Remote sensing and GIS techniques for assessing and prioritising riparian corridor restoration

[Departement_IRSTEA]Eaux [TR1_IRSTEA]QUASARE

International audience

إنجليزي. Preservation and restoration of river ecosystems requires an improved understanding of the mechanisms through which they are influenced by landscape at multiple spatial scales from the overall upstream watershed to the local riparian corridor. This research is dedicated to the statistical analysis of the role of riparian vegetation in regulating and protecting river ecological status. It is particularly focussed on the relevance of land use indicators and the spatial footprint over which they are calculated, for the assessment and implementation of efficient and realistic preservation and restoration strategies in the riparian corridor. The method was implemented over the Normandy river network (6000 km long river network; 155 ecological stations). It analysed (i) the potential of Very High Spatial Resolution (VHRS, metric) remote sensing imagery for a fine scale mapping of the Riparian Area Land Cover (RALC); (ii) the use of GIS techniques to produce spatial indicators at riparian scale; and (iii) the design of Regionalized Pressures/Impacts (RPI) models studying relationships between land cover indicators and indicators of the stream biological response. The overall objective is to provide quantitative and spatialised basis for the design of a regional restoration strategy for the riparian corridor. In a first phase, we elaborated a generic classification procedure (Tormos et al., submitted-010), that produces RALC maps in different relief, climate and geology contexts. This procedure uses information from orthophotos (0.5 m pixels) and Spot 5 XS (10 m pixels) image data, and complementary spatial thematic data. It produces a reliable detailed (62 classes) RALC map. In a second phase, we derived from this RALC map, using automated GIS techniques, several spatial indicators that quantify the proportion of a given land cover category within a wide range of riparian footprints. A footprint is a buffer with a given lateral distance to the river and given longitudinal distances upstream and downstream from the ecological station. Finally, we developed RPI models (Wasson et al., 2010) based on these spatial indicators at riparian scale over the Normandy river region. Thanks to the fine RALC mapping and the indicators computed on a spectrum of riparian footprints it was possible to identify and localize more precisely the different pressures and to demonstrate the specific influence of riparian vegetation (on a 20m wide strip on both sides of the river) on river ecological status at regional level. In the near future, new spatial indicators at riparian scale, incorporating knowledge of human impact and riparian vegetation influence on aquatic ecosystems (Tormos et al., in press-2010), and new biological indicators will be used in RPI modeling in order to prioritize riparian corridor restoration.