ArticlePDF Available

Mobility and persistence of methane in groundwater in a controlled-release field experiment

  • The Lyell Centre, Heriot-Watt University
  • G360 Institute for Groundwater Research
  • Shell International Exploration and Production Inc.

Abstract and Figures

Expansion of shale gas extraction has fuelled global concern about the potential impact of fugitive methane on groundwater and climate. Although methane leakage from wells is well documented, the consequences on groundwater remain sparsely studied and are thought by some to be minor. Here we present the results of a 72-day methane gas injection experiment into a shallow, flat-lying sand aquifer. In our experiment, although a significant fraction of methane vented to the atmosphere, an equal portion remained in the groundwater. We find that methane migration in the aquifer was governed by subtle grain-scale bedding that impeded buoyant free-phase gas flow and led to episodic releases of free-phase gas. The result was lateral migration of gas beyond that expected by groundwater advection alone. Methane persisted in the groundwater zone despite active growth of methanotrophic bacteria, although much of the methane that vented into the vadose zone was oxidized. Our findings demonstrate that even small-volume releases of methane gas can cause extensive and persistent free phase and solute plumes emanating from leaks that are detectable only by contaminant hydrogeology monitoring at high resolution.
Content may be subject to copyright.
PUBLISHED ONLINE: 27 MARCH 2017 | DOI: 10.1038/NGEO2919
Mobility and persistence of methane in
groundwater in a controlled-release
field experiment
Aaron G. Cahill1,2, Colby M. Steelman1, Olenka Forde2, Olukayode Kuloyo3, S. Emil Ru3,
Bernhard Mayer3, K. Ulrich Mayer2, Marc Strous3, M. Cathryn Ryan3, John A. Cherry1
and Beth L. Parker1*
Expansion of shale gas extraction has fuelled global concern about the potential impact of fugitive methane on groundwater
and climate. Although methane leakage from wells is well documented, the consequences on groundwater remain sparsely
studied and are thought by some to be minor. Here we present the results of a 72-day methane gas injection experiment into
a shallow, flat-lying sand aquifer. In our experiment, although a significant fraction of methane vented to the atmosphere,
an equal portion remained in the groundwater. We find that methane migration in the aquifer was governed by subtle grain-
scale bedding that impeded buoyant free-phase gas flow and led to episodic releases of free-phase gas. The result was lateral
migration of gas beyond that expected by groundwater advection alone. Methane persisted in the groundwater zone despite
active growth of methanotrophic bacteria, although much of the methane that vented into the vadose zone was oxidized. Our
findings demonstrate that even small-volume releases of methane gas can cause extensive and persistent free phase and
solute plumes emanating from leaks that are detectable only by contaminant hydrogeology monitoring at high resolution.
Unconventional petroleum development1has generated con-
troversy regarding the occurrence, magnitude and effects
of fugitive methane leakage2,3 from imperfectly sealed well
bores4,5. Identifying the frequency, origin and pathways of fugitive
methane in the subsurface6,7 is paramount to understanding its po-
tential impacts on our fresh groundwater resources. Numerous field
techniques have been developed in an attempt to detect methane
leakage8–10, while scenario-based numerical modelling has been
used to assess subsurface behaviours11. However, no field research
has specifically or comprehensively addressed methane migration
and fate in the subsurface at high spatiotemporal resolution to more
rigorously understand the dominant mechanisms and their interac-
tions. Consequently, evidence-based conceptual understandings of
mechanisms controlling fugitive methane distribution in freshwater
aquifers remain weak despite expert panel recommendations to
examine potential environmental impacts12,13.
Methane leakage from oil and gas wells occurs in an estimated
1–3% of unconventional wells due to imperfect well bore seals2,14.
The seals are formed of cement and are geochemically unstable in
the saline formations through which the wells are emplaced15,
resulting in eventual deterioration and leakage. However,
controversy arises because the typical measurements conducted at
well heads consider only gas leaking up along the well casings to
the atmosphere16, and neglect the potential for methane escaping
the well casing into geologic formations and migrating within the
subsurface, potentially impacting potable groundwater resources.
A series of papers that relied only on sampling existing domestic
and farm water wells in active shale gas plays in Pennsylvania8,17
claim that methane occurrence was common within a kilometre
of the well; however, other studies claim this methane is explained
by processes not related to shale gas development18–20. Much of this
debate originates from inherent ambiguities in measurements ob-
tained from domestic water wells19 that are merely receptors of con-
taminants, as opposed to properly engineered groundwater moni-
toring networks to understand source–receptor relationships19,21.
Here, we present the results of a multi-disciplinary field
experiment that examined the movement, impacts and fate of free-
gas methane injected into a shallow freshwater aquifer comprised
of nearly homogeneous, Pleistocene beach sand located at CFB
Borden, Ontario, Canada, a well-established aquifer research facility
(Fig. 1a). We considered surface gas efflux and ground-penetrating
radar (GPR) measurements to track subsurface free-gas movement,
along with aqueous chemistry (including dissolved gases and
C isotope ratios of methane) and microbial characterization to
assess the nature and extent of groundwater quality impacts in the
well-characterized Borden aquifer22,23 under natural conditions of
geologic heterogeneity, groundwater flow and hydrochemistry.
Hydrogeologic monitoring construct
Methane (51 m3or 36.4 kg at standard temperature and pres-
sure) was injected using electronic mass flow control valves simul-
taneously into two 45-angled drive-point wells over 72 days at
two depths situated at 4.5 and 9 m below ground surface (Fig. 1b).
A four-phase injection schedule was used with rates of 0.17, 1.0
and 4.3 m3d1(Supplementary Table 1), which represent the low
to medium range in surface casing vent flows observed in Alberta
and British Columbia, Canada14. Monitoring was performed for
245 days after start of injection across a network of depth-discrete
monitoring points (Fig. 1b,c). The groundwater monitoring network
was designed on the basis of previous Borden aquifer24 and CO2gas
1G360 Institute for Groundwater Research, University of Guelph, Ontario N1G 2W1, Canada. 2Department of Earth, Ocean and Atmospheric Science,
University of British Columbia, British Columbia V6T 1Z4, Canada. 3Department of Geoscience, University of Calgary, Calgary, Alberta T2N 1N4, Canada.
