Vol. 7, No. 1, 11-20, 2008

Abstract
The scope of this study was to find a simple and robust technique to analyse a 16-year time series (totalling 576 decades) of NOAA-AVHRR derived Green Vegetation Fraction (GVF). The bio-physical properties of the observed vegetation canopy are described as a function of its composition in terms of a seasonally changing vegetation component and a permanent vegetation component. The principal idea behind the analysis is to use a simple model of an annual vegetation growing cycle per pixel which is fitted against the available time sequence of data, and interpret the parameters of the fit on the one hand, and the residuals of the original versus the fitted data on the other hand. For simplicity reasons this part is represented by a sine curve with a fixed wavelength of one year. This model allows splitting of the timely resolved vegetation signal into two components in vegetation appearance. One represents a "permanent background" throughout the year which is the off-set between the 0-level representing the absence of vegetation cover and the minimum of the modelled seasonal change. The second represents the difference between the maximum and the minimum vegetation cover modelled every year. This technique was applied to the entire Mediterranean region covered by a NOAA-AVHRR time series. The derived proportions of permanent and seasonal vegetation components were finally interpreted in relation to the European CORINE land cover class 'Olive grove', assessing the variation of permanent and seasonal vegetation components as a function of management intensity. This led to a distinction of different olive grove management intensity classes within the limits of the CORINE class. The olive class was chosen as test case because of its well known linkages between the evergreen component represented by the olive trees and the more or less pronounced presence of annual herbaceous understorey.

ISSN 1729-3782