A connectivity Compound C index was computed according to the method developed by Borselli et al. (2008) to outline the spatial linkages and the potential connection between the sediment eroded from hillslopes by runoff processes and the different storage areas identified within catchments. These areas may either store sediment temporarily (i.e., reservoirs, lakes or local depressions in the floodplain) or definitively (i.e., outlets). Considering the lack of specific-event data such as soil erosion rates, discharge and suspended sediment concentrations, this index of connectivity
based on GIS data tended to describe the general hydro-sedimentary behaviour of the investigated catchments. To calculate this index, landscape morphological characteristics and recent land use patterns were derived
from high resolution databases. The potential of various land use surfaces to produce or store sediment was also assessed. The calculation was conducted on a Digital Elevation Model (DEM) with a 10-m regular grid provided by the Geospatial Information PCI 32765 Authority of Japan (GSI) from the Ministry of Land, Infrastructure, Transport and Tourism (http://www.gsi.go.jp/). This DEM was computed by the GSI from data obtained by LIDAR airborne monitoring surveys. Values of the weighting cropping and management parameter (the so-called ‘C-factor’), originally used in the USLE equation (USDA, 1978), were determined based on data found in the literature (Borselli et al., 2008, Kitahara et al.,
2000 and Yoshikawa et al., 2004) and applied to the different land use classes observed in the catchments and determined by a multitemporal and multispectral classification of SPOT-4 and SPOT-5 satellite images. SPOT-4 20-m resolution images dated from May 5, June 3 and September 10 2010, and SPOT-5 10-m resolution images dated from March 18, April 13 and 24, 2011. Differences in spectral responses (reflectances) between land uses allowed their spatial discrimination using ENVI 4.8 software. Then, based on their respective vegetal cover density during the spring Protirelin season and their implications on soil sensitivity to erosion, three main land uses were identified (i.e., forests, croplands and built-up areas). Additionally, surface water areas (i.e., rivers, lakes, reservoirs) were delineated. The land use map was validated by generating a set (n = 150) of random points on the map and by comparing the classification output with the land use determined visually on available aerial photographs of the study area. Hydrological drainage networks were derived from the GSI 10-m regular grid DEM using hydrologic analysis tools available from ArcGIS10 (ESRI, 2011).