© 2017 Macmillan Publishers Limited, part of Springer Nature. All rights reserved.
Clay aquitard
6−9 cm d−1
Sand aquifer
Vadose zone
Low dissolved oxygen
increasingly anaerobic
with depth
Clay aquitard
4.5 m bgs
9.0 m bgs
CH4 injection
GPR profiles
Extended monitoring network
CH4[aq] and PTDG
CH4 chambers
Sampling screens
M5 M2 M7
2 m
4 m
5 m
6 m
8 m
1 m
Gas sparging screens
Surface eux chambers
C1/C3 C2
Dissolved oxygen
1 m
Figure 1 | Hydrologic setting and experimental set-up of the methane injection experiment. a, Borden aquifer consisting of a thin vadose zone and a
relatively homogeneous sandy aquifer underlain by a clay aquitard. The upper injector is located in the aerobic zone and the lower one in the anaerobic
zone where there is slightly higher baseline dissolved methane and increased sulfate. b, Cross-section of groundwater monitoring network adjacent to
injection points. bgs, below ground surface. c, Plan view of monitoring network including multilevel piezometers, surface eux chambers and geophysical
profile lines. The injection control centre consisted of a mobile laboratory, methane canisters and computer-controlled mass flow valves.
injection experiments25. Methane efflux was monitored at surface
and within the vadose zone (Fig. 1b,c), while GPR was used to
monitor the migration of the gas phase in the aquifer (Fig. 1c).
The aquifer consists of 9m of horizontal, interconnected lenses
ranging from silt to coarse-grained sand with limited heterogeneity,
underlain by a silty-clay aquitard. Concentrations of dissolved
methane are naturally low (0.02 to 0.2mg l1, where 1 mg l1
of methane equals 0.0027 ccSTP l1) and baseline methane is
biogenic with potential methane oxidation (δ13C-CH4ranges from
80 to 48h). The total injected methane has the capacity to
impact 2,282.5 m3of aquifer on the basis of a solubility of 45.6 mg l1
(Supplementary Information).
Methane migration and fate in the aquifer
Our results show rapid and widespread horizontal migration
of free-gas methane along bedding coupled to buoyancy-
driven upward migration in the Borden aquifer, resulting in
an extensive, continuous, dispersed zone of dissolved methane and
hydrochemistry changes (Figs 2–4 and Supplementary Fig. 2). Total
dissolved gas pressure (PTDG), which qualitatively indicates the sum
of dissolved and free-phase gases in the subsurface26, evolved in
an episodic manner (Fig. 2c) with pressure increases preceding
build-up of [CH4](aq)(dissolved methane concentration). [CH4](aq)
increased across all depths after a pause in gas injection (days 39
to 44), and after the injection stopped on day 72 (Fig. 2b), to
concentrations that indicate explosion risk following ex-solvation
into a confined space (>10 mg l1as the accepted lower explosion
limit)27. Following resumption of methane injection on day 44,
PTDG increased or stabilized while [CH4](aq)decreased below the
explosion risk limit (Fig. 2b,c). This temporary decrease in [CH4](aq)
is potentially a result of displacement of CH4-charged water by, or
mixing with un-impacted groundwater.
The injection of pressurized gas into a permeable geologic unit
typically results in the formation of free-gas migration pathways or
channels28. During gas injection, methane transfer from gaseous to
aqueous phase appears limited during periods of free-gas migration
or channelling29. Following termination of methane injection, free-
gas pressure decreases allow groundwater to re-saturate the pore
space and increase surface area for transfer between gas and aqueous
phases, resulting in the observed [CH4](aq)increases (Fig. 2b).
While free-gas migration channelling is recognized as limiting
subsurface gas retention and efficacy of air-sparging remediation
strategies30, its importance in the migration and fate of fugitive
methane from energy resource development has not yet been
rigorously explored.
The highest concentrations and spatial extent of [CH4](aq)
occurred on day 113 (41 days following termination of gas
injection), with concentrations >25 mg l1(Supplementary Fig. 1).
After 245 days, methane concentrations >10 mg l1were observed
at 2 and 6 m depths, demonstrating persistence of injected methane
(Supplementary Fig. 1). Attenuation of [CH4](aq)and general water
quality impacts were minimal at the time of last sampling, showing
slight increases in alkalinity (4.3 to 5.3 meq l1), decreases in pH
(7.3 to 7.0), and consistent increases in cation concentrations,
including trace metals (for example, Al, Ni, Sr), directly around (that
is, 1 to 2 m distance) the injection zone (Supplementary Table 2).
The evolution of pH, alkalinity and aqueous Ca2+(Supplementary
Fig. 2) in the upper portion of the aquifer suggests that microbial
aerobic methane oxidation is generating CO2, thereby leading to
carbonic acid formation and acid-soluble (for example, carbonate)
mineral dissolution. On the basis of previous CO2experiments31,
carbonic acid-induced groundwater acidification may explain the
increase in trace metals via dissolution and/or ion exchange
processes. Although trace metals did not exceed drinking water
limits over our monitoring period, these increasing trends may be of
more significant concern given expected lifespans of shale gas wells14
or aquifers with different mineralogy. Given that the Borden sand
is relatively inert (comprised primarily of quartz)22, groundwater
© 2017 Macmillan Publishers Limited, part of Springer Nature. All rights reserved.
