Extraction of the phenological signal per plant functional type from medium resolution observations

摘要

The objective of this study is to describe the main plant functional type (PFT) based on plant phenology from the CYCLOPES (Carbon cYcle and Change in Land Observational Products from an Ensemble of Satellites) cover fraction biophysical product (fCover). The 10-days composite at 1km resolution derived from VEGETATION sensor will thus be used. The method proposed to describe the phenology in mixed pixel situations is based on a desaggregation technique applied on a moving window of few hundreds of pixels. A parametric model describing the dynamics of fCover is adjusted for each PFT in the window. The solution of this high dimensional system is regularized using prior information on the distribution of the parameters of the phenological model, as well as on the land cover. The technique is be first tested and demonstrated over an area close to the city of Toulouse (southern France) by using a time series of high resolution Spot-HRV images.