717
THE ANGULAR VEGETATION INDEX: AN ATMOSPHERICALLY
RESISTANT INDEX FOR THE SECOND ALONG TRACK SCANNING
RADIOMETER (ATSR-2)
S. E. PLUMMER, P. R. NORTH AND S. A. BRIGGS
British National Space Centre,
Remote Sensing Applications Development Unit,
Monks Wood, Abbots Ripton,
Cambridgeshire, PE 17 2LS
UK
Tel: 44 487 3381 Fax: 44 487 3277
ABSTRACT:
A vegetation index which is resistant to atmospheric scattering, the Angular Vegetation Index (AVI),
is defined. This index has been developed for the three optical channels which will be provided by the
second Along Track Scanning Radiometer (ATSR-2) although it is similarly applicable to any sensor
with green, red and near infrared wavebands. A sensitivity analysis is performed using a coupled soil-
vegetation-atmosphere model which combines a soil model, the PROSPECT leaf model, SAIL canopy
model with hot-spot and the Lowtran-7 atmospheric package. Preliminary results for a canopy with a
spherical leaf angle distribution, for one solar position and at the nadir view indicate that the AVI is
resistant to the effects of variation in atmospheric optical depth and aerosol size distribution. Based on
example light and dark soils AVI is also unaffected by soil brightness. Variation in leaf chlorophyll
concentration does, however, affect AVI but only beyond an LAI of 2.
KEY WORDS: Atmospheric resistance, ATSR-2, vegetation index, simulation
1. INTRODUCTION
This paper presents preliminary work on the development and testing of a new vegetation index for use
with data from the second Along Track Scanning Radiometer (ATSR-2). The ATSR-2 instrument will
be launched on ERS-2 in 1995 as a follow-on to the ATSR sensor on ERS-1. It will match the
characteristics of ATSR but will also carry three experimental optical channels centred at 555, 659 and
865 nm with a bandwidth of 20 nm. Additionally the ATSR-2 provides a dual look capability along
track using a conical scan mechanism with a spatial resolution of 1 km at nadir and 2x3 km for the 55'
forward view. The three channels are positioned to maximise the difference between vegetation and
background but also to avoid major atmospheric absorptions. These optical data will offer an
alternative to the current Advanced Very High Resolution Radiometer (AVHRR), for regional and
global vegetation monitoring.
The Normalised Difference Vegetation Index (NDVI), which is a combination of the of measurements
in the near-infrared and red wavelengths, provides a measure of changes in vegetation. However, when
calculated from top-of-the-atmosphere (TOA) radiances NDVI is highly sensitive to variation in
optical thickness and when calculated from measurements made with the AVHRR sensitive to
absorption by columnar water vapour. The dominant atmospheric effect, however, is due to
atmospheric scattering by atmospheric aerosols and gases. Further advances in quantitative remote
sensing of vegetation, therefore, require the development of either accurate atmospheric correction or
redefinition of NDVI such that it is insensitive to atmospheric effects.
Recently a number of authors (Kaufman and Tanrd 1992, Pinty and Verstraete 1992, Phil pot 1991)
have suggested indices to either redefine the NDVI to make it less sensitive to atmospheric effects or
provide alternatives to NDVI. The proposal by Philpot (1991) requires sufficient wavebands to
generate a second derivative for each waveband used to develop the ratio algorithm. The examples
given use high spectral resolution data rather than that available from ATSR-2. The Pinty and