Injection rate (m3 d−1)
Injected gas (m3)
[CH4](aq) (mg l−1)
6 m
2 m
Pre-injection R1 R2 Post-injectionR3 - 4
−60 −30 0 30 60 90 120 150
PTDG (mmHg)
Time (d)
4 m
8 m
Injection rate
Cumulative volume
Explosion risk limit
Figure 2 | Aqueous methane [CH4](aq) concentrations and total dissolved gas pressure (PTDG) of the groundwater during methane injection and
recovery periods. a, Experimental timeline showing the pre-injection period (49 to 0 days), the injection rates employed (R1–R4) including cumulative
volume of methane injected (0 to 72 days) and the post-injection period (72 to 150 days). b,c, Average [CH4](aq) (b) and PTDG (c) with depth around the
injection zone (that is, samples from points M2, M5, M6 and M7; Fig. 1c and Methods). [CH4](aq) and PTDG evolve in a staggered manner with pressure
increases preceding elevated [CH4](aq). Loss of gas injection pressure induced depth-discrete increases in [CH4](aq).
quality impacts in this experiment will be relatively low compared
with aquifers containing more reactive minerals.
Methane gas efflux was detected at ground surface within hours
of starting the injection. An equally rapid cessation of surface
effluxes was observed when injection was stopped. The surficial CH4
emission area was dependent on the injection rate and resulted in a
‘hotspot’ offset 1m from the injection point (Fig. 3a). Continuous
long-term chambers (LTCs) captured a high degree of temporal
variability in surficial gas effluxes (Fig. 3b). Peak methane efflux
rates at the three LTCs also varied spatially, even though the
measurements were taken only 1 to 2 m apart (Supplementary
Fig. 5). A conservative estimate based on the integration of all
methane efflux rates over the experimental time suggests that at
least 30% of the injected methane was emitted into the atmosphere
(Supplementary Table 3); an additional portion was probably
emitted as CO2after CH4oxidation in the vadose zone. Substantial
13C enrichment of gas-phase methane in the vadose zone resulting
in δ13C values as high as 20 to 15hprovides strong evidence
for microbially mediated methane oxidation.
GPR measurements show that methane gas preferentially
accumulated within a sequence of horizontally layered and
interconnected sand lenses (Supplementary Fig. 6). The maximum
GPR response occurred on day 65 (Fig. 4), where despite limited
heterogeneity in the aquifer, methane migrated much faster
and farther than that predicted by advective groundwater flow.
This period of maximum lateral extent was accompanied by a
systematic decline in [CH4](aq)and PTDG around the injection zone
(Fig. 2b,c); here, rapid lateral expansion of free-phase gas in the
direction of groundwater flow aided by horizontal bedding led
to a temporary reduction in gas pressure and displacement of
CH4-charged groundwater around the injection. These combined
observations suggest that vertical buoyancy-driven methane gas
migration and/or dissolution into groundwater over a period of
days to weeks may influence the lateral extent of free-gas migration;
however, the time and distance scales will depend on site-specific
gas leakage rates and subtle variations in aquifer properties. In
the case of our shallow unconfined aquifer, known to be relatively
homogeneous in the context of single-phase fluid flow, fugitive free-
phase methane spread more extensively than presupposed even for
relatively low leakage rates and short time scales. Lateral migration
will probably be more significant in deeper confined or partially
confined aquifers subject to higher real-world leakage rates and
stronger anisotropy in a multi-formational sedimentary sequence.
The potential for microbial attenuation of the released methane
was investigated by screening the groundwater for methanotrophic
bacteria with 16S ribosomal RNA gene amplicon sequencing.
Prior to methane injection, methanotrophic microorganisms (for
example, Methylococcaceae) were not detected. Methylotrophic
microorganisms, which degrade one-carbon compounds but cannot
oxidize methane (for example, Methylophilaceae), constituted <2%
of the overall community and decreased in abundance with depth
(Fig. 5a). Methane release strongly altered the microbial community
© 2017 Macmillan Publishers Limited, part of Springer Nature. All rights reserved.
56 57 58 59 60 61 62 63 64 65 66 67 68
Continuous CH4 eux
(µmol m−2 s−1)
Time (d)
−3 −2 −1 0 1 2 3 4
−3 −2 −1 0 1 2 3 4
Distance from injection (m)
Distance from injection (m)
Day 68
Day 56
C3 C3
Groundwater flow
CH4 eux (µmol m−2 s−1)
Groundwater flow
C1 C2 C1 C2
Figure 3 | Measured methane eux to atmosphere during injection.
a, Spatial distribution of methane eux on days 56 and 68 of the
experiment. b, Continuous methane eux measurements provided by the
LTCs C1, C2 and C3 between days 56 and 68. Euxes were episodic
throughout the experiment despite constant injection rates during each
injection phase. These spatiotemporal variations in eux demonstrate the
challenges associated with flux measurements and calculations,
particularly those based on standard monitoring methods such as soil gas
surveys, which have a much lower spatial resolution and are conducted
only periodically.
structure (Fig. 5a) and stimulated growth of both aerobic
methanotrophic Methylococcaceae and aerobic methylotrophic
Methylophilaceae despite low oxygen concentrations throughout
the aquifer (Supplementary Fig. 2). These two clades became very
abundant at 2 and 4 m depth, with stable or increasing total cell
numbers of up to 1.1 ×105cells ml1(Fig. 5b) indicating active
populations. Their growth of the two clades also appeared to be
coupled, as noted for oxygen-limited methane oxidation32,33. An
increase in the relative abundance of sulfate-reducing bacteria,
related to Peptococcaceae, indicated that methane release led to
anoxic conditions at 6 and 8m, potentially a result of aerobic
methane oxidation. This population shift was associated with a
sharp decline in total cell counts at 6 m depth. Taxa capable of the
anaerobic oxidation of methane were not detected, which, together
with oxygen limitation, explains why microbial methane oxidation
was relatively ineffective (Supplementary Fig. 2). However, methane
injection strongly and persistently disturbed indigenous microbial
communities, with no sign of recovery after 253 days.
During injection, the δ13C values of dissolved methane in
groundwater rapidly converged towards the carbon isotope ratio
of the injected methane (42h). After 245 days, the δ13C-CH4
remained unchanged and showed little indication of microbial
conversion in the saturated groundwater zone (Supplementary
Fig. 3). Isotope ratios indicate that methane was not significantly
degraded within our monitoring period; however, more significant
degradation, and therefore, potential water quality changes are
expected to occur over longer time scales and with varying leakage
rates, sediment composition and/or redox conditions.
