Landfill and water quality : Analysis: How low-discharge wells powered with solar energy may be useful for slowing plume growth

Landfill managers are committed to preventing leachate from spreading beyond the site to safeguard public health and the environment. Slowing or halting the growth of plumes allows landfill managers to assess natural attenuation and other remedial alternatives. Emerging sustainable technologies for remediating soil and groundwater contaminated by landfills include permeable reactive barriers, electrokinetic methods, microbial action and injecting solubilising agents or air bubbles (Ye et al., 2019).
In the event of a breach being documented, low-discharge wells powered with solar energy may be useful for slowing or stopping plume growth. Extracted groundwater can be treated above ground and reinjected into the subsurface. Liner perforations in landfills have the potential to generate slim contaminant plumes that can be retrieved using low-discharge wells. Furthermore, it may be feasible to transform passive detection wells, which are mandatory at landfill sites, into low-discharge pumping wells to stem a contaminant plume. This article provides an outline and assessment of a plan that transforms detection wells into pumping wells, to counteract the spread of small-scale contaminant plumes from an imaginary lined landfill.

A diagrammatic method (Hudak, 1998) was utilised to develop an inactive monitoring system for a notional landfill (Figure 1). Flow pipes with even width were superimposed onto the landfill's area. The central axis of flow pipes contained monitoring wells, situated 10 m down the slope of the landfill. The central axis of flow pipes contained monitoring wells, situated 10 m down the slope of the landfill. A mass transport model (Zheng and Wang, 1999) was utilised to determine the optimal width of flow tubes, i.e. the minimum number of wells necessary to detect a contaminant plume originating from any given point source within the landfill.
The finite-difference model was constructed with one layer, 570 columns (trending north-south), 340 rows (trending east-west) and a total of 193,800 cells. The nodes located at cell centres were spaced 0.50 metres apart along both columns and rows. The hydraulic head registered 5.000 m and 2.155 m in the western and eastern extremities, respectively. The measurements were taken from a datum at the model's base. No fluid progressed through the northern or southern boundaries. The hydraulic gradient towards the east had an average of 0.01.

Other variables, which represent an alluvial aquifer (API, 1989), included a hydraulic conductivity of 1.0 m per day, effective porosity of 0.25, longitudinal dispersivity of 1.0 m, transverse dispersivity of 0.1 m, and effective molecular diffusion coefficient of 0.00001 m2 per day. During mass transport simulations, point sources contained a concentration of 100 mg/L, while plume boundaries held a concentration of 1.0 mg/L. The flow simulations employed a preconditioned conjugate gradient solver and the mass transport simulations utilized a generalized conjugate gradient solver. The mass balance errors for both were less than 0.03%.
Following the establishment of an initial detection network, the model generated contaminant plumes emerging from five random point sources (one at a time) within the landfill's footprint. The plume's geometry was examined through the model output at the initial detection time and the moment it first made contact with the site boundary.

Throughout subsequent remediation trials, the source remained active. Upon the initial detection of a contaminant plume at a well site, it was repurposed as an extraction well, while the furthest cross-gradient wells were repurposed as injection wells. Each injection well pumped at half the rate of the extraction well, but in the opposite direction. Multiple simulations were conducted with different pumping rates to determine the minimum required to stabilize the plume onsite. Stability is characterized by the leading tip of the plume ceasing to advance while the source and pumping wells continue to operate. The model output also identified the plume geometry at stability and the time required to achieve stability.

The initial detection network comprised 12 monitoring wells positioned near the two downgradient sides of the landfill. Five randomly placed sources were located at different points within the landfill's footprint. In various simulations, a narrow contaminant plume developed from each point source and reached one or more detection wells before arriving at the downgradient site boundary. Contamination plumes were initially spotted between 270 days and 720 days. These plumes then advanced towards the site boundary, occurring between 4,000 days and 4,800 days.

The contamination plume situated in Case 4, located closest to the landfill's downgradient boundary, was detected quite early, resulting in timely conversion of detection wells to pumping wells. Furthermore, the plume detected in Case 4 was comparatively small, resulting in a shorter stabilization time when subjected to pumping . Conversely, the sources in Cases 1 and 5 were situated further upgradient within the landfill and therefore yielded larger plumes at initial detection. The plume in Case 5 was marginally bigger, necessitating a higher pumping rate, which ultimately resulted in quicker stabilization when compared to Case 1. Reducing the time needed for stabilization, the altered extraction well in Case 5 was more accurately positioned along the ambient hydraulic gradient to focus on higher levels of pollutant concentrations.

The point source in Case 3 was situated in close proximity to the first well, which was a cross-gradient well detecting the plume, and along the surrounding hydraulic gradient. Since one cross-gradient well was transformed into an extraction well, solely one cross-gradient well continued to remain, which was transmuted to an injection well. The rate of injection was equivalent to the extraction well but in the opposite direction. This alternative proved to be efficacious as the associated plume stabilised at an earlier stage and was relatively insignificant upon stabilisation.

After the pumping process commenced, the duration required to attain plume stabilization varied between 3,200 days (Case 3) and 5,400 days (Case 1) (refer to Table 1). The plumes exhibited sluggish growth towards stability, but they did not penetrate the boundary of the site. When the plumes finally stabilized, they were at their narrowest downstream of the extraction well, which arrested the spread of contaminants that would have otherwise widened the plume (refer to Figure 2). Additionally, injection wells that were positioned in the cross-gradient direction pushed the contaminants towards the extraction well, which augmented the process of plume narrowing. The fixed, foremost end of a stabilised plume indicates an equilibrium between the opposing influences of the extraction well and the ambient hydraulic gradient. Once it has attained stability, maximum concentrations are typically detected in a confined zone surrounding the source and the portion of the landfill boundary positioned in the downgradient direction (Figure 2).The required flow rate to stabilise a plume extended from 0.3 m3/d (Case 4) to 0.5 m3/d (Case 5) (Table 1, Figure 2). Case 4 required less time to stabilise because the plume was small on initial detection, and the transformed extraction well almost aligned (along the ambient hydraulic gradient) with the source. In addition, early recognition and efficient alignment in Case 4 culminated in a smaller plume at stabilization (Figure 2).

In practice, stabilizing a contaminant plume prevents offsite migration while giving landfill managers time to assess a situation and apply appropriate remediation measures. Remedial action might involve repairing a leak, monitored natural attenuation of a leak and associated plume, or additional measures outside the footprint of a landfill to reduce the size of a plume. While potentially effective in some cases, low-discharge pumping wells would not work as effectively for contaminants with low solubility, or for wide contaminant plumes. Cases outlined above involved modeling a contaminant plume at first detection, to find a minimum pumping rate to stabilize it on site. However, in practice, the location of a point source, and the geometry of an associated plume, is rarely known at first detection. Samples from an initial array of detection wells can aid in defining a plume. The location of the first well that detects a contaminant plume helps narrow down a plausible (upgradient) subarea containing the point source. Additional wells and samples may be necessary to define an initial plume for modeling remedial action.

This study examined the capability of passive detection wells converted to low-discharge wells for stabilizing narrow leachate plumes originating from lined waste impoundments. The first well contacting a plume was converted to extraction, and the farthest cross-gradient wells on either side of the extraction well were converted to injection. At low pumping rates, this conversion strategy effectively stabilized five plumes emerging from small random leaks in a hypothetical landfill. The strategy outlined here may be worthwhile at some sites contaminated by small leaks in lined impoundments.