The Earth Observer



July/August 1996, Vol. 8, No. 4

The VEGETATION Programme

- Gilbert Saint (gilbert.saint@cst.cnes.fr), VEGETATION Programme Scientist, CNES DPI/OT, 18 avenue E. Belin, 31055, Toulouse, France

The VEGETATION Programme was undertaken under a partnership between the European Commission, France, Belgium, Sweden, and Italy, to set up a system to monitor the biosphere at medium resolution, taking advantage of the SPOT satellite series with the capability to have simultaneous high-spatial-resolution acquisition with the same spectral bands. It will be ready for launch in December 1997.

Introduction

The overall objectives of the VEGETATION system are to provide accurate measurements of the basic characteristics of vegetation canopies on an operational basis:

The VEGETATION system, consisting of a satellite-borne sensor and its associated ground segment, will provide long-term basic measurements adapted to biosphere studies. Opportunities for scale integration are provided by the combination with the main SPOT instrument (HRVIR) which allows high spatial resolution for detailed modeling activities or multilevel sampling procedures. Availability of data to different types of users is facilitated through the centralization of reception and archiving of global data sets. The launch date (nominally December 1997) and duration of the system (5 years of estimated life time for a first model and continuation on future SPOT satellites) are adapted to a systematic and extensive long-term monitoring of the biosphere.

Clearly, this system will benefit from detailed studies based on other systems that are dedicated to specific studies of the characteristics of remote sensing measurements or to their relationships with surface or process parameters. It must be envisaged that the evolution of the mission specifications will have to take into account results of such studies to provide improved characterization of the biosphere state and dynamics.

Mission Objectives

Surface parameter mapping: This is the basic requirement, especially for climate and meteorological studies, where boundary conditions have to be prescribed as in the case of General Circulation Models (GCMs) or forecasting models. Factors such as albedo, surface roughness, resistance to heat exchange (sensible and latent) are important variables for these models, and they can be either determined directly from the measurements or inferred from identification of land cover. The seasonal and long-term variations of such variables are related to vegetation dynamics. The capability to identify, through these variations, the physical characteristics of land cover is a key to accurate prescription of these variables. Scales addressed in GCM or forecasting models (typically about 100 km) require that land cover and its variability must be determined with a sampling of about 8 to 10 km: the basic spatial resolution needed for identification of land cover and its variability is 1 km.

Agricultural, pastoral, and forest production: Since the beginning of the land surface satellite remote sensing era (1972), important projects (for example LACIE, AGRISTARS for USDA, MARS for CEC, TREES for JRC/ESA...) have been set up to develop methodologies and strategies to use remote sensing data either for mapping of land use in anthropogenized or natural ecosystems or for estimation of production potential. Their specific objective was to determine the evolution of production. This objective had to be adapted to the management of crop production for agricultural exporting countries, to the monitoring of pastoral resources and their dependence on meteorological evolution, to the evaluation of possible global impacts of deforestation, and more generally to the need for information related to political or social orientations and decisions.

Terrestrial biosphere monitoring and modeling: The contribution of the continental biosphere to the biogeochemical cycles (exchanges of carbon and other trace gases) and to water and energy exchanges is one of the objectives of the development of global models. Interaction with human activities is also one of the main points to be studied, because the effect of human pressure on the biosphere might be one of the means by which man is acting on climate in the long term. Biosphere processes and land cover characterization are the basis for quantification: estimates of land cover variables as well as the dynamics of these variables have to be made in order to obtain a good understanding of these processes upon which models may be built. Predictions of impact of climate change on the biosphere and of interactions of the biosphere with the climate (either due to natural factors or to human pressure) can only be inferred from quantification and formalization of the mechanisms by which vegetation cover and ecosystems function. Multilevel series of models have to be developed and linked, ranging from ground studies, local parameterization and exchange models to regional or global dynamics and interaction models. Remote sensing of the vegetation as shown above offers a unique tool for these developments, providing the specification of the systems be adapted to each particular need.

VEGETATION System Characteristics

Radiometry

Spectral characteristics:
Spectral bandsWavelengthSurface reflectance range
BLUE 0.43 - 0.47 µm 0.0 - 0.5
RED 0.61 - 0.68 µm 0.0 - 0.5
NIR 0.78 - 0.89 µm 0.0 - 0.7
SWIR 1.58 - 1.75 µm 0.0 - 0.6

Radiometric resolution (NE delta p):
BLUE 0.003 for the entire range
RED 0.001 up to reflectance of 0.10, linear increase up to 0.003 for reflectance of 0.5
NIR, SWIR 0.003 for the entire range

Intra-image consistency within an entire image, corresponding to a NE delta p of 0.005 for any reflectance value.

Calibration accuracy:
Interband and Multitemporal: better than 3%
Absolute: better than 5%

Geometry

Spatial resolution:

Geometric accuracies:
Local distortion less than 0.3 pixel
Multispectral registration 0.1 km desired, 0.3 km specified
Co-location with HRVIR 0.3 km for simultaneous acquisitions
Multitemporal registration 0.3 km desired, 0.5 km specified over one year
Location accuracy better than 500 m desired, 1000 m specified

Spatial coverage:
About 90% of the equatorial areas are imaged each day, the remaining 10% being imaged the next day. For latitudes higher than 35° (North and South), all regions are acquired at least once a day.

Operation specifications:
Equator crossing time (descending node): 10:30 local solar time

Image transmission:
All spectral bands at full spatial resolution acquired on terrestrial areas will be stored on-board in a solid-state- memory, allowing the use of only one receiving station to which data will be transmitted in X band. All the spectral bands will also be transmitted in L band, for possible regional receiving stations.

VEGETATION Products

These standard products have been defined by the International Users' Committee. They are adapted to the particular missions described above and coherent as much as possible with the needs of existing projects. To illustrate the special characteristics of the instrument, high priority was given to design products that would allow direct multitemporal registration as well as simple superposition with simultaneously acquired high-resolution data.

VGT-P Products

These products are adapted for users for whom the physical quality of the data is important. They correspond to data which would have been acquired by an ideal instrument: they are corrected for system errors (misregistration of the different channels, calibration of all the detectors along the line-array detectors for each spectral band) and resampled to geographic projections for multitemporal analysis as well as for comparison with high-resolution data. The accuracies given above apply to this data level. Annotations giving full information on applied corrections (calibration information, geometric parameters taking into account attitude and position on the orbit), or for further non-system corrections ("standard" atmosphere parameters) are attached to the data sets.

VGT-S Products

These products are most probably the data sets which will be frequently used operationally: they correspond to VGT-P data to which corrections have been applied using the annotations and for which some syntheses are provided :

To adapt to the evolution of users' needs as well as to the validation of new algorithms, a procedure to regularly update the processing system is requested: it should provide capabilities to include new methods for data correction, synthesis, etc., as soon as they are commonly accepted by the user community.

Support to users will be provided to facilitate the use of VEGETATION data: a catalogue with browsing capability on the data quality (cloudiness) will be accessible through the usual networks. Validated software templates for the common operations for data handling and standard correction will be made widely available.

The Future

To prepare evolution and continuity on the long term, the design of the next system is already being considered, with the objectives:

The next instrument will be launched on SPOT 5, planned around 2002.

Detailed information can be found on the World Wide Web at URL : http://www-vegetation.cst.cnes.fr:8050/.

[Table of Contents]
[Previous]
[Next]