Although the rates and duration of methane injection
(0.17–4.32 m3d1over 72 days) were relatively conservative in
Δlog10(instantaneous amplitude)
Isosurface = 0.429
[GPR response above 5 m depth]
Extended groundwater network
injection points
y-direction (m)
0246810 12 14 16 18 20
x-direction (m)
Depth (m)
GW flow
2 m
4 m
6 m
8 m
Monitored points:
PTDG/CH4 concentration
Surface eux chambers
Figure 4 | GPR response associated with gas-phase methane
accumulation extending downgradient from the injection points. The
iso-surface boundary (orange zone) demarcates a laterally extensive
volume of high-amplitude reflection events on day 65 (Methods and
Supplementary Fig. 6), in the direction of groundwater flow impacted by
methane gas accumulation beneath and along distinct sedimentary
interfaces to a maximum sensing depth of 5 m below ground surface.
Methane gas migrated >17 m downgradient, which is significantly greater
than an estimate of 6 m based on aquifer hydraulic properties
and gradients22.
comparison with documented cases of leakage (Supplementary
Information) and leakage scenarios deemed critical by regulatory
agencies (that is, >300 m3d1)14, the lateral extent of the impacted
groundwater zone relative to the depth of injection was substantial
(4:1), significantly exceeding the observed methane emission
footprint at surface. This large impact zone, as indicated by the
GPR, was facilitated by small-scale geologic layering with subtly
varying permeability, where the horizontal direction parallel to
bedding represents the orientation of highest permeability. Here,
the thin, finer, silty sand layers impeded upward migration of
free gas, thereby trapping and shunting the gas laterally between
coarser, more permeable interconnected sand layers, resulting in
a hydrodynamically dispersed plume consisting of free-phase and
dissolved methane (Fig. 6).
Free and dissolved-phase methane persisted in the groundwater
throughout the 245-day monitoring period. Even a small mass of
injected methane over a relatively short period of time (36.4 kg
over 72 days) persisted due to lack of measurable microbial
degradation in the aquifer volume investigated (600 m3);
although, there was an appreciable perturbation to the indigenous
microbial community in the groundwater zone. Hence, observed
methane attenuation must for the most part be attributable to other
processes; for example, efflux to the vadose zone during injection
followed by advection, dispersion and diffusion.
On the basis of these observations, fugitive methane from a leaky
petroleum well casing may have a strong propensity to migrate hori-
zontally in deeper aquifers (Fig. 6). Here, anisotropy would facilitate
lateral migration and retention of CH4in the groundwater zone,
while buoyancy would enhance dispersion and upward migration
to the vadose zone. The preferential mobility of free-phase gas along
horizontal bedding planes will result in a laterally extensive and dis-
persed methane gas zone extending in the direction of groundwater
flow, accompanied with hydrochemical alterations of groundwater.
We show that both vertical and lateral gas migration mechanisms
are important even when anisotropy is weak and the contribution of
© 2017 Macmillan Publishers Limited, part of Springer Nature. All rights reserved.
Relative abundance (%)
Other bacteria
2 m depth 6 m depth 8 m depth4 m depth
Microbial population
(cells × 105 ml−1)
Time (d)
Time (d)
Figure 5 | Microbial response to injected methane gas in groundwater
samples. a, Relative sequence abundances of bacterial taxa from
monitoring well M6 (1 m downgradient of injection zone) determined by
16S rRNA gene amplicon sequencing. Methane was injected from
day 0 to 72 (dashed line). Methanotrophs are shown in red, methylotrophs
in orange, and sulfate reducers in yellow. Other taxa possibly involved in the
cycling of sulfur and iron are shown in blue and brown, respectively. b, Total
microbial cell counts at monitoring well M6 shown as average counts of ten
independent microscopic fields; error bars are within two standard
deviations of the mean.
each to groundwater flow will be site specific, depending primarily
on anisotropy of the geologic medium. In bedrock units, fracture
network conditions (aperture, orientation, length) will control the
anisotropy and direction and magnitude of fluxes.
Regional groundwater flow systems
In sedimentary rocks, which are known to have very small effective
porosities, advection of fluids is controlled by fractures and their
connectivity. Therefore, for a given gas leakage rate, free gas will
migrate much farther and faster in fractured sedimentary rock
compared with a granular sand aquifer21 for a given specified
gas injection (leakage) rate. Furthermore, the likelihood of leaks
occurring deeper below multiple stratigraphic layers would be
accompanied by stronger anisotropy, enhancing the rate of lateral
migration parallel to bedding compared with upward migration
orthogonal to bedding, but strongly influenced by the presence of
long, through-going vertical pathways, possibly faults.
Current groundwater monitoring requirements in areas of
energy resource development have perceived groundwater impacts
to methane leakage as secondary to atmospheric emissions.
However, our study shows that gas-phase migration from a high-
pressure source in the saturated zone is an equivalent, if not, more
significant process relative to atmospheric emissions. Present-day
monitoring efforts do not consider the groundwater resource in its
entirety and involve only periodic sampling from existing, sparsely
located domestic wells, which serve as receptors at risk, rather than
adequate monitors for groundwater resource impact evaluation.
Current surface and subsurface monitoring efforts for shale gas
development are thus insufficient to meaningfully detect or assess
methane impacts to atmosphere and groundwater.
Petroleum well
Clay aquitard
Vadose zone
sand aquifer
Fine sand/silt lense
Groundwater flow
Free-phase methane flux
Free-phase methane
Dissolved-phase methane
50 m
CH4 + 2O2
CO2 + 2H2O
sand layers
Figure 6 | Conceptual model of a continuous methane leak in an
unconfined freshwater aquifer. (1) A compromised energy well casing
releases methane; (2) laterally extensive free-phase gas migration
controlled by subtle geologic variability; (3) variable methane eux at the
surface with potential for explosion risk and greenhouse gas emission; and
(4) a temporally persistent dissolved methane plume. Aerobic oxidation
of methane occurs in the vadose zone via microbes, but degradation is
negligible in the saturated zone over 245 days. While vertical and lateral
gas migration occurs simultaneously the relative magnitude of fluxes
and style of resultant free-gas plume will depend on site-specific
geologic conditions.
