¹ú²úÂ鶹

Identifying leak locations and mapping contaminant plumes enables effective source control and remediation. Using vertical and horizontal wells together may effectively leverage the relative merits of each technology.
By Paul F. Hudak

Countless landfills worldwide have contaminated soil and groundwater with leachate. Modern landfills have liners, but small punctures or tears in geomembranes allow leachate to escape outward. Timely leak detection enables effective source control and remediation. Ideally, lysimeters or moisture sensors in vadose zone monitoring networks detect escaped leachate before it reaches the water table. However, many facilities lack sufficient vadose zone networks to timely detect leak(s) originating anywhere in a landfill’s footprint. Moreover, if vadose zone monitor(s) show leakage, a need arises to determine the extent of groundwater contamination.

Small leaks produce narrow contaminant plumes that are difficult to detect in groundwater with a limited number of conventional (vertical) monitoring wells. Potentially, sampling data from horizontal wells constructed with directional drilling could augment information from vertical wells. Horizontal wells can be continuous, with entry and exit holes at ground level, or blind, with no exit hole (EPA, 2004). A casing and screen can be pushed into a blind hole or pulled through a continuous hole. This article outlines a strategy for estimating the location of a leak with a vertical groundwater monitoring network augmented with horizontal wells.

Methods
A graphical approach (Hudak, 1998) was used to design a vertical groundwater monitoring network for a hypothetical landfill above an alluvial sand aquifer (API, 1989) (see Figure 1). With graphics software, flow tubes of equal width were overlaid on the footprint of the landfill. Wells were located 10 m hydraulically downgradient of the landfill, along the centerlines of flow tubes. A block-centered, finite difference mass transport model (Zheng and Wang, 1999) was used to find the maximum width of flow tubes (minimum number of wells) required to detect a contaminant plume originating from small (1-m2) point source within the landfill.

The model consisted of one unconfined layer with 570 columns and 340 rows, with a 0.5 m node spacing. Hydraulic head was constant at the western and eastern boundaries, producing an average hydraulic gradient of 0.01 eastward; no flow crossed other boundaries. Other parameters were:
• Hydraulic conductivity = 1 m/d
• Effective porosity = 0.25
• Longitudinal dispersivity = 1.0 m2/d
• Transverse dispersivity = 0.1 m2/d
• Effective molecular diffusion coefficient = 0.00001 m2/d
• Source concentration = 100 mg/L
• Plume boundary concentration = 1 mg/L

Simulations used the preconditioned conjugate gradient solver for flow and generalized conjugate gradient solver for mass transport, which produced mass balance errors less than 0.03 percent.
A heuristic was applied to model-generated contaminant plumes emerging from five random point sources (one at a time) (see Figure 2). In practice, modeling is not required, but may be helpful to apply the heuristic:
Identify vertical wells (a and b) on either side of the first well detecting contaminant. Extend upgradient construction lines from a and b through the footprint of the landfill to define the zone (Z) containing the leak.

Construct a horizontal well along transect a-b. Collect and analyze samples from discrete intervals of the well to identify the point (e) with the highest concentration.

Extend a longitudinal transect (parallel to the hydraulic gradient) through e, with tentative (yet unknown) endpoints c and d. Based on site constraints, find the entry point along c-d that is closest to the landfill. Typically, the entry point is three to five times the depth of a horizontal well (EPA, 2004). Extend the horizontal well upgradient of the entry point, collecting and analyzing samples to identify the point with the highest concentration (L) and upgradient plume boundary (c).

If a sample from the most downgradient point in (3) exceeds the plume boundary concentration, extend a horizontal well downgradient of the entry point, and sample to identify the downgradient plume boundary.

Discussion
In separate cases, wells M7, M3, M12, M4, and M11 first detected five model-generated contaminant plumes. Elapsed time between onset of leakage and first detection ranged from 270 (M12) to 720 (M11) days. For each case, with sufficient sampling and analytical data, the heuristic identified the leak location, L.

