NOAA-AVHRR/NDVI profiling on CORINE land cover classes

In the framework of the Monitoring Agriculture with Remote Sensing project (MARS) satellite derived information are largely used to derive crop information and indicators for production assessment. Among the different data sets available project, indicators on vegetation from NOAA-AVHRR (Normalised Difference Vegetation Index) is the most used within agro-meteorological analysis. It is defined as: NDVI=(ρ2-ρ1)/(ρ2+ρ1). Where ρ2 and ρ1 are respectively the reflectance (%) in the near infrared and in the red channels. Some efforts have been recently made within the MARS project to extract information from the NDVI, which can be linked to crops. In this framework two major lines are being followed: the derivation of growth parameters through energy balance models and the derivation of indicators to fit statistical models. A challenge in both the approaches is the decomposition of the NDVI vegetation mixed values into more reliable values for single vegetation targets. The final objective is to better exploit the time series of NDVI, linking them as most as possible to crop growing conditions, extracting indicators which can be related more closely to crop yield performances and fitting forecasting models. To achieve this objective the vegetation mixed indicator must be deseggregated to a crop specific indicator. The use of an independent land cover map increases the information content of the extracted NDVI profiles. In this context the use of CORINE land cover data as a stratification layer is introduced and some examples are given