological interpretation in the hydrogeological modelling.