A quantitative method for the characterisation of karst aquifers based on spring hydrograph analysis

Abstract

This paper presents a method for characterizing flow systems in karst aquifers by acquiring quantitative information about the geometric and hydraulic aquifer parameters from spring hydrograph analysis. Numerical sensitivity analyses identified two fundamentally different flow domains, depending on the overall configuration of aquifer parameters. These two domains have been quantitatively characterized by deducing analytical solutions for the global hydraulic response of simple two-dimensional model geometries.During the baseflow recession of mature karst systems, the hydraulic parameters of karst conduits do not influence the drainage of the low-permeability matrix. In this case the drainage process is influenced by the size and hydraulic parameters of the low-permeability blocks alone. This flow condition has been defined as matrix-restrained flow regime (MRFR). During the baseflow recession of early karst systems and fissured systems, as well as the flood recession of mature systems, the recession process depends on the hydraulic parameters and the size of the low-permeability blocks, conduit conductivity and the total extent of the aquifer. This flow condition has been defined as conduit-influenced flow regime (CIFR).Analytical formulae demonstrated the limitations of equivalent models. While equivalent discrete-continuum models of early karst systems may reflect their real hydraulic response, there is only one adequate parameter configuration for mature systems that yields appropriate recession coefficient. Consequently, equivalent discrete-continuum models are inadequate for simulating global response of mature karst systems. The recession coefficient of equivalent porous medium models corresponds to the transition between matrix-restrained and conduit-influenced flow. Consequently, equivalent porous medium models yield corrupted hydrographs both in mature and early systems, and this approach is basically inadequate for modelling global response of karst aquifers.

A fractal approach to the recession of spring hydrographs

“…In contrast, we assume that the highly permeable system, which may consist of fractures or tubes (both called conduits in the following), has infinite permeability and is directly connected to the spring where the discharge is measured, so that the highly permeable conduits are held at constant hydraulic head. This is reasonable for mature karst aquifers [Kovács et al, 2005], but also applies to porous fractured aquifers if the permeability contrast between matrix and fractures is sufficiently high. Thus, the total outflow is determined by the flow from the lowly permeable regions to the highly permeable conduits.…”

“…It has been suggested [Kovács et al, 2005] to use the point of inflection in the hydrograph, i.e., the steepest descent, as an estimate for the end of recharge. We have followed this idea.…”

Groundwater management of a highly dynamic karst by assessing baseflow and quickflow with a rainfall-discharge model (Dardennes springs, SE France)

“…Separating the hydrograph of karst springs has been the subject of numerous studies over several decades, using recession analysis (e.g., Atkinson, 1977;Baedke and Krothe, 2001;Fiorillo, 2014;Fu et al, 2016;Geyer et al, 2008;Kovács and Perrochet, 2008;Kovács et al, 2005;Mangin, 1975;Padilla et al, 1994). Recession analysis can be applied at the flood scale or at the annual scale.…”

Influence of aquifer heterogeneity on karst hydraulics and catchment delineation employing distributive modeling approaches

“…While the low hydraulic conductivity of the matrix is limiting groundwater flow in this part of the catchment, the ability of the conduits to conduct water becomes limiting close to the Gallusquelle spring and causes water to flow out of the conduits and back into the matrix. According to the classification after Kovács et al (2005) the flow regime in this part of the model area thus is conduit influenced. Scenario 4 shows a significantly better fit for the hydraulic gradient distribution (Fig.…”

“…5e). This corresponds to the matrix-influenced flow regime according to Kovács et al (2005), where the discharge is controlled by the matrix rather than by the conduits. The effect is not strong enough to completely avoid an overestimation of hydraulic heads in the east and an underestimation in the central part and in the west (Fig.…”