Mapping Land Cover in Northern Eurasia Using a Hierarchical Land Cover Classification System
Damien
Sulla-Menashe, Boston University, dsm@bu.edu
(Presenting)
Alessandro
Baccini, Woods Hole Research Center, abaccini@whrc.org
Mark
Friedl, Boston University, friedl@bu.edu
Curtis
Woodcock, Boston University, curtis@bu.edu
Olga
Krankina, Oregon State, olga.krankina@oregonstate.edu
Northern Eurasia encompasses a vast landmass with tremendous land cover and ecological diversity and importance. We describe an approach to mapping land cover in this region using a classification system based on the FAO-LCCS model. Specifically, we have implemented a hierarchical classification strategy where each pixel is classified using a set of nested classes. In this framework, each level in the classification hierarchy is based on vegetation structure, leaf type, or phenology. In addition, the system explicitly distinguishes land cover from land use, which is mapped as a separate attribute. To perform the mapping we use two MODIS collection 5 products: (1) the 500-m nadir BRDF-adjusted reflectance (NBAR) product, and (2) the 1-km MODIS Land Surface Temperature (LST) product, both measured at 8-day intervals. As part of the classification strategy we merge remote-sensing classification results with information derived from potential vegetation modeling. Our initial results indicate that this hierarchical methodology leads to a successful classification result. More generally, the approach we have taken provides a methodology that can be used for large-scale land cover mapping tasks outside of Northern Eurasia.
NASA Carbon Cycle & Ecosystems Active Awards Represented by this Poster: