il 2004
Sm | Bio-geophysical land surface parameters derived from ADEOS-2/POLDER-2 multi-angular
693. measurements
nn R. Lacaze*', F. Maignan®, B. Miras!, F. M. Bréon?, M. Weiss), F. Baret*
nt. J.
: : Médias-France, CNES-BPI 2102, 18 av. E. Belin, 31401 Toulouse Cedex 9, France — (lacaze, miras )(a)medias.cnes.fr
* Laboratoire des Sciences du Climat et de l'Environnement, CEA, 91191 Gif sur Yvette, France — maignan(álsce.saclay.cea.fr
Univ. fmb@cea.fr
* Noveltis, Parc Technologique du Canal, 2 av. de l’Europe, 31520 Ramonville, France
ues for |
| ^ INRA/CSE Agroparc, 84014 Avignon, France (baret, weiss)@avignon.inra.fr
KEY WORDS : Ecosystem, Environment, Vegetation, Monitoring, Retrieval, Global, Parameters
ABSTRACT:
Monitoring of terrestrial vegetation from satellites at global and regional scales requires accurate and frequent measurements of
| surface reflectance. In this context, the POLDER instrument leads a key improvement providing, at high temporal resolution,
measurements of the Bi-directional Reflectance Distribution Function corrected for atmospheric effects. In the frame of the ADEOS-
| 2/POLDER-2 project, advanced Land Surface Level 3 algorithms have been developed to retrieve the spectral Directional-
| Hemispherical Reflectances (DHR), the Normalized Difference Vegetation Index (NDVI) corrected for directional effects, the Leaf
Area Index (LAI) and the Fraction of Vegetation Cover (FVC). The retrieved parameters are validated following a procedure based
| upon the analyze of the spatial and temporal variability, the comparison with available equivalent products derived from other
sensors by similar or different approaches, and finally the comparison with up-scaled in-situ measurements collected over selected
| areas.
|
RESUME:
Le suivi de la végétation continentale à l'échelle global et régionale demande des mesures fréquentes et précises de la réflectance de
surface. Dans ce contexte, l'instrument POLDER apporte une information inédite en fournissant des mesures journaliéres de la
Fonction de Distribution de la Réflectance Bi-directionnelle (FDRB) corrigées des effets atmosphériques. Dans le cadre du projet
ADEOS-2/POLDER-2, des algorithmes améliorés de restitution des paramétres biophysiques des surfaces continentales ont été
développés. Ils estiment les réflectances hémisphériques-directionnelles, l'indice de végétation NDVI corrigé des effets directionnels,
l'indice foliaire LAI et la fraction de couverture végétale FVC. Les paramétres ainsi restitués sont ensuite validés selon une
procédure qui repose sur l'analyse de leur variabilité spatiale et temporelle, la comparaison à des produits équivalents issus d'autres
capteurs par des méthodes similaires ou différentes, et enfin la comparaison à des mesures terrains spatialisées acquises sur des sites
sélectionnés.
| 1. INTRODUCTION term observations of land surface at appropriate spatial and
temporal scale.
| The regular increase in greenhouse gases due to anthropogenic
| emissions in the atmosphere may have a major impact on the The advent of the POLDER sensor providing multiple spectral
images at high angular resolutions has lead to investigate the
| improvement of land surface parameters retrieval using multi-
angular measurements. The POLDER radiometer has been
| designed to measure the directional and polarized reflectances
| of the surface-atmosphere system (Deschamps et al., 1994). The
instrument concept consists in a camera composed of a two-
| dimensional CCD detector array, wide field of view telecentric
optics and a rotating wheel carrying spectral and polarized
filters. During a single satellite overpass, a surface target is
Earth' s climate in the forthcoming decades. In order to reduce
the uncertainties in forecasting climatic changes, it is necessary
to better understand the processes involved in interactions
between land surface, aerosols, clouds, radiation and
atmospheric circulation.
The exchanges between the land surface and the atmosphere are
| largely controlled by the physiological characteristics of the
surface, which determine energy, water and momentum fluxes.
In climatic and weather forecasting models, surface processes
are represented by soil-vegetation-atmosphere transfer scheme,
which simulate the fluxes at the atmospheric boundary layer
(Sato et al., 1989; Calvet et al., 1998). This imposes strict
requirements in order to implement global and regional
description of soil and vegetation properties which control
physical, biological, and hydrological surface processes. Even
though climatic and meteorological models are based on large
grid cells, a good knowledge of the subgrid landscape
variability is necessary to accurately account for surface fluxes
of radiation, heat, moisture, and momentum (Henderson-sellers
and Pitman, 1992; Noilhan et Lacarrére, 1995). In this context,
space-borne sensors play a key role by providing global, long
scanned up to 14 times under different viewing angles. The
view illumination directional configuration changes each day as
the orbit shifts. Therefore, after a few days, assuming clear
atmospheric conditions, the measurements provide a sample of
the Bi-directional Reflectance Distribution Function (BRDF)
within the sensor field of view.
The POLDER-1 sensor onboard the Japanese satellite platform
ADEOS-1 has measured the reflectance of the entire land-
atmosphere terrestrial system from November, 1996 to June,
1997. Its successor, POLDER-2, has been launched on the
ADEOS-2 platform in December, 2002. It provided BRDF
measurements from April to October, 2003.