Our experimental field results show that subtle variations in
aquifer properties can have a major impact on fugitive methane
distribution, and that surface flux to the atmosphere is not indicative
of the lateral extent of methane-induced groundwater quality
impacts. Current efforts by oil and gas companies to locate and
mitigate casing vent flow leaks or breaches in casings would identify
entry points into the subsurface, and thus, facilitate groundwater
impact monitoring strategies. Most importantly, we demonstrate
that meaningful characterization and effective monitoring of
fugitive methane in freshwater aquifers can be achieved through
a multi-disciplinary and multi-scale monitoring network that
intercepts and delineates a fugitive methane contaminant plume
emanating from a leaky petroleum well.
Methods, including statements of data availability and any
associated accession codes and references, are available in the
online version of this paper.
Received 6 October 2016; accepted 20 February 2017;
published online 27 March 2017
1. Malakoff, D. The gas surge. Science 344, 1464–1467 (2014).
2. Vidic, R. D., Brantley, S. L., Vandenbossche, J. M., Yoxtheimer, D. & Abad, J. D.
Impact of shale gas development on regional water quality. Science 340 (2013).
3. Jackson, R. E. et al. Groundwater protection and unconventional gas
extraction: the critical need for field-based hydrogeological research. Ground
Water 51, 488–510 (2013).
4. Dusseault, M. B., Gray, M. N. & Nawrocki, P. A. Why Oilwells Leak: Cement
Behavior and Long-Term Consequences (Society of Petroleum Engineers, 2000).
5. Dusseault, M., Jackson, R. & MacDonald, D. Towards a Road Map for
Mitigating the Rates and Occurrences of Long-Term Wellbore Leakage Special
Report (Univ. Waterloo, 2014).
6. Osborn, S. G., Vengosh, A., Warner, N. R. & Jackson, R. B. Methane
contamination of drinking water accompanying gas-well drilling and hydraulic
fracturing. Proc. Natl Acad. Sci. USA 108, 8172–8176 (2011).
© 2017 Macmillan Publishers Limited, part of Springer Nature. All rights reserved.
7. Darrah, T. H., Vengosh, A., Jackson, R. B., Warner, N. R. & Poreda, R. J. Noble
gases identify the mechanisms of fugitive gas contamination in drinking-water
wells overlying the Marcellus and Barnett Shales. Proc. Natl Acad. Sci. USA 111,
14076–14081 (2014).
8. Heilweil, V. M. et al. Stream measurements locate thermogenic methane fluxes
in groundwater discharge in an area of shale-gas development. Environ. Sci.
Technol. 49, 4057–4065 (2015).
9. Heilweil, V. M. et al. A stream-based methane monitoring approach for
evaluating groundwater impacts associated with unconventional gas
development. Ground Water 51, 511–524 (2013).
10. Humez, P. B. et al. Occurrence and origin of methane in groundwater in
Alberta (Canada): gas geochemical and isotopic approaches. Sci. Total Environ.
541, 1253–1268 (2016).
11. Roy, N., Molson, J., Lemieux, J.-M., Van Stempvoort, A. & Nowamooz, A.
Three-dimensional numerical simulations of methane gas migration from
decommissioned hydrocarbon production wells into shallow aquifers. Wat.
Resour. Res. 52, 5598–5618 (2016).
12. Shale Gas Extraction in the UK: A Review of Hydraulic Fracturing DES2597
(The Royal Society, The Royal Academy of Engineering, 2012).
13. Environmental Impacts of Shale Gas Extraction in Canada. The Expert Panel on
Harnessing Science and Technology to Understand the Environmental Impacts of
Shale Gas Extraction (Council of Canadian Academies, 2014).
14. Nowamooz, A., Lemieux, J. M., Molson, J. & Therrien, R. Numerical
investigation of methane and formation fluid leakage along the casing of a
decommissioned shale gas well. Wat. Resour. Res. 51, 4592–4622 (2015).
15. Ingraffea, A. R., Wells, M. T., Santoro, R. L. & Shonkoff, S.B. C. Assessment and
risk analysis of casing and cement impairment in oil and gas wells in
Pennsylvania, 2000–2012. Proc. Natl Acad. Sci. USA 111, 10955–10960 (2014).
16. Caulton, D. R. et al. Toward a better understanding and quantification of
methane emissions from shale gas development. Proc. Natl Acad. Sci. USA 111,
6237–6242 (2014).
17. Jackson, R. B. et al. Increased stray gas abundance in a subset of drinking water
wells near Marcellus shale gas extraction. Proc. Natl Acad. Sci. USA 110,
11250–11255 (2013).
18. Molofsky, L. J., Connor, J. A., Wylie, A. S., Wagner, T. & Farhat, S. K. Evaluation
of methane sources in groundwater in northeastern Pennsylvania. Ground
Water 51, 333–349 (2013).
19. Jackson, R. E. & Heagle, D. J. Sampling domestic/farm wells for baseline
groundwater quality and fugitive gas. Hydrogeol. J. 24, 269–272 (2016).
20. Siegel, D. I., Azzolina, N. A., Smith, B. J., Perry, A. E. & Bothun, R. L.
Methane concentrations in water wells unrelated to proximity to existing oil
and gas wells in northeastern Pennsylvania. Environ. Sci. Technol. 49,
4106–4112 (2015).
21. Parker, B. L., Cherry, J. A. & Chapman, S. W. Discrete fracture network
approach for studying contamination in fractured rock. AQUA Mundi 3,
101–116 (2012).
22. Cherry, J. A., Gillham, G. W., Anderson, E. G. & Johnson, P. E. Migration of
contaminants in groundwater at a landfill: A case study: 2. Groundwater
monitoring devices. J. Hydrol. 63, 31–49 (1983).
