Bilan du suivi des sites
aerosol optical depth in
ing, 12(6): 1147-1163.
oronment, 25:295-309.
pectra, Remote Sensing
ctional soil reflectance,
i using NOAA AVHRR
: normalized difference
mal of Remote Sensing,
i the daylight spectrum.
Ecological Studies 75,
:lian grassland model to
mg of Environment).
Optical properties of the
mal of applied ecology,
;lian grassland model to
ing of Environment).
ierivation of vegetation
-403.
predicting bidirectional
resolution satellite data,
n of results for sahelian
ric correction of satellite
l
végétale d'une steppe à
i génieur, Université des
reduction végétale d'une
7.
el for albedo or surface
nsing, 11(8): 1495-1499.
lating the soil-vegetation
, ll(3):447-459.
spheric effects on NOAA
13.
>eschamps, P.Y. (1990),
rtrtun: the 5S code, Int.
irith for NOAA-AVHRR
Sensing, 30(2):231-248.
nsing of total dry matter
note sensing of total dry
61-474.
eflectance modeling: the
COMPARISON AT GLOBAL SCALE OF C02 CONCENTRATION
MEASUREMENTS WITH RESULT OF A TERRESTRIAL BIOSPHERE MODEL
DRIVEN BY REMOTE SENSING:
SOME TEACHINGS ABOUT THE REQUIRED QUALITY OF SATELLITE DATA.
By P.MAISONGRANDE 1 , A.RUIMY 2 , G.DEDIEU 1 and B.SAUGIER 2
1 Laboratoire dEtudes et de Recherches en Télédétection Spatiale,
Unité mixte CNES-CNRS, 18 avenue Edouard Belin, 31055 Toulouse Cedex, France
2 Laboratoire dEcologie Végétale, Université de Paris Sud-C.N.R.S., Bât 362, 91405 Orsay, France.
ABSTRACT
The terrestrial biosphere acts in the global carbon cycle through the process of Net Ecosystem Productivity
(NEP) which is the balance of vegetation Net Primary Productivity (NPP) and heterotrophic decomposition.
The model of NPP is based on a parametric model derived from the Monteith's approach (1977), where PAR
absorption coefficient is derived from the Normalized Difference Vegetation Index ( NDV1 ). This preliminary
study is focused on the sensitivity of model results to the quality of remotely sensed data. It is especially
important to distinguish NPP (or NEP) temporal variations resulting from the actual changes of the vegetation
and artefacts due to satellite measurements. As far as this distinction is possible, we also tried to quantify the
impact of atmospheric corrections on NDVI, NPP and NEP. This approach allows us to estimate the influence
of atmospheric decontamination when comparing cumulated NEP fluxes with regard to the in situ atmospheric
C02 concentration measurements.
KEYWORDS : Primary productivity, Satellite, Atmospheric correction, Vegetation.
1- INTRODUCTION
In this paper, we present global estimates of Gross Primary Productivity (.GPP), NPP and NEP of terrestrial
ecosystem as derived from a diagnostic model driven by satellite measurements. We specifically address the
impact of satellite data quality on seasonal and interannual productivity estimates. Our ability to accurately
monitor NEP globally and for long time period is of paramount importance to better understand the role of
terrestrial ecosystems in the global carbon cycle and its possible evolution.
Net Ecosystem Productivity is the balance of NPP and heterotrophic decomposition (soil
respiration, SR). NPP, the rate of carbon uptake by the vegetation, is the balance between GPP and autotrophic
respiration. These processes are driven by climatic variables such as solar radiation, temperature, precipitation.
The model for GPP, NPP and NEP that we use in this analysis is based on a parametric model derived from
Monteith's NPP model (1972, 1977), and K umar and Monteith's (1981) model for remote sensing of crop
growth : NPP is computed as the product of incident solar radiation by various efficiencies, the efficiency of
solar radiation absorption being related to a satellite-derived vegetation index. It has been first used on a global
scale by Heimann and Keeling (1989). The main advantage of this approach is that the use of satellite-derived
vegetation index impose strong constraints on the temporal and spatial evolution of NPP estimates.
We applied Monteith's model to GPP and we computed autotrophic respiration independently.
Indeed, plant GPP and respiration are primarily driven by different climatic variables : GPP depends
essentially on solar radiation, while respiration is mainly driven by temperature. The model has been run for 6
years (1986-1991) using NOAA-AVHRR Global Vegetation Index (GV1 product, Tarpley et al., 1984).
For these years, estimates of NEP are compared with measurements of CO2 concentrations in
the atmosphere. This qualitative comparisons give us some indications about the ability of the model to
simulate spatial, seasonal and interannual variations of NEP fluxes. We also attempted to distinguish temporal
variations due to satellite data quality from those resulting from actual change in vegetation productivity. As far
as this distinction is possible, we tried to quantify the impact of atmospheric correction on NDVI, NPP and
NEP.
