CIESIN Thematic Guides The global-change land-sciences research community is increasingly stressing the need for repetitive, globally consistent, remotely sensed land-cover data of sufficiently high spatial resolution to enable comprehensive but detailed analysis of land cover. In Improved Global Data for Land Applications, Townshend (1992) explains that the spatial resolution must allow for observing landscape features within dominant vegetation types that, though local in scale, are critically important for global change studies involving biomass distribution, biogeochemical cycling, and the hydrological cycle. Such features include wetlands, bogs, flood plains, evidence of biomass burning, ecosystem transition gradients and boundaries, and evidence of land-cover change. Repeated observations enable monitoring the dynamics of seasonal variations in the biophysical processes of land cover and changes in the amounts and distribution of land cover.
These requirements have prompted the specification for and initiated the development of regional and global land data sets of 1-km spatial resolution using the full-resolution National Oceanic and Atmospheric Administration (NOAA) Advanced Very High Resolution Radiometer (AVHRR) data of 1.1 km. These data provide a compromise between the need for high ground resolution and the large volumes of data associated with detailed monitoring activities. While the high resolution Landsat and Systeme Probatoire d'Observation de la Terra (SPOT) sensors are more appropriate for detailed analysis of land-cover and land-cover change, their lack of daily re-visit capabilities, relatively limited ground coverage, and the frequent occurrence of cloud cover make highly repetitive, comprehensive coverage for detailed monitoring unattainable.
The daily, synoptic coverage of the AVHRR sensor generally includes a significant amount of cloud cover, which precludes extensive analysis of land cover from the acquired data for any one day. To overcome this obstacle, procedures have been developed to produce maximum Normalized Difference Vegetation Index (NDVI) composited data sets. In these data sets, pixels have been selected from daily AVHRR scenes for a specified number of days to minimize cloud cover over extensive regions. Such a composited data set is more conducive to extensive analysis of land cover. Generally, AVHRR data are composited for a succession of time periods throughout a growing season or a year to create a time series of mostly cloud-free data sets.