Visualizing sparse gridded data sets

Gridded data sets with many missing values pose a problem because most visualization algorithms fail when presented with incomplete cells. We discuss visualization methods that handle this problem. Our primary interest is developing 3D images for Next-Generation Radar (Nexrad), a weather radar that makes a series of conical scans. Most of the time it has extremely sparse returns. Current visualization techniques for Nexrad simply discard the 3D nature of the data set and provide 2D plots of the grid's lowest layer, leaving the missing data colored black or transparent. Most standard visualization packages either fail or give incorrect visualizations in 3D because of the unusual nature of the data set.