Multispectral satellite observations and mechanistic vegetation models connected to the global carbon cycle

You are here

Context and objectives

This research was preparatory to the ADEOS (Advanced Earth Observing Station) polar platform of the Japanese Space Agency (NASDA). The general theme of the research was the evaluation of the possibilities offered by new sensors such as POLDER (Polarization and Directionality of Earth Reflectances) to use remote sensing data for global carbon cycle modelling. More specifically, the objective was to define and test a methodology for the assimilation of multispectral observations in global biospheric models such as those developed in the Laboratory for Planetary and Atmospheric Physics (LPAP) at the University of Liège.

Project outcome

Expected scientific results

A quantitative comparison between ground observations (climate and soil data, ecosystem distribution), vegetation indices derived from remotely sensed reflectance data and the results from a model of soil hydrology and leaf area index have been performed. A satisfactory agreement was obtained after adjustment of the model parameters to the observations. The maximum LAI had to be adapted to meet the characteristics of the dominant vegetation types found in the study region. Similarly, a threshold minimum amount of precipitation was defined in order to trigger the development of vegetation after a long drought period in a realistic way. This threshold value was adjusted to take the increase of NDVI after the first rainfalls into account.

After these adjustments, the agreement between model and NOAA/AVHRR or SPOT/HRV NDVI values is very promising. However, there remain large differences between observed and modelled NDVI, mostly due to a poor knowledge of bare soil reflectances.
From the preliminary quality assessment of the HAPEX-Sahel AVHRR dataset made in the initial stage of the present study, it can be concluded that the problem of correcting satellite images for atmospheric effects is very complex and not fully mastered up to now. Such corrections are nevertheless required when the "top of the atmosphere )" reflectances need to be converted into “top of the canopy” values.
Project leader(s): ULg - Laboratory for Planetary and Atmospheric Physics
Location: Region:
  • Central Africa,Sahel