Exploring geological and topographical controls on low flows with hydrogeological models.

This study investigates how catchment properties influence low-flow dynamics. With 496 synthetic models composed of a bedrock and an alluvial aquifer, we systematically assess the impact of the hydraulic conductivity of both lithologies, of the hillslope and of the river slope on catchment dynamics. The physically-based hydrogeological simulator HydroGeoSphere is employed, which allows obtaining a range of low-flow indicators. The hydraulic conductivity of the bedrock Kbedrock , a proxy for transmissivity, is the only catchment property exerting an overall control on low flows and explains 60% of the variance of Q95/Q50. The difference in dynamics of catchments with same Kbedrock depends on hillslope gradients and the alluvial aquifer properties. The buffering capacity of the bedrock is mainly related to Kbedrock and the hillslope gradient. We thus propose the dimensionless bedrock productivity index BPI that combines these characteristics with the mean net precipitation. For bedrock only models, the BPI explains 82% of the variance of the ratio of Q95 to mean net precipitation. The alluvial aquifer can significantly influence low flows when the bedrock productivity is limited. Although our synthetic catchment set-up is simple, it is far more complex than the available analytical approaches or conceptual hydrological models. The direct application of the results to existing catchments requires nevertheless careful consideration of the local geological topographic and climatic conditions. This study provides quantitative insight into the complex interrelations between geology, topography and low-flow dynamics and challenges previous studies which neglect or oversimplify geological characteristics in the assessment of low flows.