Laboratory and numerical study of hyporheic flow in karst conduits and its effect on sediment-entrapped DNAPLs

Contamination of karst aquifer, particularly by organic compounds, such as the Dense Non-aqueous Phase Liquids (DNAPLs) constitutes a threat for water supply and the environment. Due to the presence of large openings in karst systems and the high density of DNAPLs they may frequently reach conduits and accumulate at low points such as siphon structures that likely also contain sediments. Hence sediment-entrapped DNAPLs are commonly encountered and present as a continuous phase of organic liquid denoted as DNAPL pool. DNAPLs present in conduits are particularly problematic for the water quality as most of the flow in karst systems transits through conduits before reaching springs. The longevity of DNAPL pools will depend on the flow of water through the sediment which is referring to as karst hyporheic flow. In contrast to surface water, hyporheic exchange has received little attention so far, although it has potentially a strong influence on the fate of contaminants in karst systems. The thesis addressed this lack of knowledge. The objective of the thesis was first to verify whether karst hyporheic flow could occur between the conduit and sediments, and second, to investigate how the hyporheic flow mediates the dissolution of sediment-entrapped DNAPL pools. A laboratory analog resembling a siphon conduit partly filled with sediments was built to investigate hyporheic flow and DNAPL pool dissolution. Experiments with the laboratory analog and numerical modeling are the main research approaches throughout this study. <br> First, the driving force and characteristics of hyporheic exchange between the conduit and sediment system was investigated using the laboratory analog model. Influences of conduit flow (expressed as Reynolds number <i>Re</i>), conduit angle and sediment properties (grain size/permeability) on the flow pattern and magnitude of hyporheic exchange were explored. Sediment-source tracer tests were carried out at varied flow rates for sediments of variable grain size to characterize hyporheic flow patterns and to evaluate solute transport processes (by the mean travel time MTT and dispersion of tracer plume σ²). Numerical modeling with a two-domain approach assuming pressure continuity across the conduit-sediment interface was employed to simulate the sediment-source tracer tests and flow fields. Results of the sediment-source tracer tests showed that MTT and σ² decreased with increasing conduit <i>Re</i> and increasing sediment permeability, and were smaller for the steeper conduit angle. Flow pattern in the sediment was characterized by zones of forward and reverse flow induced by the conduit bends. Simulation of the pressure fields showed that flow in sediment was driven by the pressure gradients along the sediment and water interface (SWI), imposed by the overlying conduit flow. While in streambeds, pressure gradients are usually induced by the sediment surface topography, in the current investigated system enhanced pressure gradients occurred on top of a flat sediment surface due to the conduit bends. The observed and simulated flow patterns in different sediments were identical under the same conduit Re for the same conduit angle. However, the magnitude of flow in the sediment varied with sediment permeability for the same conduit <i>Re</i> and geometry. The calculated exchange flux <i>q_int</i> by numerical simulations increased linearly with the conduit <i>Re</i> for all types of sediment and conduit angle, and was higher for a steeper conduit angle. The relationships established among <i>q_int</i> and the various controlling factors also suggested that sediment properties had the strongest influence on the exchange flux. <br> Based on these insights into the karst hyporheic flow, the dissolution of a sediment-entrapped DNAPL pool facilitated by hyporheic flow was further investigated by the same analog model and numerical simulations. Influences of conduit flow and conduit angle on the dissolution rate were evaluated. Dissolution experiments with a well-defined DNAPL pool consisting of a surrogate compound were carried out to quantify the mass transfer rate. Soluble tracer test with a solidified pool was performed to visually trace the solute plume dissolved from DNAPL pool surface, and was compared to the DNAPL pool dissolution reproduced by numerical modeling. The mass transfer coefficient <i>k_m</i>, calculated with the effluent concentration at steady state from dissolution experiments was used to quantify the mass transfer rate. The experimental <i>k_m</i> were compared with the values calculated by numerical simulations. The dissolution rate for the conduits-sediment system was also compared to the well investigated pool dissolution in a semi-infinite porous media as a reference case. The observed <i>k_m</i> increased with the conduit <i>Re</i> and was higher for the steeper conduit angle, which were in agreement with the results from numerical simulations. Both the observed and simulated <i>k_m</i> increased linearly with the average groundwater flow velocity <i>U</i> in the sediment. However, <i>k_m</i> were higher compared to the values predicted by a 2D analytical model which is conventionally applied for the semi-infinite porous media. The higher <i>k_m</i> was likely due to the hyporheic flow pattern characterized by upward curved streamlines in karst conduit sediment. The curved streamlines carry dissolved compounds away from the DNAPL/water interface, while in a semi-infinite porous media upward transport only occurs by transverse dispersion. <br> In conclusion, the study confirmed the occurrence of hyporheic flow between karst conduit and sediment through a laboratory analog investigation. Pressure gradient along the sediment and water interface drove the exchange of flow and solute between the conduit and sediment domains. Conduit flow condition, sediment properties and conduit geometry/angle played a combined role in the hyporheic exchange between conduit and sediment. The pattern of hyporheic flow was controlled by the conduit flow rate and conduit angle. The magnitude of flow in sediment was determined by sediment permeability. The exchange flux increased with the conduit flow rate and sediment permeability and was higher in steeper angle conduit. Sediment grain size/permeability had the strongest influence on the magnitude of exchange flux in contrast to conduit flow and conduit angle. The mass transfer rate of hyporheic flow-mediated DNAPL pool dissolution was found to increase linearly with the average groundwater flow velocity in sediment. The study highlights that a higher mass transfer rate/shorter longevity is to be expected in the shallow karst conduit sediment than in a semi-infinite porous medium due to the influence of the hyporheic flow pattern. The study of potential karst hyporheic flow and relevant processes has implications on the sequestration, transportation and transformation of contaminants in karst aquifer. The exchange of flow and solutes between conduit and sediment potentially plays a significant role in the biogeochemical and karstification processes as well. While the study clearly highlighted the mechanism of hyporheic exchange and its potential role on contaminant fate, further studies are required that approach field conditions in more detail. In particular, the upper limit of the flow rate should be extended to the turbulent range, effect of sediment movement of contaminant fate should be considered and the effect of more complex conduits geometries and sediment assemblies should be explored.

Keywords: karst hyporheic flow, conduit and sediment, DNAPL pool dissolution, exchange flux, mass transfer coefficient Thèse de doctorat : Université de Neuchâtel, 2013