Local residents often view former dumpsites as nothing more than harmless, rolling green fields, but hidden danger might lie just beneath the surface: landfill gas, containing potentially explosive concentrations of methane.

By Gordon K. Parish, Renee Pionk, Kenneth Davis, and Abigail Hendershott

Figures to supplement this article may be viewed in pdf format by clicking here

As our cities continue to sprawl into rural areas, they might encroach on more abandoned waste sites, placing new homeowners at risk from landfill gas (LFG) migrating through the subsurface. The sudden explosion of a home in Rochester Hills, MI, in the middle of the night on April 21, 2000, dramatically illustrated the potential hidden dangers associated with these types of sites. The explosion was triggered when a pilot light ignited methane that had seeped into the basement from a former dump nearby.

In late 1999, the Michigan Department of Environmental Quality (MDEQ) and Malcolm Pirnie Inc. discovered the presence of potentially explosive concentrations of methane in soils beneath several homes adjacent to an abandoned landfill site located north of Kalamazoo, MI. Malcolm Pirnie Inc., acting under the MDEQ's emergency contracting procedures, designed and supervised the installation of an emergency perimeter extraction system to intercept and withdraw LFG at the property boundary. Contour plots of methane concentration data collected monthly during the first year of operation illustrate a decrease in subsurface methane concentrations beneath the homes from as much as 40% before startup in July 2000 to nondetect by June 2001.

This case study emphasizes the need for regulators and consultants to recognize the potential explosive dangers presented by these abandoned landfill sites, and it underscores the close communication, coordination, and cooperation between the MDEQ, Malcolm Pirnie, and the local residents during all phases of this project: the preliminary gas migration assessment, the public meetings to disseminate information and household methane detectors to the local residents, the expedited design and installation of the emergency extraction system, and the subsequent monitoring of the extraction system. The development of detailed geologic cross-sections, methane concentration isopleths, and absolute pressure contours reveal how the installation of a simple, well-positioned gas extraction system might have averted an unthinkable catastrophe.

Site History

The KAV Company Sanitary Landfill was a rural dumpsite operated by a group of private haulers from the Kalamazoo area. The landfill is located on an 80-ac. parcel north of Kalamazoo, and the limits of solid waste placement are believed to occupy approximately 52 ac. This unlined facility reportedly accepted general refuse and construction debris from approximately 1958 to January 31, 1983, when a temporary injunction ordered the cessation of waste acceptance and disposal at the landfill. The KAV Company filed for bankruptcy protection under Chapter 11 in December 1984, and the MDEQ eventually was granted permanent access to the landfill property.

Because the site predated the promulgation of Michigan's Act 641 (Michigan's Solid Waste Management rules, which incorporated the RCRA Subtitle D requirements, subsequently updated and replaced by Act 451, Part 115), the MDEQ placed the facility under the jurisdiction of the Environmental Response Division (ERD) and directed closure activities in accordance with state environmental remediation rules (Act 451, Part 201), which are designed to be protective of human health and the environment.

Preliminary LFG Migration Assessment

In 1998, the MDEQ retained Malcolm Pirnie Inc. to provide various services related to the site, including groundwater monitoring, landfill cover evaluation, and the assessment of the potential for horizontal LFG migration. The contractor supervised an Earthprobe investigation that included field screening of subsurface soils for LFG and the installation of 20 temporary gas monitoring probes around the periphery of the landfill and on neighboring properties. The gas monitoring probes were constructed within the Earthprobe boreholes using 1-in. diameter polyvinyl chloride (PVC) casing with 12-ft. lengths of slotted screen. Depths of installation were set based on field screening results but generally varied from approximately 20 to 40 ft. below grade.

The results of the preliminary LFG migration assessment, completed by the fall of 1999, indicated that potentially explosive concentrations of LFG had migrated more than 1,000 ft. beyond the western landfill property line. Deeming that LFG migration represented a potential threat to the local residents, the MDEQ approved the design and installation of a pilot-scale, active, perimeter LFG extraction system and organized a public meeting to disseminate information and provide methane detectors to the residents living between the western property line and Doster Road, located approximately 1,500 ft. west of the landfill property.

Geological Setting

The site is located near the western margin of a glacial end moraine with glacial ground moraine deposits east of the landfill and glacial outwash and lake plain deposits less than 0.5 mi. west of the landfill. The ground elevation varies from approximately 990 ft. above mean sea level (msl) along the northwest perimeter of the property to approximately 940 ft. above msl along portions of the eastern and southwestern property line. Groundwater, which generally flows to the west, has been identified at an elevation of approximately 890 ft. above msl in the vicinity of the landfill.

