Examining the accuracy of three-dimensional isosurfaces as an analysis or operational forecast tool

As graphics processing has become more robust and affordable, the ability to examine three-dimensional isosurfaces of any particular meteorological field has evolved from an offline illustration exercise to a realtime analysis product. Common and popular platforms that employ isosurfacing include the Gibson Ridge "GR Analyst", the Unidata Integrated Data Viewer, Matlab, Vis5D/VisAD, and WDSSII. Displaying isosurfaces of model parameters or radar reflectivites allows a user to visualize and more easily understand the three dimensional structure of that particular feature as opposed to attempting to mentally construct the 3D model from two-dimensional planes, and provides a common model for multiple inspectors to examine and collaborate.

Now that the ability to construct such displays as quickly as data become available, a question needs to be asked about what the user is actually looking at, and how precisely can a 3D feature be relied upon? How much is an isosurface resembling the atmosphere, the observing platform, or even the objective analysis scheme used to create the isosurface representation? In this presentation we explore some potential pitfalls for those who use three-dimensional isosurfacing in their analyses.