As technological advances give rise to increasingly more powerful computers, applied mathematicians are able to simulate phenomena of considerable complexity using nurmerical models. One of the problems is how to transform the volminuous output data from a numerical form to a visual one, which is easier for humans to comprehend. This article analyzes a case study from the field of meteorology. A weather simulation, program produces numerical output which defines the values of several variables as functions of the three spatial parameters and time. The output of this model is then used as the input to our grphic rendering system, which displays these complex data sets using stereo animation, employing novel techniques to add realism in the animation sequences. We take advantages of the special geometrical configuration to reduce the computation time by about two orders of magnitude. The value our approach is that our techniques for representing four-dimensional data sets can easily be extended to display phenomena in other disciplines.