23. Sudicky, E. A. & Illman, W. A. Lessons learned from a suite of CFB Borden
experiments. Ground Water 49, 630–648 (2011).
24. Tomlinson, D. W., Thomson, N. R., Johnson, R. L. & Redman, J. D. Air
distribution in the Borden aquifer during in situ air sparging. J. Contam.
Hydrol. 67, 113–132 (2003).
25. Cahill, A. G., Marker, P. & Jakobsen, R. Hydrogeochemical and mineralogical
effects of sustained CO2contamination in a shallow sandy aquifer: a field-scale
controlled release experiment. Wat. Resour. Res. 50, 1735–1755 (2014).
26. Manning, A. H., Kip Solomon, D. & Sheldon, A. L. Applications of a total
dissolved gas pressure probe in ground water studies. Ground Water 41,
440–448 (2003).
27. Eltschlager, K. K., Hawkins, J. W., Ehler, C. & Baldassare, F. Technical Measures
for the Investigation and Mitigation of Fugitive Methane Hazards in Areas of
Coal Mining (Office of Surface Mining Reclamation and Enforcement, 2001).
28. Thomson, N. R. & Johnson, R. L. Air distribution during in situ air sparging:
an overview of mathematical modeling. J. Hazar. Mater. 72, 265–282 (2000).
29. McCray, J. E. Mathematical modeling of air sparging for subsurface
remediation: state of the art. J. Hazards Mater. 72, 237–263 (2000).
30. Johnson, P. C., Das, A. & Bruce, C. Effect of flow rate changes and pulsing on
the treatment of source zones by in situ air sparging. Environ. Sci. Technol. 33,
1726–1731 (1999).
31. Cahill, A. G., Jakobsen, R., Mathiesen, T. B. & Jensen, C. K. Risks attributable to
water quality changes in shallow potable aquifers from geologic carbon
sequestration leakage into sediments of variable carbonate content. Int. J.
Greenhouse Gas Control. 19, 117–125 (2013).
32. Beck, D. A. C. et al. A metagenomic insight into freshwater methane-utilizing
communities and evidence for cooperation between the Methylococcaceae and
the Methylophilaceae. PeerJ 1, e23 (2013).
33. Oshkin, I. Y. et al. Methane-fed microbial microcosms show differential
community dynamics and pinpoint taxa involved in communal response.
ISME J. 9, 1119–1129 (2015).
We acknowledge funding from the Natural Sciences and Engineering Research Council
of Canada (NSERC), Strategic Project Grant no. 463045-14. Additional support was
provided by the AITF/Eyes High Postdoctoral Fellowship programme (S.E.R.), the
Campus Alberta Innovation Chair Program (M.S.), the Canadian Foundation for
Innovation (M.S.), the Alberta Small Equipment Grant Program (M.S.), NSERC
Discovery Grant (M.S.) and a Banting Postdoctoral Fellowship (C.M.S.). Special thanks
for field assistance are given to A. Haggman, B. Ladd, D. Klazinga, T. Cheung, A. Verdin,
R. Ingleton and P. Johnson.
Author contributions
A.G.C., B.L.P. and J.A.C. conceived, designed, installed and oversaw the experiment.
A.G.C., C.M.S., O.F., O.K. and S.E.R. collected and processed all field data. All authors
interpreted the multi-disciplinary data sets. A.G.C., B.L.P., J.A.C. and C.M.S. generated
the first draft of the manuscript before all authors contributed to refinement and
Additional information
Supplementary information is available in the online version of the paper. Reprints and
permissions information is available online at Publisher’s note:
Springer Nature remains neutral with regard to jurisdictional claims in published maps
and institutional affiliations. Correspondence and requests for materials should be
addressed to B.L.P.
Competing financial interests
The authors declare no competing financial interests.
© 2017 Macmillan Publishers Limited, part of Springer Nature. All rights reserved.
Injection system configuration and methodology. The gas delivery system was
composed of two, 45-inclined sparging wells emplaced at 4.5 and 9 m depths
(termed shallow and deep). Each sparging well was constructed using a 1 m length
of porous polyethylene (PE) well screen (SCHUMASOIL) connected to a mass flow
controller and two canisters (master and slave) of industrial-grade methane by PE
tubing. Injection screens were installed using a Geoprobe (model 7822DT) direct
push system, along a vertical plane perpendicular to groundwater flow. Borden
sediment is not freestanding following use of a Geoprobe direct push tool; drilled
holes collapse almost immediately following removal of drill rods. Angled injection
wells were used to minimize migration of methane along the delivery tube whilst
maximizing gas release into the aquifer. Injection wells were connected to a control
unit containing two individual (shallow and deep) mass flow controllers (Red-y
smart GSC-C9SA-BB26) controlled by associated software (Get Red-y, Vögtlin
Instruments AG).
Gas injection was conducted in four phases at three rates as shown in
Supplementary Table 1. Total gas volume injected during the experiment was
51,350 l of industrial grade methane (34.9kg at 1.013 bar and 15 C). Following
injection, gas delivery wells were sealed by non-return valves to prevent immediate
flow back.
Aqueous chemistry sampling. Groundwater samples were collected through 32
multilevel groundwater sampling systems (M1–M32) comprising a total of 112
depth-discrete sample points (Fig. 1b). Sampling systems were constructed using
6 mm (OD) PE tubing connected to 5 cm length, geotextile-covered slotted PE
screens (Nitex monofilament screen cloth, 200 µm mesh), fixed at multiple depths
to a single polyvinyl chloride (PVC) centre-stock. Multilevel systems were
instrumented with either three screens (2, 5 and 8m depths), four screens (2, 4, 6
and 8 m depths) or five screens (2, 3, 4, 5 and 6 m depths) depending on the
location and proximity to the injection point. The groundwater monitoring
network covered an area of approximately 72 m2, spanning 10 m downgradient of
the injection zone. Multilevels were installed using a Geoprobe (model 7822DT)
direct push system at higher density around and immediately downgradient of the
injection zone at required depths. A plan view of the groundwater multilevel
network sample points is shown in Fig. 1b.
