| Abstract: | This paper describes the physical and mathematical approach followed to design a vegetation index optimized for the Medium Resolution Imaging Spectrometer (MERIS) sensor, i.e. the MERIS Global Vegetation Index (MGVI). It complements an earlier feasibility study presented elsewhere in this issue by Govaerts and collaborators. Specifically, the crucial issue of the dependency of the vegetation index on changes in illumination and observing geometries is addressed, together with the atmospheric contamination problem. The derivation of the optimal MGVI index formulae allows a comparison of its performance with that of the widely used Normalized Difference Vegetation Index (NDVI), both from a theoretical and an experimental point of view. Data collected by the MOS/IRS-P3 instrument since March 1996 in spectral bands analogous to those that will be available from MERIS can be used to evaluate the MVGI. |
