The Normalized Difference Vegetation Index

The usage of vegetation indices such as the Normalized Difference Vegetation Index (NDVI) calculated by means of remote sensing data is widely spread for describing vegetation status on large space scale. However, a big limi-tation of these indices is their inadequate time resolution for agricultural purposes. This limitation could be over-come by the ground-based vegetation indices that could provide an interesting tool for integrating satellite-based value. In this work, three techniques to calculate the ground-NDVI have been evaluated for sugar beet cultivated in South Italy in all its phenological phases: the NDVI1 based on hand made reflectance measurements, the NDVI2 calculated on automatically reflectance measurements and the broadband NDVIb based on Photosynthetically Ac-tive Radiation (PAR) and global radiation measurements. The best performance was obtained by the NDVIb. More-over, crop-microclimate-NDVI relations were investigated. In particular, the relationship between NDVI and the Leaf Area Index (LAI) was found logarithmic with a saturation of NDVI at LAI around 1.5 m2 m-2. A clear rela-tion was found between NDVI and crop coefficient Kc experimentally determined by the ratio between actual and reference measured or modelled evapotranspirations, while the relation between NDVI and crop actual evapotran-spiration was very weak and not usable for practical purposes. Lastly, no relationship between the microclimate and the NDVI was found. Key-words: actual evapotranspiration, broadband NDVI, eddy covariance, sugar beet