Physical chemistry parameters including electrical conductivity (accuracy
1µS cm1), pH (accuracy 0.02) and total dissolved gas pressure (1mmHg) were
measured with a multi-parameter water-quality sonde and flow-through cell
(Manta 2, Eureka Water Probes) using a peristaltic pump at constant, low flow rate
(approximately 40 ml min1). Dissolved gas samples were collected by filling either
biocide-treated pre-evacuated 125 ml glass vials or 40ml VOA vials, submerged in
a beaker and piercing or sealing with a septa cap, respectively (leaving no
headspace). Samples were stored upside down at 4 C and measured on a Bruker
450 Natural Gas Analyzer for dissolved hydrocarbons (C1–C3) and other dissolved
gases within 7 days. Major and minor cation concentrations were determined
following filtration (0.45 µm cellulose acetate filters) and acidification for cation
samples (concentrated HNO3) by ion chromatography or inductively coupled
plasma-mass spectrometry (ICP-MS; PerkinElmer Elan 6100DRC). Accuracy of
ICP-MS measurements is 5% with limits of detection (mmol l1) for parameters
shown as follows: Al, 1.11; Ba, 0.14; Ca, 3.74; Fe, 0.89; K, 1.27; Mg, 0.82; Mn, 0.036;
Na, 2.17; Si, 3.55; Sr, 0.03; and Zn, 0.15. Alkalinity was determined (by Gran
titration) for selected samples.
Experimental field site. The experimental site was characterized
hydro-geochemically prior to methane injection through a series of depth-discrete
background samples (three sampling events over 4 months prior to methane
injection). A historic construction waste landfill leachate plume exists at the base of
the aquifer (that is, 8–9 m depth)34 creating two distinct aqueous chemistries. The
upper zone, from the water table down to approximately 6–7 m depth, features
sub-oxic groundwater (that is, negative Eh with low to no dissolved O2) with
moderate total dissolved solids (250–450 mgl1) and electrical conductivity
(400–600 µS cm1) in the Fe/Mn reduction stage of the redox sequence and which
is not currently impacted by landfill leachate. At 8m depth, landfill leachate
impacts the groundwater chemistry as evidenced by a significant increase in total
dissolved solids (1,100–1,500 mgl1) and electrical conductivity
(1,800–2,000 µS cm1) caused primarily by increases in concentrations of major
ions and alkalinity. Concentrations of dissolved methane are naturally low
(0.02–0.2 mg l1), the methane is of biogenic origin (δ13C-CH4as low as 80h),
and δ13C-CH4values occasionally as high as 48hsuggest some occurrence of
methane oxidation in the aquifer. Total dissolved gas pressure (PTDG) is variable
across the site, ranging from 700 to 860 mmHg (standard deviation 40–50 mmHg).
Mean aqueous chemical conditions from three background samples taken 50,
21 and 2 days prior to injection are shown in Supplementary Table 2. The
water table at the experimental site was approximately 1m below ground surface
during the duration of the experiment, varying ±0.5 m from spring to
autumn, respectively.
Geophysical data collection and data processing. Ground-penetrating radar
(GPR) measurements were collected using a PulseEKKO 100 GPR system (Sensors
& Software) equipped with 200 MHz bistatic antennas and a 1,000 V transmitter.
Reflection profiles were acquired along a series of parallel 20-m-long lines spaced
0.5 m apart, and orientated parallel to groundwater flow direction; these lines were
connected by orthogonal tie lines positioned at 0.5, 8.5, 12, 15 and 18m
downgradient (Fig. 1c). Radar traces were recorded using a spatial step size of 0.1m
with the antennas spaced 1 m apart and orientated perpendicular to the survey line
direction. A 64-trace stack was used with a temporal sampling interval of 800 ps.
Data were recorded using a manual trigger and measuring tapes.
Post-acquisition processing and visualization were performed using ReflexW
(v.7) (Sandmeier Software). The processing flow consisted of a high-pass mean
dewow filter to remove d.c. shifts on individual traces; zero-time corrections; a
fixed gain function to compensate for attenuation; a bandpass frequency filter
(25–75–250–325 MHz); a notch frequency filter (40–85–85–125 MHz) to suppress
ringing (for example, near-surface efflux chambers and electrical wires); and a
three-trace horizontal filter. The instantaneous amplitude, a complex trace attribute
that removes the effects of wavelet polarity and phase35, was calculated to more
accurately measure changes in reflection strength through the injection period.
Here, the instantaneous amplitude provided a measure of changes in the dielectric
contrast along closely spaced bedding planes. The high dielectric constant of water
(κ=81) relative to mineral grains (κ=4–6) and air/gas (κ=1) (ref. 36) results in a
high reflection coefficient at the interface between contrasting stratigraphic units.
In this study, the vertical and lateral migration of free-phase methane within the
sand aquifer will be impeded at interfaces demarcating subtle changes in capillary
properties. Subsequent accumulation of methane gas along these capillary
boundaries will lead to a reduction in water saturation just below the interface as
gas displaces water; this gas will continue to desaturate the pore space until the
capillary pressure exceeds the air-entry pressure of the overlying sediment, upon
which the gas will pass through37. Spatiotemporal GPR amplitude variations along
sedimentary interfaces were used to assess gas accumulation along stratigraphic
bedding planes (Supplementary Fig. 6).
The spatial distribution of energy (that is, instantaneous amplitude) was
calculated over a 20 ×7 m area using the reflection profiles shown in Fig. 1c.
A three-dimensional cube (20 ×7×6 m) of the instantaneous amplitude was
computed using an inverse distance weighting method based on a uniformly
discretized 0.25 (x,y,z) grid, assuming a fixed radar velocity of 0.06 mns1
(ref. 24).
