Links between karst hydrogeological properties and statistical characteristics of spring discharge time series: a theoretical study

Physics-based modeling of karst systems remains almost impossible without enough accurate information about the inner physical characteristics. Usually, the only available hydrodynamic information is the spring discharge at the karst outlet. Numerous works in the past decades have used and proven the usefulness of time-series analysis applied to spring discharge, precipitations or even physico-chemical parameters, for interpreting karst hydrological functioning. The main objective of this work is to provide additional insights of to what extent the informative content of the hydrodynamic signal at karst springs is sensitive to karst aquifers internal physical properties. To address this issue, we undertake an empirical approach based on the use of both distributed and physics-based models and on synthetic systems responses. The link between karst hydraulic and physical properties was studied. For this purpose, forward modeling of flow through several simple, constrained and synthetic cases in response to precipitations is undertaken. It allows us to quantify how the statistical characteristics of flow at the outlet are sensitive to changes (1) one hydraulic parameter of the model and (2) in conduit network geometry. The matrix/conduit exchange parameter appeared clearly as a determinant model parameter in the spring discharge simulation. The auto- and cross-correlation functions seem to be of particular interest for the understanding of the karst inner physics. Indeed, these functions are always different, despite not so pronounced configuration differences. This would highlight that there is an informative content within the spring discharge time series and the usefulness of such analysis methods.