Informations et ressources scientifiques
sur le développement des zones arides et semi-arides

Accueil du site → Doctorat → Royaume-Uni → 1990 → Vegetation indices and the measurement of vegetation growth in semi-arid west African grasslands using satellite imagery

University of London (1990)

Vegetation indices and the measurement of vegetation growth in semi-arid west African grasslands using satellite imagery

Hanan, N.P

Titre : Vegetation indices and the measurement of vegetation growth in semi-arid west African grasslands using satellite imagery

Auteur : Hanan, N.P

Université de soutenance : University of London

Grade : Doctor of Philosophy (PhD) 1990

Résumé
Primary production of Sahelian grasslands can be monitored using satellite measurements of reflected red and near-infrared radiation combined in a vegetation index. This is demonstrated for northern Senegal using data from the Advanced Very High Resolution Radiometer. Atmospheric conditions affect the satellite imagery. Differences in vegetation types may affect the shoot-root ratio, causing bias in estimates of production made using above ground standing crop. The derived rangeland production map was used to estimate productivity near deep wells in the study area. No evidence was found of a reduction in total production due to overgrazing. Realisation of the full potential of satellite remote sensing in the Sahel requires an understanding of how features of the landscape influence the spatial average reflectance. A model is presented to describe the Normalised Difference Vegetation Index (NDVI) in environments where spatially separate soil and vegetation components occur. The increase in NDVI is non-linear with fractional cover of the vegetation, the extent of non-linearity depending on the brightness (sum of red and near-infrared reflectances) of the soil. This model can be inverted to obtain the NDVI of the vegetation component if information on soil spectral properties and fractional cover of the vegetation are available. Application in the Sahel was limited by variability within the vegetation components. Trees growing over bare soil reduce the spatial average red reflectance but have little effect on near-infrared reflectance. Over green herbaceous vegetation, red reflectance is less sensitive to trees but still reduced. The decrease in red reflectance causes an increase in spatial average NDVI. Substantial agreement in these results was found using radiometer measurements with a geometric model and hemispherical photographs with a radiative transfer model. A method is given which combines radiometer and hemispherical photograph data, to estimate spectral reflection and transmission by canopy material.

Proquest Dissertations & Theses

Page publiée le 24 février 2015, mise à jour le 25 novembre 2019