The shallow unsaturated soils, ranging from 30 to 60 ft. thick along the landfill's western property line, are primarily of glacial moraine origin and consist largely of poorly sorted mixtures of sand, silt, and clay. This shallow zone appears to comprise two general units: an upper deposit of interbedded silts and clays and a lower deposit of poorly sorted silty sand. The interbedded silts and clays range in thickness from approximately 20 to 30 ft. Beneath the interbedded silts and clays lies a deposit of poorly sorted, silty sand, which contains discontinuous lenses of silty clay. Perched water was observed on several of these clayey lenses. This unit varies from approximately 20 to 30 ft. thick in the vicinity of the extraction system.

The deeper unsaturated soils appear to be of glacial outwash or lake plain origin and consist primarily of well-sorted, rounded, medium sands overlying more poorly sorted deposits of sand and gravel. Because the well-sorted sands are relatively uniform and clean, they likely possess greater permeability than the shallow soils. During well installation, field personnel observed that methane concentrations were more consistently greater in this deep zone than in the shallow zone.

Site-specific historical geologic data were obtained from the boring logs of groundwater monitoring wells, but these borings had been logged based on observations of the cuttings rather than on soil samples collected from distinct intervals. Therefore, detailed geologic data were collected during installation of the emergency gas extraction system and monitoring network, and field personnel adjusted screened intervals and well locations based on their observations.

Extraction System Description

Seven 4-in.-diameter PVC LFG extraction wells (GPWs) were installed on a spacing of approximately 150 ft. along the northwest perimeter of the landfill property. Each extraction well is screened from approximately 15 ft. below grade to 5 to 10 ft. above the groundwater table (Figure 1). The extraction wells were installed using 6.25-in. hollow-stem augers, and split-spoon samples were collected in accordance with the Standard Penetration Test (ASTM D-1586) procedures to characterize the subsurface geology. At each boring location, subsurface gas concentrations were collected and recorded at 10-ft. intervals. A coarse silica sand pack was placed around each screen to a depth of approximately 12 ft. below grade, and the remainder of the annular space was sealed with bentonite slurry above a 2-ft. layer of bentonite chips. The wells were completed in a flush-mounted vault and fitted with valves for flow control.

Each extraction well was connected to a buried, 8-in.-diameter PVC header pipe connected to a pair of 5-hp, trailer-mounted blowers, which induce a vacuum within the wells and extract the LFG from the subsurface soils. Each blower is equipped with a flow meter, which measures the instantaneous volumetric flow rate in standard cubic feet per minute (scfm). Each blower passes the extracted gas through a pair of 1,000-lb. carbon filters, operating in series, before discharge to the aosphere. These carbon filters contain a mixture of activated carbon and H₂Solution to remove trace volatile organic compounds and hydrogen sulfide that might be present.

Monitoring Network

Eight trilevel monitoring points (TLs), consisting of a cluster of three 1-in.-diameter PVC wells nested within a single borehole, were installed at locations approximately 75 ft. north and south of each extraction well (Figure 1). Each well was installed using 4.25-in. hollow-stem augers. As with the extraction wells, split-spoon samples as subsurface gas concentrations were collected to characterize the geology and assist selection of screen intervals. Each well was completed with a 10-ft. screen set at depths selected based on depth to groundwater, observed geology, and measured methane gas concentrations. A coarse silica sand pack was placed around each screen, at intervals from approximately 2 ft. below the bottom of the screen to approximately 2 ft. above the top of the screen. The annular space between each well screen, and above the uppermost well screen, was sealed with bentonite slurry sandwiched between bentonite chips. Personnel fitted each wellhead with an airtight, valved assembly that includes an internal, 0.25-in.-inside-diameter polyethylene tube extending down to the well screen. To reduce the potential aesthetic inconveniences to local property owners, the monitoring points were set inside flush-mounted protective casings.

The TLs were designed to allow the collection of gas composition and subsurface pressure data necessary to evaluate whether the emergency LFG extraction system has induced and maintained a continuous vacuum across the length of the system, as well as the effect within each geologic unit.

