3.7 Vulnerability of Karst Aquifers to Contamination

Karst aquifers are recognized as being especially vulnerable to contamination. Contaminants sourced from above-ground human activities and land uses, such as accidental chemical spills or releases, agricultural chemical applications and livestock waste disposal easily and rapidly reach the aquifer via surface runoff entering sinkholes and sinking streams. Contaminants that infiltrate soils or are released from subsurface sources such as buried wastes, landfill leachate, and underground storage tanks and pipelines, may move rapidly and/or slowly through the epikarst into deeper parts of the aquifer. Once contaminants reach the saturated zone, they may move rapidly and/or slowly throughout the aquifer depending on the distribution of zones of higher and lower hydraulic conductivities in the bedrock matrix, and the hydraulic properties of fractures and conduits.

Heterogeneities created by the multiple porosity and permeability of karst aquifers, and especially by multiple discrete conduit and fracture flow paths, are the major factor contributing to the overall higher vulnerability of karst aquifers to contamination, and to the difficulties encountered in detecting, assessing, and remediating contaminant occurrences (Field, 1993). Contaminant transport, storage, fate, and remediation are especially complicated where the karst aquifer is extremely heterogeneous and/or the contamination involves mixtures of contaminants having different physio-chemical characteristics, such as a combination of non-aqueous-phase liquid (NAPL) and dissolved contaminants (Figure 31). The conventional conceptual model of a single contaminant plume defined by concentration gradients and spreading gradually through an aquifer by advection and diffusion does not represent contaminant transport in karst. Contaminant transport through conduits occurs rapidly, often over long distances, and contaminants within them may be dispersed into multiple “plumes” of varying concentration that take unpredictable or unknown pathways to the aquifer’s discharge boundaries at wells, springs, or surface waters. Contaminants may be rapidly diluted within or flushed from conduit systems, but residual contaminants may be stored and released gradually or effectively immobilized within lower-permeability, slower-velocity zones of the karst aquifer matrix (Green et al., 2006). Contaminant flow paths, flow velocities, residence times, travel times and concentrations may change significantly under different hydrologic conditions.

Figure showing contamination of a karst aquifer

Figure 31  Hypothetical representation of multiple storage and transport pathways following two spills of dense non-aqueous phase liquid (DNAPL) into a heterogeneous karst aquifer showing migration and pooling of free-product (i.e., DNAPL that is not mixed with water) and the associated zones of dissolved DNAPL that form as groundwater flows around the free-product. Modified from Wolfe and others (1997).

Vesper and others (2001) broadly summarize the factors involved in contamination of karst aquifers by:

  • inorganic and organic water-soluble compounds;
  • LNAPLs – Light Non-Aqueous Phase Liquids that are slightly soluble organic liquids, such as gasoline, that are less dense than water;
  • DNAPLs – Dense Non-Aqueous Phase Liquids that are low to slightly soluble organic liquids, such as chlorinated solvents like trichloroethylene (TCE), that have a higher density than water;
  • pathogens including microbes and viruses;
  • metals; and,
  • trash.

Their key points are concisely highlighted in this excerpt:

“Transport of the contaminants through the aquifer is by a variety of mechanisms depending on the physical and chemical properties of the contaminant…Water soluble compounds…move with the water. But rather than forming a plume spreading from the input point, the contaminated water forms linear stringers migrating down the conduit system toward the discharge point. LNAPLs…float on the water table and can migrate down the water table gradient to cave streams where they tend to pond behind obstructions. DNAPLs…in contrast, sink to the bottom of the aquifer. In the conduit system, DNAPLs pond in low spots at the bottom of the conduit and infiltrate sediment piles. Transport of both LNAPL and DNAPL is dependent on storm flow which can force LNAPL through the system as plug flow and can move DNAPLs by mobilizing the sediment piles. Pathogens…are transported through the karstic drainage system because of the absence of filtration and retain their activity for long distances. Metals (i.e. chromium, nickel, cadmium, mercury, and lead) tend to precipitate as hydroxides and carbonates in the neutral pH, carbonate rich water of the karst aquifer. Metal transport is mainly as particulates and as metal adsorbed onto small particulates such as clays and colloids…”.

Trash and sediment are relatively underrecognized contaminants uniquely associated with environmental and water-quality degradation in karst (Mahler and Bennett, 1999; Mahler et al., 2007). Often introduced directly into the conduit system by storm-induced runoff entering open sinkholes and swallow holes, these contaminants may be transported by turbulent flow far into the karst aquifer system where they may act as subsurface sources of leachable chemical and microbial pollutants. High concentrations of suspended sediment or turbidity, especially during storm events, is a hallmark of sinkhole-dominated karst recharge, and is a surrogate indicator of karst water vulnerability to contamination by surface runoff. In many karst aquifers, even those unaffected by anthropogenic contaminants, groundwater quality is naturally degraded by suspended sediment. It is often a factor limiting groundwater resource development, affecting the water quality of water-supply wells and springs, and is a recurrent or ongoing problem contributing to higher costs for water treatment and water well maintenance. The mineralogical composition and grain-size of the sediment are important co-factors in the transport, storage, and fate of many contaminants in karst aquifers. For example, in areas undergoing increasing urbanization, high organic carbon content and high specific surface area increases the potential of sediments to transport metals and organic chemical contaminants (Mahler et al., 1999). Core samples of layered sediment deposits in conduits provide useful data to track temporal and spatial changes in karst water quality and pollutant loading and provide a record of changing anthropogenic activity in a karst basin (Feist et al., 2020).

