Mapping European land cover with NOAA-AVHRR data
This study describes a methodology to integrate low and high spatial resolution satellite data selected from the Monitoring Agriculture by Remote Sensing (MARS) data archives with Geographic Information System (GIS) and statistical information for mapping European land cover at a one to one million scale. The methodology follows recent findings which demonstrate that the combined use of surface temperature and spectral vegetation indices can successfully discriminate regional land cover classes. A six month time series of surface temperature data and the normalized difference vegetation indices can successfully discriminate regional land cover classes. A six month time series of surface temperature data and the normalized different vegetation index is thus extracted from Advanced Very High Resolution Radiometer - Local Area Coverage (AVHRR-LAC) data and input into an unsupervised classification scheme. The AVHRR data are stratified into ecosystem regions according to the Forest Information from Remote Sensing (FIRS) regionalization, in order to reduce the impact of climatic gradients and the phasing of seasons across Europe. Unsupervised classifications are performed independently in each ecosystem region. Post classification labelling is performed in each region using preclassified high spatial resolution MARS test image data. The classification results are validated with independent European Statistical Office (EUROSTAT) land cover statistics for western and central Europe on a Nomenclature of Territorial Units for statistics (NUTS)-2 level.
Bibliographic Reference: EUR 16464 EN (1996) 40pp., FS, free of charge
Availability: Available from the Public Relations and Publications Unit, JRC Ispra, I-21020 Ispra (IT), Fax: +39-332-785818
Record Number: 199710078 / Last updated on: 1997-02-28
Original language: en
Available languages: en