Improving the quantification of land cover pressure on stream ecological status at the riparian scale using High Spatial Resolution Imagery

摘要

The aim of this paper is to demonstrate the interest of High Spatial Resolution Imagery (HSRI) and the limits of coarse land cover data such as CORINE Land Cover (CLC), for the accurate characterization of land cover structure along river corridors and of its functional links with freshwater ecological status on a large scale. For this purpose, we compared several spatial indicators built from two land cover maps of the Herault river corridor (southern France): one derived from the CLC database, the other derived from HSRI. The HSRI-derived map was obtained using a supervised object-based classification of multi-source remotely-sensed images (SPOT 5 XS-10 m and aerial photography-0.5 m) and presents an overall accuracy of 70 %. The comparison between the two sets of spatial indicators highlights that the HSRI-derived map allows more accuracy in the quantification of land cover pressures near the stream: the spatial structure of the river landscape is finely resolved and the main attributes of riparian vegetation can be quantified in a reliable way. The next challenge will consist in developing an operational methodology using HSRI for large-scale mapping of river corridor land cover,, for spatial indicator computation and for the development of related pressure/impact models, in order to improve the prediction of stream ecological status.