Investigating Runoff Generation in a Headwater Forested Catchment Using a Typology of Nonlinear Storage-Discharge Relationships

We investigated the "switching on and off" of runoff generation in a small forested catchment, the Hermine, characterized by moderately steep surface topography and an extensive but discontinuous fragipan-like horizon at an average depth of 0.75 m within the soil mantle. A water table dataset (94 wells, 50 sampling campaigns) was used to derive point-scale perched groundwater (PGW) storage variables illustrating the presence or absence of surface water, the soil deficit to saturation, the location of dry, filling and spilling subsurface hollows, and the water table elevation above an underlying filled subsurface hollow. A typology of storage-discharge relationships was also built to differentiate locations driven by threshold-crossing between two extreme states from other locations experiencing a continuum of hydrologic states. We found that spatial patterns of PGW storage were initially important during a rainstorm to trigger a response at the catchment outlet but then remained unchanged even though streamflows continued to increase up to an order of magnitude higher. While discontinuous storage-discharge relationships were observed at the point-scale, continuous relationships were rather observed at the aggregated catchment-scale, thus implying that the emergent catchment behaviour could be described as a continuum of hydrological states. At the point-scale, storage-discharge relationships were used to assess the spatial heterogeneity of runoff processes as their associated shapes and thresholds indicated the timing of either preferential flow activation or subsurface connectivity enhancement. At the catchment scale, a critical outflow parameter of 1 L.s-1 was found to transcend small-scale process heterogeneity as it signalled the triggering of the catchment internal stormflow generation mechanisms. Our typology of nonlinear relationships was especially useful for grouping similar runoff responses both in space and time. We therefore make a plea for cross-site comparisons of "nonlinear hydrological relationships" as their different shapes can inform us on regime tipping points at the catchment scale and spatially variable processes at sub-catchment scales.