Implicit in the heuristic are key assumptions. The method assumes a linear groundwater flow field. For curved flow paths, adjustments may allow for estimations of L, for example, transect c-d would be curved. The method also assumes that landfill operators identify the first vertical well by detecting a contaminant plume through frequent monitoring. Additionally, horizontal wells must be drillable and monitored at discrete intervals along the intake, for example, with inflatable and mobile straddle packers (Holloway and Waddell, 2008).

Horizontal wells are more technical and costly than vertical wells. In addition to offset considerations noted previously, horizontal wells must be sufficiently deep to remain saturated during periods of drought or groundwater extraction.

The approach is best suited to a vertical network with sufficiently close spacing to detect most contaminant plumes that might emerge from a landfill. A coarse or haphazard layout may allow for offsite contamination prior to first detection. Moreover, if a well in a haphazard vertical network fortuitously detected a contaminant plume in a timely fashion, and neighboring wells were far away, the leakage zone (Z) would be too large; extensive drilling and monitoring would be needed to narrow down the actual location of the leak.

Though limited, the heuristic may be useful at some sites and has advantages over conventional vertical monitoring approaches. For example, horizontal wells can propagate beneath buildings and other surface obstructions. In addition to monitoring, horizontal wells can expedite groundwater remediation by targeting the longitudinal axis (c-d) of a contaminant plume, allowing for greater contact with contamination than vertical wells. Both types of wells can extract contaminants or deliver reagents to an aquifer, in addition to monitoring groundwater quality.

Once a leak is located, it should be repaired, though making repairs can be difficult. Ideally, waste is removed, and the geomembrane is repaired directly. If removing waste is not feasible, targeted injection grouting with conventional grout or expansive polymers (Ke et al., 2024) may stabilize the release. Otherwise, leachate production may be slowed by reducing the amount of water entering the landfill. Attempts at stabilization should involve continued monitoring to track changes in source strength and natural attenuation of the contaminant plume. If a plume is moving too quickly, more aggressive remedial action such as groundwater extraction or permeable reactive barriers may be appropriate.

Effective Source Control
An established technology for aquifer remediation, horizontal wells may have unrealized monitoring applications: to approximate leak locations and map contaminant plumes in groundwater. Identifying leak locations and mapping contaminant plumes enables effective source control and remediation. Using vertical and horizontal wells together may effectively leverage the relative merits of each technology. | WA

Paul F. Hudak is a Professor at the Department of Geography and the Environment, University of North Texas. He received a B.S. in geology from Allegheny College, M.S. in geology from Wright State University, and Ph.D. in geography, emphasis in water resources, from the University of California, Santa Barbara. He has also practiced geotechnical consulting in Pennsylvania, Ohio, California, and Texas. His current research interests include groundwater monitoring and remediation, geologic hazards, urban heat islands, and wetland mitigation. He can be reached at [email protected].

References
API (American Petroleum Institute) (1989). Hydrogeologic Database for Groundwater Modeling. American Petroleum Institute, Washington, D.C.
EPA (U.S. Environmental Protection Agency) (2024). Horizontal Remediation Wells. Available at: https://clu-in.org/techfocus/default.focus/sec/
Horizontal_Remediation_Wells/cat/Overview/.
Holloway, O.G., Waddell, J.P. (2008). Design and Operation of a Borehole Straddle Packer for Ground-Water Sampling and Hydraulic Testing of Discrete Intervals at U.S. Air Force Plant 6, Marietta, Georgia. U.S. Geological Survey Open-File Report 2008-1349, 1-24.
Hudak, P.F. (1998). Configuring detection wells near landfills. Ground Water Monitoring and Remediation, 18(2), 9396.
Ke, Q., Guo, C., Wang, F., Chu, X., Zhai, K. (2024). Investigations on the seepage characteristics of polymer grouting body for repairing HDPE geomembrane defects based on LF NMR. Construction and Building Materials, 414, 135004.
Zheng, C., Wang, P.P. (1999). MT3DMS: A Modular Three-Dimensional Multi-Species Transport Model for Simulation of Advection, Dispersion, and Chemical Reactions of Contaminants in Groundwater Systems; Documentation and User’s Guide. U.S. Army Engineer Research and Development Center, U.S. Army Corps of Engineers, Contract Report SERDP-99-1, Vicksburg, Mississippi.