Date of this Version
Annals of the Association of American Geographers, 85(2), 1995, pp. 339-355
Research on global change has been hindered by deficiencies in the availability Wand quality of land-cover data (Mather and Sdasyuk 1991; Townshend 1992). To address this deficiency, the U.S. Geological Survey (USGS) and the University of Nebraska- Lincoln have collaborated in developing a method of land-cover characterization that is suitable for research on global change and on regional patterns of land cover (Loveland et al. 1991; Brown et al. 1993). This methodology is based upon statistical analysis of multidate, meteorological satellite imagery acquired by the National Oceanic and Atmospheric Administration's (NOAA) Advanced Very High Reso- lution Radiometer (AVHRR) sensor complemented by ancillary spatial data. The product of this analysis-a multi-level, digital, geographically referenced land-cover database (hereafter referred to as the database) covering the coterminous United States-serves as a prototype for a global land-cover database which is currently under development. The study of global change requires improved regional frameworks (for example, Turner, Moss, and Skole 1993; Mather and Sdasyuk 1991). The land-cover characterization strategy developed in this study is based upon regionalization of the seasonal expression of vegetative development. This approach is well-suited for global-change research because of the explicit manner in which critical biophysical conditions are used to define and characterize land-cover regions. Moreover, the regionalization process presented here has the advantages of replicability, computational manageability, flexibility, and global applicability.