Global land cover classification using ADEOS-II/GLI global mosaic data
2009
Soyama, N.(Tenri Univ. Nara (Japan)) | Tsujimoto, H. | Muramatsu, K. | Furumi, S. | Daigo, M. | Ochiai, F.
This paper describes the algorithm used to produce a global land cover product using data sets from the second Advanced Earth Observing Satellite/Global Imager (ADEOS-II/GLI). Although ADEOS-II was in operation for only a short period, we were able to obtain data sets for about 8 months, from March to October 2003. The aim of our classification system is to produce useful information for monitoring changes in vegetation cover conditions caused by climate changes. To achieve this aim, we determined classification conditions using only the GLI data sets, and constructed a land cover classification system for monitoring truly unique vegetation responses. We started with seven classes based on the International Geosphere-Biosphere Program, and expanded this to 23 classes. Each classification condition was determined using the universal pattern decomposition method (UPDM) coefficients and a modified vegetation index based on the UPDM that was obtained by applying the UPDM to GLI data. Comparisons between information from the FLUXNET Project site and our global land classification system (NWUGLC) at a regional level showed many areas where classes agreed, such as deciduous forest, cropland, and savanna; exceptions were evergreen forest and grassland. One possible reason for the disagreement may be the difference in class definitions between the FLUXNET site and NWUGLC. ADEOS-II/GLI data sets do not contain information relating to surface objects; therefore our system could not distinguish between forest and flat fields covered with thick vegetation. The next JAXA project (second-generation GLI) will provide more useful data that will help in producing accurate global land cover data for the study of global environmental changes.
Show more [+] Less [-]AGROVOC Keywords
Bibliographic information
This bibliographic record has been provided by Agriculture, Forestry and Fisheries Research Information Technology Center