Coordinating methodologies for scaling landcover classifications from site-specific to global: steps toward validating global map products

The MODIS sensor to be launched on the EOS-AMplatform will be the most important sensor for globalvegetation mapping. Among the programmatic goals forthe MODIS sensor are to assess and track changes inland use/landcover, leaf area index (LAI), and net pri-mary productivity (NPP). For these products to be usedin global models, they must be rigorously validated withsite-specific data products. This article presents a reviewof some of the problems facing a regional- to global-scalevalidation effort and presents strategies for coordinatingthe land-cover classification process across multiple sites.We suggest the Enhanced Thematic Mapper (ETM+) asthe source of remotely sensed data for validation, andthat the IGBP 17-class land-cover classification system beused to provide a link between more complex site-specificsystems and global-scale data products. We further rec-ommend that the best site-specific land-cover classifica-tions be obtained, using whatever ancillary data arefound to be useful, as a basis for validation. In addition,we propose ways in which ambiguities in translation ofclasses, from specific to general systems, may be identi-fied. Finally, we stress that even though standardizationof methodology among sites may not be appropriate tothe goal of obtaining the best possible land-cover products, there should be standardization of error analysisand metadata reporting. ©Elsevier Science Inc., 1999