Nineteen bilevel monitoring points (BLs), consisting of a cluster of two 1-in.-diameter PVC wells nested within a single borehole, were installed generally west of the gas extraction system. These points were installed and constructed in the same fashion as the TLs. BL-1 through BL-6 were installed along a line approximately 500 ft. west of the extraction system. BL-15 and BL-18 were placed south and north of the extraction system. The remaining BLs were located on upgradient and downgradient sides of residences.

The BLs were designed to provide a means of monitoring gas composition and subsurface pressure at various distances from the landfill, particularly near the homes along Doster Road. Deep-well screens were set within the well-sorted medium sand, and shallow-well screens were primarily set within the poorly sorted silty sand, but those close to the homes, such as BL-7 (shown in Figure 1), were set somewhat shallower to provide a better indication of gas concentrations in the vicinity of the basements.

Field Procedure

Gas composition and pressure were analyzed using CES-LANDTEC GEM 500 and GEM 2000 gas extraction monitors. The GEMs measure concentrations of methane, carbon dioxide, and oxygen in percent by volume and measure pressure in inches of water column relative to aosphere (gauge pressure). Malcolm Pirnie personnel calibrated the GEMs before each monitoring event for the gas components, and the gauge pressure was rezeroed to aosphere no longer than every two hours during monitoring to correct for changes in aospheric pressure.

For this project, methane concentrations were the most critical parameter. Carbon dioxide, oxygen, and the balance (assumed to be nitrogen) provided supplemental information useful for evaluating the integrity of the measurement. Pressure data were essential to verify that the extraction system maintained a continuous vacuum along the property line to prevent further migration of LFG. Because it took at least a full day to conduct one round of monitoring, gauge pressures had to be converted to absolute pressures in order to evaluate network-wide pressure data. Field personnel also recorded pressure data along a single line of monitoring points (GPW-4, TL-4, BL-3, BL-7, and BL-8) at the beginning, middle, and end of monitoring events. These data were used both to establish a baseline pressure drift throughout the event and to allow an assessment of subsurface pressure versus distance from the extraction well.

LFG Extraction System Results

The two blower systems consistently provided a combined flow of approximately 300 scfm. The composite methane concentration in the extracted gas was greater than 50% in July 2000 and remained greater than 40% through October 2000. During this time period, the recorded methane in extraction wells GPW-2 through GPW-7 persisted at concentrations of 40% or greater, but the methane concentration at GPW-1 dropped to less than 20%. By January 2001, however, the composite methane concentration in the extracted gas had dropped to approximately 20%, and methane concentrations at either end of the system (extraction wells GPW-1, GPW-2, GPW-6, and GPW-7) had decreased to less than 15%. Since June 2001, the composite methane concentration has remained approximately 10%.

LFG Monitoring Results

The methane concentration isopleths presented in Figures 2 through 4 illustrate the reduction in methane concentrations within the deep soil zone during the first year of operation of the extraction system. At system startup, methane concentrations were greater than 60% at a distance of more than 500 ft. west of the landfill property line and up to 40% beneath the homes along Doster Road (Figure 2). Beneath the homes there was a discernable decrease in methane concentrations by September 2000, though concentrations nearer the extraction system generally persisted. By January 2001, however, observed methane concentrations had decreased to less than 40% in the immediate vicinity of the extraction wells and to less than 12% beneath the homes (Figure 3). Observed methane concentrations across the monitoring network continued to drop steadily each month, eventually reaching concentrations of less than 20% in TL-4 and nondetect beneath the homes by June 2001 (Figure 4).

Methane concentrations also have generally decreased in the shallow soil zone, but because the shallow soils are more heterogeneous and contain more fines, this decrease has occurred much more slowly and does not plot as consistently on a site plan. During the first year of operation, shallow soil methane concentrations at BL-7s and BL-9s, located just upgradient from the two homes directly west of the extraction system, decreased from approximately 20% in July 2000 to less than 15% in January 2001 and to less than 5% by June 2001.

Subsurface Pressure Results

Following system startup, Malcolm Pirnie conducted a radius-of-influence test to verify that the system achieved the minimum required radius of influence required (75 ft. at each extraction well) to maintain control of the migrating gas. Because this project involved an emergency response to a potential hazard, the system and monitoring network were designed to mitigate an emergency condition and to provide data useful in determining whether the migrating gas had been cut off and contained at the property line. Although the network of monitoring wells does not allow the determination of a precise radius of influence, it is sufficient to allow verification that the minimum required radius of influence has been achieved. During the radius-of-influence test, TL monitoring wells demonstrated an immediate drop or rise in pressure when the system was alternatively started and stopped. Under normal operating conditions, the observed apparent induced vacuum at each extraction well was approximately 2 in. of water column. At each deep TL monitoring well, the observed apparent induced vacuum was as much as 1 in. of water column.

