A Cheat Sheet for Modelling with Hydraulic Conductivity
Water quality impacts to surface water and groundwater receptors can be a major risk of mining operations. Managing that risk requires modelling, monitoring and management. One of the major uncertainties in developing a thorough water quality model is understanding the partitioning of seepage through a mine rock or tailings landform. Does the water report as surface water or as groundwater recharge?
A key control of this partitioning is the hydraulic conductivity (K) of geologic media underlying mine waste landforms. Similarly, hydraulic conductivity is an important control on mobilization and transport of contaminants through saturated mine rock and tailings. Hydraulic conductivity is a physical property which describes the ease with which a porous material can transmit fluid (usually water) through its pore spaces. It differs from permeability in that it is dependent on both the material and the fluid characteristics. Hydraulic conductivity is a parameter that is heavily relied on for use in numerical models to estimate travel times, velocity, and peak concentrations of contaminants of concern (COPC) reaching an environmental receptor. This article present considerations for one simple, yet effective, method to estimate in situ hydraulic conductivity and help conceptualization of water flowpaths at your site.
The reliability of hydraulic conductivity estimates is dependent on three primary factors:
1) How the tests were completedin the field
2) Estimation parameters selected during data processing
3) Appropriate interpretation by the solver
Error introduced at any of these stages may lead to under or over-estimation at the magnitude scale. Keep reading for tips to avoid these errors in your tests and improve your model results.
In the Field
Insitu horizontal hydraulic conductivity may be estimated in saturated deposits under confined or unconfined conditions. ‘Slug tests’ are a common and effective method to estimate in situ hydraulic conductivity in saturated zones. Other type of K tests could include pump tests, infiltration testing, or permeameter testing.
An in situ slug test begins by injecting or withdrawing a volume (typically a 1m3 solid slug) into the water column to generate an instantaneous change in pressure head. The change in displacement over time is monitored using a pressure transducer (logger). Appropriate test selection, range of logger, and reading interval of the logger is crucial to execute a reliable K test.
Some common areas where error can be introduced into the test include:
Logger Range: Loggers should be correctly selected based on anticipated height of water column. Ranges include 5, 10, 20, 30, 100, and 200 m head, with accuracy decreasing with increasing pressure head.
Reading interval: Reading intervals should be selected based on the local lithology. This test may be limited by the soil texture and particle size. Care should also be taken in confirming adequate data points are available based on the interval chosen. The newest loggers have capacity for 100,000 readings . For example, a reading interval of 5 seconds is appropriate for a slow recovery unit such as silty clay and will record measurements for approximately 5.8 days.
Under-developed wells: Fine sediments clogging the screen prevents the movement of water into the well, resulting in an artificially low water table and low recovery in that location leading to under-estimated K values.
Barometric effects: To account for atmospheric pressure effects on the water table, a logger capable of measuring both water level and atmospheric pressure should be used, or a second logger can be installed in the vicinity of the well. The results of the test will be compensated for measured barometric effects.
Static Measurement: Accurate measurement of static water level prior to the test is crucial. Factors to keep in mind which could influence the static water level include barometric effects and underdeveloped wells as discussed above, as well as how the well is capped. If the cap on a monitoring well is secured too tight and does not allow any venting, the static water level measurement required to initiate a slug test will not be representative due to pressure buildup within the well. Best practice when time permits is to open the wells and allow them to equilibrate to static conditions before initiating the test.
Parameter selection is dependent on which analytical solution is being used, under what conditions (confined vs. unconfined), and accuracy of physical measurements taken during execution of the test.
The most common mathematical solutions for overdamped tests include Hvorslev, Bouwer-Rice (B-R), and Kansas Geological Survey (KGS). Underdamped tests are typically solved using the Butler method. Each solution has been developed for specific aquifer properties, and care should be taken in selecting which solution to use.
Hvorslev was developed for use in confined aquifers, however, it can be used in unconfined aquifers provided there is sufficient distance between the water table and the well screen. A tendency to overestimate K due to omission of storativity effects in unconfined aquifers has been documented in literature .
B-R is a modification of the Hvorslev method, developed to account for the effect of partially penetrating conditions , unconfined conditions, and wells that are screened across the water table.
KGS is our preferred analytical solution. It is the most robust solution as it can account for skin effect, anisotropy, and specific storage. The KGS method can be applied to both confined and unconfined aquifers making it the most widely applicable solution, provided the water table does not intersect the well screen.
Butler is a solution developed for underdamped tests characterized by very high K resulting in an oscillatory response (Figure 3). A pneumatic slug tester should be used in these cases to capture the data required. In these tests, recovery is too fast for manual readings to be taken with any confidence.
To summarize some key consideration in the interpretation of K test, interpretation will be discussed in terms of B-R and KGS solutions. Assuming a test is completed in an unconfined aquifer in which the water table intersects the well screen, the B-R solution may be considered as a starting point. A filter pack correction is assumed for a porosity of 30%. Figure 2 illustrates the curve generated from the hypothetical test. Note the “double straight-line effect” which is an indication that B-R is an appropriate solution to use in this case.
Interpretation of this solution involves curve matching and feasibility assessment. When curve matching to the appropriate portion of the curve, an ideal head range between normalized head 0.2 and 0.3 m/m is typically a reasonable starting point. In the below example, the ideal head range captures later time recovery, and would not be considered the best part of the curve to fit. To assess if the estimated K value is feasible, consider what is known about the site, like lithology, groundwater flow direction, and confined/unconfined conditions.
Figure 3 shows an example of a reliable KGS solution in an unconfined aquifer with water level above the well screen. Curve matching in this case is typically computed by a program and involves less interpretation than the B-R solution above.
In this example, the Ss value may be used to assess the reliability of the estimate. An Ss < 1E-6 m/s may indicate an underdeveloped well or other potential issues. Adjustments can be made to initial displacement, anisotropy ratio, and/or well radius relative to borehole radius (to account for skin effects) until Ss provides a reasonable magnitude estimate. In most programs, Ss can also be forced set to a minimum value of 1E-6 m/s.
Okane applies a comprehensive approach for developing hydrological and water quality models for greenfield and operating sites. The hydrological and water quality performance of a mine waste landform are interpreted within the context of the site-specific hydrological and hydrogeological settings, and with appropriate characterization of materials governing COPC fate and transport.
Okane takes great care in developing and interpretating numerical models that serve as powerful tools for understanding water flow and quality in mine-impacted watersheds. Hydraulic conductivity is a key input in developing quantitative estimates of water flow and quality. In the development of 2D and 3D numerical models, horizontal hydraulic conductivity is especially important as it dictates the lateral movement of water through the subsurface towards potential environmental receptors. Used in conjunction with other properties such as hydraulic gradients and COPC source and sink terms, accurate estimation of hydraulic conductivity is an important step towards developing a water quality model that informs the potential risk to water quality and effective strategies to mitigate this risk.