This paper describes various 3D visualization methods applied to a simulation of a hurricane. The data consists of a 3D grid of six variables over a number of time steps. The previous method used to analyze the data was line printer graphics using asterisks and numeric output, which was of limited value due to the large amount of data produced. The visualization at the Ohio Supercomputer Graphics Project (OSGP) brought together the scientist, animation specialists, and computer scientists to produce 3D animations of the data. The initial step is the conversion of the simulation data into a useful format. The data is then mapped using various methods for graphical display. One method is a traditional polygon representation. The polygons are generated from the relative humidity data and displayed as clouds using a scanline polygon renderer. The second method uses orthogonal ray tracing to volumetrically render multiple variables, such as temperature change vs. relative humidity. The final method uses wind velocity data displayed as particles. The resulting animations have proven very useful in analyzing the data. The polygonal animation revealed errors in the simulation data that were previously overlooked. The volumetric rendering showed relationships between different variables. The particle animation clearly indicated the swirling patterns and the anti-cyclonic outflow generated by the hurricane, a key indicator that the simulation is producing accurate data.
Robert E. Marshall,
Peter G. Carswell,
"Alternative Views of a Hurricane", Proc. SPIE 1083, Three-Dimensional Visualization and Display Technologies, (11 September 1989); doi: 10.1117/12.952899; https://doi.org/10.1117/12.952899