An operational methodology for riparian land cover fine scale regional mapping for the study of landscape influence on river ecological status
Abstract
Preservation and restoration of river ecosystems require an improved understanding of the mechanisms through which they are influenced by landscape at multiple spatial scales and particularly at river corridor scale considering the role of riparian vegetation for regulating and protecting river ecological status and the relevance of this specific area for implementing efficient and realistic strategies. Assessing correctly this influence over large river networks involves accurate broad scale (i.e. at least regional) information on Land Cover within Riparian Areas (LCRA). As the structure of land cover along rivers is generally not accessible using moderate-scale satellite imagery, finer spatial resolution imagery and specific mapping techniques are needed. For this purpose we developed a generic multi-scale Object Based Image Analysis (OBIA) scheme able to produce LCRA maps in different geographic context by exploiting information available from very high spatial resolution imagery (satellite or airborne) and/or metric to decametric spatial thematic data on a given study zone thanks to fuzzy expert knowledge classification rules. A first experimentation was carried out on the Herault river watershed (southern of France), a 2650 square kilometers basin that presents a contrasted landscape (different ecoregions) and a total stream length of 1150 Km, using high and very high multispectral remotely-sensed images (10m Spot5 multispectral images and 0.5m aerial photography) and existing spatial thematic data. Application of the OBIA scheme produced a detailed (22 classes) LCRA map with an overall accuracy of 89% and a Kappa index of 83% according to a land cover pressures typology (six categories). A second experimentation (using the same data sources) was carried out on a larger test zone, a part of the Normandy river network (25 000 square kilometers basin; 6000 km long river network; 155 ecological stations). This second work aimed at elaborating a robust statistical eco-regional model to study links between land cover spatial indicators calculated at local and watershed scales, and river ecological status assessed with macroinvertebrate indicators. Application of the OBIA scheme produced a detailed (62 classes) LCRA map which allowed the model to highlight influence of specific land use patterns: (i) the significant beneficial effect of 20-m riparian tree vegetation strip near a station and 20-m riparian grassland strip along the upstream network of a station and (ii) the negative impact on river ecological status of urban areas and roads on the upstream flood plain of a station. Results of these two experimentations highlight that (i) the application of an OBIA scheme using multi-source spatial data provides an efficient approach for mapping and monitoring LCRA that can be implemented operationally at regional or national scale and (ii) and the interest of using LCRA-maps derived from very high spatial resolution imagery (satellite or airborne) and/or metric spatial thematic data to study landscape influence on river ecological status and support managers in the definition of optimized riparian preservation and restoration strategies.
- Publication:
-
AGU Fall Meeting Abstracts
- Pub Date:
- December 2010
- Bibcode:
- 2010AGUFM.H43G1344T
- Keywords:
-
- 1813 HYDROLOGY / Eco-hydrology;
- 1834 HYDROLOGY / Human impacts