ABSTRACT
ABSTRACT: Rangeland ecosystems are important elements of Mediterranean landscapes and provide a variety of goods and services, ranging from resources for livestock production systems to conservation of flora and fauna. The sustainable use of areas dominated by livestock grazing requires finding the sometimes delicate balance between productivity and corresponding stocking rates. However, given the land use change processes of the past decades, it is not sufficient to define average regional stocking rates. The access to rangelands often occurs with such a spatial diversification that environmental degradation might occur although average stocking rates are not exceeded. This not only affects the local carrying capacity, but might impose feedbacks on the hydrological functioning of headwater areas and downstream valleys. This is but one example illustrating the need for spatially differentiated information on ecological state variables for supporting the required spatial management of resources. This paper explores options for deriving leaf area index as a major biophysical variable from earth observation satellite data. Inverse reflectance modelling of Landsat-TM data is constrained with fractional cover estimates obtained from very high spatial resolution Quickbird imagery, thereby increasing the possibilities to obtain woody vegetation LAI estimates from open canopies. Results underline the complexity of canopy reflectance signals in heterogeneousMediterranean landscapes. The spatial diversity and variation was well reproduced by the model inversion, and woody LAI integrated to site level was found consistent with field-based measurements. On the other hand, the estimation shrub LAI was found to be strongly affected from the background signal introduced in the model, which affected results in particular in regions with low woody vegetation cover.
