Vegetation assessment in East Africa using MGVI and Red Edge Position from ENVISAT MERIS data

摘要

The actual condition of vegetation cover is represented by its spectral signature. Photosynthetic active vegetation is characterized by a low reflectance in the red and a high reflectance in the near infrared spectra. This specific feature is used to calculate different vegetation indices. One index is the MERIS Global Vegetation Index (MGVI), the other one is the Red Edge Position (REP), which represents the point of maximum slope of reflectance of green vegetation between 670 nm and 780 nm. With a spectral resolution of 15 bands and a narrow band setting in the red and near infrared spectra ENVISAT MERIS seems particularly suitable for the Red Edge Position or the MERIS Terrestrial Chlorophyll Index (MTCI), which is a surrogate of REP and designed for a better correlation with canopy chlorophyll. In this study the variation of the MGVI, REP and MTCI is analyzed in comparison with seasonal rainfall pattern and their suitability to characterize different vegetation types (forest, woody savanna, grassland, agricultural fields). A maximum likelihood classification was performed for the vegetation classes using different multitemporal vegetation index data sets and the resulting accuracies were compared. The results show a) very high correlation between REP and MTCI, b) generally, a relative high correlation between MGVI and either REP or MTCI, c) a similar seasonal variation of REP, MTCI and MGVI for most classes, d) a better discrimination for high chlorophyll classes using REP and a better discrimination for low chlorophyll classes using MGVI and e) a high potential for ecosystem monitoring using a combination of the different indices.