Strategy to document heterogeneity in soil properties and its impact on water transfers from slope to catchment scales

Heterogeneity in soil properties has been identified to impact water\ntransfers at different scales from vertical column, hillslopes to watershed. Thus Distributed physically based hydrological models require distributed hydraulic characteristics to quantify these impacts. To characterize soil properties and their heterogeneity, a multi-scale sampling strategy was proposed based on distributed information including electromagnetic survey maps, topography and land use coverage. Each identified units are characterized by there hydraulic properties including in situ infiltration tests. This strategy was applied over the Ara Catchment (12km2) in northern Benin. It has been instrumented in the\nframework of the AMMA-Catch experimental network in West Africa, to better determine water resources and to investigate possible hydrological retro-action on monsoon cycle. From hydrological point of view, distributed soil hydraulic properties are supposed to impact water transfers and watershed dynamics all along the monsoon cycle. To document this heterogeneity, an electrical conductivity map and geological survey was used as starting points to identify the ground structures which align with the north-south direction with a dip angle\nof 20° east. A total of 20 pits have been opened to document the 0-2m\nhorizons, and 2 more for the 0-5m horizons. 3 pits were digged within each geological structure areas at the surface. In each pit, the retention and hydraulic conductivity curves of each pedological horizon were characterized with three replicates. This database is used to document the variability of these properties and to produce soil hydraulic property maps. Using the variability information, we tested their impact with the Parflow-CLM 3D distributed model. It was run in an homogeneous configuration and compared with a data controlled heterogeneous configuration. The latest is prepared using a turning band algorithm to distribute soil hydraulic properties.