At this site, observed subsurface pressures have been influenced noticeably by the shallow silty/clayey end moraine deposits. During monitoring events, subsurface gauge pressures have been observed to lag the aospheric pressure, which tends to rise and fall more rapidly than the subsurface pressure. As the aospheric pressure rises, the subsurface will experience low pressure relative to the aosphere. Air then infiltrates into the subsurface until the pressures equilibrate. Conversely, when aospheric pressure falls, the subsurface will experience high pressure relative to the aosphere until equilibrium is reached. For this reason, gauge pressures do not necessarily provide an accurate representation of the subsurface pressure distribution when the data are collected over a long time interval. Therefore, pressure data were converted to absolute pressures and plotted the results for each monitoring period.

To verify that the system consistently maintains this minimum radius of influence with variations in aospheric conditions, subsurface pressures were collected and plotted during each monitoring event. A typical absolute pressure contour plot for the deep soil zone is presented in Figure 5. Deep soil zone pressure data indicate that individual extraction wells are inducing a minimum radius of influence sufficient to maintain a vacuum along the extraction system. As represented in Figure 5, the apparent induced vacuum at each extraction well is approximately 3-4 in. in. of water column. The recorded absolute pressure readings at the TL monitoring wells are approximately 1.0 - 1.5 in. of water column less than those at BL-1 through BL-6.

During each monitoring event, personnel also recorded pressure readings from control wells, as described under Field Procedure. These control readings not only document subsurface pressure drift throughout a monitoring event but also provide data that can be used to plot absolute pressure versus distance from extraction well GPW-4. Documenting the subsurface pressure drift during an event allows at least a qualitative assessment of the subsurface pressure distribution for a given monitoring event. A large pressure drift throughout the day (for example, a large rise in subsurface pressures in the wake of a low-pressure front) would indicate that the spatial distribution of pressures is skewed over the time it took to collect the readings. Documenting the absolute pressure versus distance for a line of wells collected within a 10- or 15-minute span factors out any such drift.

The absolute pressure versus distance from extraction well GPW-4 during the July 2002 monitoring event is presented in Figure 6. In this figure, a logarithmic curve has been fitted through the first three points for each recorded time interval. Although the number and spacing of monitoring points associated with this system does not allow a precise interpretation of the radius of influence, it is sufficient to demonstrate that there is an apparent induced vacuum at the TLs of approximately 1.0 - 1.5 in. of water column.

Conclusions

The discovery of potentially explosive concentrations of methane gas in soils beneath homes as far as 1,500 ft. from an old, abandoned landfill site underscores the need for regulators and consultants to be wary of potential dangers associated with these kinds of sites. In this case, after discovering the potential hazard, the MDEQ implemented emergency measures that might have helped avert a disaster. During the design and construction of the emergency LFG extraction system, the MDEQ distributed methane detectors to local residents and kept them informed of site activities.

The geology of the area influenced not only the migration pathway of the LFG but also the effectiveness of the extraction system. The shallow, low-permeability soils associated with the glacial end moraine restricted gas migration but also restricted short-circuiting, enhancing the influence of the extraction system in the deeper, more permeable, well-sorted sand, which has been the main migration pathway for the LFG.

Based on continued monitoring, the emergency LFG extraction system effectively cut off the continued migration of LFG from the landfill and appears to have extracted a significant amount of gas that had already migrated toward the residences to the west. Methane concentrations in the deep zone began dropping shortly after startup, and within the first year, concentrations underneath the residences dropped from as much as 40% to nondetect. Methane concentrations in the shallower, less permeable soils were more irregularly distributed and decreased more slowly but, by the end of the first year of operation, had fallen to less-than-explosive concentrations beneath the residences. The emergency LFG extraction system has successfully cut off the continued westward migration of LFG and effectively removed the main source of methane recharge into shallower soils.

Gordon K. Parish, P.E., C.P.G.; Renee Pionk; and Kenneth Davis, P.E., are consulting engineers with Malcolm Pirnie in East Lansing, MI. Abigail Hendershott is a senior environmental-quality analyst with the Michigan Department of Environmental Quality, Remediation and Redevelopment Division, Grand Rapids district office.

MSW - May/June 2003