Terrain analysis.

Research on representation and computation of terrains has been an active area.

• TIN: Triangulated surfaces were suggested in GIS in 1978, and dubbed TINs or Triangulated Irregular Networks [119]. Delaunay triangulations are often suggested (and used) as good TINs. Other (data-dependent) triangulations may sometimes be more appropriate.
• SIMPLIFICATION: Kumler's monograph [86] compares elevation accuracy of TINs to DEMs and finds that, byte for byte, Digital Elevation Models are better. The simplification methods he uses, however, are probably not the best possible. (Also, error measure that reflect curvature or symmetric volume difference should be considered.) Simplifications with some sort of hierarchical structure are especially useful. See Heckbert and Garland's survey [68] for many citations to geometry and graphics papers on surface simplification methods.
• DRAINAGE: For many natural resource management tasks, the watershed-or region that drains to the same point-is the natural management unit. Much of the work on computing drainage networks is raster based; some geometers are looking into preprocessing TINs for drainage queries.
• VIEWSHEDS: Some forestry operations are now required to preserve the visual impact of the forest. Thus, ``viewshed'' algorithms become necessary. Location problems for microwave and cell phone relays also involve line-of-sight constraints imposed by the terrain.
• FLYBY: Fast graphics engines give the capability to fly-by data models to get a better feeling for the actual terrain (eg, systems such as PCI's EASY/PACE).