Scientists routinely accomplish small-scale geospatial modeling in the raster domain, using high-resolution datasets for large parts of continents and low-resolution to high-resolution datasets for the entire globe. Direct implementation of point-to-point transformation with appropriate functions yields the variety of projections available in commercial software packages, but implementation with data other than points requires specific adaptation of the transformation equations or prior preparation of the data to allow the transformation to succeed. It seems that some of these packages use the General Cartographic Transformation Package (GCTP; Elassal, 1987) or similar point transformations without adaptation to the specific characteristics of raster data (Usery and others, 2003a).|
Usery and others (2003b) compiled and tabulated the accuracy of categorical areas in projected raster datasets of global extent. Based on the shortcomings identified in these studies, geographers and applications programmers at the U.S. Geological Survey (USGS) expanded and evolved a USGS software package, mapIMG, for raster map projection transformation (Finn and Trent, 2004). Daniel R. Steinwand of the USGS? Earth Resources Observation and Science Center, originally developed mapIMG, basing it on GCTP (Usery and others, 2001). Through previous and continuing efforts at the USGS, this program has been transformed from an application based on command line input into a software package based on a graphical user interface for Windows, Linux, and other Unix (officially trademarked as UNIX) machines.
[Summary provided by the U.S. Geological Survey.]