This thesis examines the application of scientific visualisation to the analysis of polarimetric radar data sets. The research contained herein forms part of a larger body of work that studies the application of scientific visualisation to the analysis of large multi-valued datasets. Visualisation techniques have historically assumed a fundamental role in the analysis of patterns in geographic datasets. This is particularly apparent in the analysis of remotely sensed data, which, since the advent of aerial photography, has utilised the intensity of visible (and invisible) electromagnetic energy as a means of producing synoptic map-like images. Progress in remote sensing technology, however, has led to the development of systems which measure very large numbers of intensity 'channels', or require the analysis of variables other than intensity values. Current visualisation strategies are insufficient to adequately represent such datasets, whilst retaining the synoptic perspective. In response to this, two new visualisation techniques are presented for the analysis of polarimetric radar data. Both techniques demonstrate how it is possible to produce synoptic image suitable for the analysis of spatial patterns without relying on pixel based intensity images. This allows a large number of variables to be ascribed to a single geographic location, and thus encourages the rapid identification of patterns and anomalies within datasets. The value of applying the principals of scientific visualisation to exploratory data analysis is subsequently demonstrated with reference to a number of case studies that highlight the potential of the newly developed techniques.