The article “Threat Down Below: Polluted Caves Endanger Water Supplies, Wildlife” by Streater (2009) highlights a number of fascinating cases of contamination in karst and their impacts on caves, groundwater, and karst ecosystems. Historically, in rural areas sinkholes have been used to dispose of: trash; industrial and agricultural liquid wastewaters and slurries; highway stormwater runoff; and household septic wastewater. In the United States, wells categorized as Environmental Protection Agency Class V, inject non-hazardous fluids underground and have been drilled in many karst areas to enhance drainage from soil-mantled sinkholes, dispose of highway runoff, or mitigate karst-related flooding (Zhou, 2007). In conduit systems that have free air space, volatile gases may exsolve from contaminated water or LNAPL and migrate to the upper parts of the aquifer, through the vadose zone to the soil or surface, creating hazardous explosive or noxious conditions. An example of this was documented by White and others (2018). Karst areas are notable for higher concentrations of radioactive, carcinogenic radon gas (Rn86), a naturally-occurring element generated by decay of uranium-containing minerals. Radon gas is often detectable in spring waters and at high concentrations in the air of limestone caves as well as buildings underlain by karstic carbonate rocks (Hakl et al., 1997; Peano et al., 2011).

Point-source (for example, sewage or industrial plant outfall) and non-point source (for example, diffuse contamination from cattle field) contaminant releases in karst terranes may result in rapid and devastating affects to drinking-water supplies (Field, 2004). One widely-reported, tragic incident involved microbial contamination of public water-supply wells in Walkerton, Ontario, Canada, in the spring of 2000, in which over 2,300 people were sickened, and seven died, as a result of groundwater contamination by Escherichia coli (E. coli) and Campylobacter jejuni bacteria. Subsequent investigations revealed that the source of the contamination was livestock waste entering the karstic limestone aquifer through rapid infiltration of surface runoff. Field testing using dye-tracing methods demonstrated the occurrence of flow velocities greater than 300 m/d and contaminant travel times of 5 to 26 hours in an area that conventional groundwater model applications had been delineated as having a 30-day time-of-travel. The groundwater model simulations did not include karst features, thus did not represent the flow system. Obviously, conventional groundwater modeling mis-represented the time require for contaminants to reach points of contact because karst features were not included in the model (Worthington et al., 2002).

This type of outcome—where groundwater simulation fails to accurately represent karst flow and contaminant transport characteristics—should be expected if the karst flow system is poorly conceptualized and the mathematical model does not properly represent the potential multiple porosity and permeability scale effects created by preferential flow paths within conduits or thick macro-porosity preferential flow layers. Conventional groundwater flow models typically use values reflecting total porosity of the rock matrix to represent effective porosity. Effective porosity represents only the porosity of interconnected conduits and preferential flow layers which is generally a much smaller value than total porosity. Contaminant travel time is linearly related to effective porosity, so over estimation of porosity increases calculated travel time. Effective porosity is most accurately determined from tracer testing (Kuniansky et al., 2001; Davis et al., 2010).

Ewers (2006) discusses the importance of properly collecting and interpreting hydrogeologic mapping data needed to determine conduit flow and contaminant transport characteristics, and the use of special techniques such as dye-tracing tests to identify groundwater flow directions, and connections between potential sources and receptors of contaminants. Field (2004) advocates for the use of multiple quantitative tracer tests initiated from potential source locations to obtain dye-breakthrough curves. These estimated solute-transport parameters that are needed to predict the fate of contaminant releases. In karst aquifers, quantitative groundwater tracing is the single-most demonstrably reliable method of obtaining information about hydraulic geometry of conduits, groundwater flow velocities and residence times, as well as other insights into contaminant transport characteristics that cannot be acquired using conventional methods such as potentiometric-surface mapping and aquifer tests (Mull et al., 1988a; Field and Nash, 1997). Dye-tracer tests are an effective method of estimating the behavior of soluble conservative contaminants (and to lesser extent non-soluble and reactive contaminants) whose transport characteristics are dependent on groundwater flow velocities. However, other tracer agents are needed, and are available, to simulate transport of bacteria, colloids, non-soluble and particle contaminants. Examples are provided by Benischke (2021) and Bandy and others (2016). Tracer properties should be carefully evaluated and matched to the known or anticipated type of contaminant transport under investigation. More discussion of water tracing is provided in Section 5.3, Water Tracing Tests.

Exercise 8 invites the reader to read about poluted caves and consider the types of contaminants found in the caves.


Introduction to Karst Aquifers Copyright © 2022 by Eve L. Kuniansky, Charles J. Taylor, and Frederick Paillet. All Rights Reserved.