Monitoring Leaf Area Index of Mediterranean oak woodlands: comparison of remotely-sensed estimates with simulations from an ecological process-based model

摘要

Annual vegetation abundance mapping was carried out within the DeMon II European project over a period of 12 years (1984-1996). The project relied on advanced satellite-based methods for spatial and temporal monitoring of Mediterranean oak woodlands by means of a series of Landsat Thematic Mapper (TM) satellite data. A standardized approach developed previously focuses on the Languedoc site, Hautes Garrigues, a typical sensitive Mediterranean region, but now recovering after centuries of grazing and agricultural activities. After geometric and radiometric rectification of nine full Landsat TM scenes with a refined correction in a smaller area of 75 km x 75 km, a GIS database was created containing satellite data, thematic maps of vegetation, geological maps, climatic data and field measurements. An empirical relation between radio-metric ground truth measurements and satellite derived Normalized Difference Vegetation Index (NDVI) allows us to derive Leaf Area Index (LAI). An ecological process-based model (Forest BGC) has been adapted to simulate ecosystem processes in a satisfying way at a local scale. Consistent results were obtained from remote sensing data and from simulations at a local scale, suggesting the possible use of remote-sensing data to monitor vegetation abundance changes at a regional scale. Without considering human disturbances, it can be noted that not much variation of LAI induced by natural factors can be detected over the considered 12-year period.