Soil piping: detection, hydrological functioning and modelling

Soil piping remains a relatively unexplored phenomenon despite its substantial impacts on watershed-scale water and sediments transfer in numerous locations around the world. In very general terms, soil piping refers to the formation of sub-surface pipes or tunnels due to the erosive action of water flowing through the soil. Presumably initiated by biopores or desiccation cracks, these natural pipes are often considered as the largest category of macropores (sufficiently large for water to sculpt their form) and have the potential to provide subterranean networks with the greatest hydrological connectivity (Jones et al., 2010). Present in a wide range of pedoclimatic environments across the globe (Verachtert et al., 2011), these networks act as conduits for water, sediments, gases and solutes (Smart et al., 2012). Over the past forty years, research has demonstrated the important contribution of natural soil pipes in water transfer and especially their role in accelerating subsurface flows and expanding stormflow contributing area. However, these studies do not yet allow to fully assess their hydrological functioning. According to Jones et al. (2010), this deficiency is due to a lack of continuous measurements and the complex characterization of pipe networks (their underground and heterogeneous nature makes the number of pipes, their dimensions, positions and connectivity very difficult to estimate). Following these observations, we propose, in this research project, to develop an integrated methodology that will allow a better understanding of natural soil pipes hydrological behavior at the hillslope scale in the specific context of Loess-derived soils (Gueule valley, Sippenaeken, Belgium). This novel methodology involves: 1. Non-invasive, high-resolution imaging using ground-penetrating radar (GPR) using the latest advances in signal processing and object detection in order to characterize the pipe network (number of pipes, dimensions, positions and connectivity) as well as the position of any impervious layer. 2. The installation of flow meter and piezometer networks to fully characterize the link between pipeflow and the water table (which is crucial but neglected in almost every research until today). These monitoring networks should present high temporal and spatial resolutions. Combining this information at the hillslope scale should not only allow to better understand the hydrological behavior of a pipe network in Loess-derived soil but also to model its hydrological functioning.