Surficial efflux monitoring methods. Continuous methane effluxes were
monitored with long-term LI-COR chambers (LI-8100-104) connected to a
LI-COR Multiplexer (LI-8150) (LI-COR), cavity ring down laser spectroscopy, and
extended range, Ultraportable Greenhouse Gas Analyzer (UGGA) (Los Gatos
Research). The spatial distribution of gas effluxes was measured with a portable
LI-COR chamber (LI-8100-103) connected to the UGGA38.
Flux measurements were completed by placing chambers on PVC collars
(0.2 m ID) inserted into the soil prior to commencing injection. The portable
chamber was used in a 5 ×6 m grid and three long-term chambers were placed
directly above (Chamber 1) and offset by 1–2 m (Chambers 2 and 3) from the
methane injection (Supplementary Fig. 4). All chamber measurements lasted 3min
and an automated cycle switched between chambers every 10 min. Effluxes were
calculated and are reported for selected times in micromoles per square meter per
second (LI-COR, 2015).
Stable carbon isotope ratio measurements. Multilevel soil gas monitoring wells
were installed (Supplementary Fig. 4) for periodic collection of soil gas samples and
subsequent stable carbon isotope (δ13C) ratio analyses of methane and CO2. Ports
were installed at depths of 10, 30 and 50cm. Wells were constructed of 1/800
gas-impermeable tubing with mesh screens at the bottom and gas-tight fitting and
septa at the top. Samples were collected with gas-tight syringes (Valco Instruments)
and stored in pre-evacuated 12 ml vials (Labco). Gas samples for isotope analyses
were transferred directly using a gas-tight syringe (CH4>3,000ppmv) or via a
pre-concentration device (CH4<3,000ppmv) into a Thermo Scientific Trace GC
Ultra where methane was isolated and subsequently quantitatively converted to
CO2at 1,000 C using a GC-Isolink. The CO2was subsequently introduced into a
Thermo Finnigan MAT253 isotope ratio mass spectrometer via a Conflo IV
system, and carbon isotope ratios of methane were determined in continuous flow
mode. Carbon isotope ratios of methane are reported in the standard delta notation
(h) relative to VPDB (Vienna PeeDee Belemnite) with a precision of better than
±0.5h.δ13C values of CO2were determined with the same system with a
precision of better than ±0.3h.
Microbiological sample filtration. Groundwater from the Borden aquifer for
DNA analyses was collected in 1 l Nalgene HDPE bottles, which were completely
filled and shipped in iced coolers to the Energy Bioengineering and
Geomicrobiology laboratories at the University of Calgary. On arrival (5–7 days
© 2017 Macmillan Publishers Limited, part of Springer Nature. All rights reserved.
after sampling), a 250 ml aliquot of each sample was filtered through a 0.22µm
GTTP Isopore membrane filter (Merck Millipore) attached to a sterile 500 ml
polysulfone bottle top filter holder (Nalgene). The supernatant was collected in
sterile 250 ml Schott bottles and subsequently filtered through a 0.1 µm Supor-100
membrane filter (Pall). The filters were stored at 20 C until DNA extraction.
Microbial cell numbers. Unfiltered groundwater (5ml) was fixed with sterile
formaldehyde (final concentration, 4% v/v) and filtered through a sterile 0.1 µm
Merck Millipore VCTP membrane filter. Cell counts from well M6 were performed
using a DAPI (40,6-diamidino-2-phenylindole) stain39 and fluorescence
microscopy. Direct cell counts were performed and averaged for 10 fields of view.
DNA isolation and quantification. The 0.22 and 0.1µm filters from each sample
were cut into thin strips with a sterile scalpel and transferred in duplicate into
autoclaved 2 ml bead beating tubes. Each tube contained 0.5g of 0.1 mm and 0.5 g
of 2.5 mm zirconia–silica beads (BioSpec Products). High-molecular-weight DNA
was extracted from each filter using an SDS–chloroform protocol as follows. First,
300 µl each of 0.1 M phosphate buffer (pH 8), 10% SDS lysis buffer (pH 8), and 24:1
chloroform/isoamyl alcohol solution were added in sequence to the tubes. The
tubes were loaded onto a bead mill homogenizer (Bead Ruptor 24, Omni) for 45 s
at speed setting 5.5. The tubes were centrifuged at 20,000gfor 5min and the
supernatant was transferred to a sterile 1.5 ml microcentrifuge tube. Ammonium
acetate (7 mol l1) was then added to achieve a final concentration of 2.5 M. The
1.5 ml tubes were inverted gently by hand for 10s and centrifuged at 20,000gfor
7 min. The supernatant was transferred into a new sterile 1.5 ml microcentrifuge
tube to which 0.54 ml of 100% isopropanol was added, then stored at 20 C
overnight for DNA precipitation. DNA was pelleted by centrifugation at 4 C for
30 min. The isopropanol was decanted off and the tubes dried inside a biosafety
cabinet for up to 2h. The DNA in each set of duplicate tubes was re-suspended and
pooled by dissolving the pellet in 30 µl of sterile nuclease-free water. DNA
quantification was performed with a Qubit 2.0 Fluorometer (Thermo Fischer
Scientific). DNA samples were stored at 20 C. The constitution of 10% SDS lysis
buffer (100 ml) was as follows: 48ml MilliQ water, 2 ml of 5M sodium chloride,
50 ml Tris solution (pH 8), and 10 g SDS. With the exception of isopropanol and
chloroform/isoamyl alcohol solution, all reagents were filter-sterilized with a
0.22 µm syringe filter.
DNA amplification and Illumina 16S rDNA amplicon sequencing. The V3–V4
region of 16S rDNA was amplified in a single-step PCR using a KAPA HiFi
HotStart reaction kit and primers Pro341F/Pro805R targeting all prokaryotes40 and
including Illumina adapters (Illumina). The primers were complementary to
standard Illumina forward and reverse primers. The reverse primer also contained
a 6-bp indexing sequence. PCR reaction conditions for DNA amplification were as
follows: initial denaturation at 95C for 3min, followed by 32 cycles of 95 C for
30 s, 55 C for 45 s, and 72 C for 60 s, followed by a final step of 72 C for 5min.
The amplicon products of triplicate PCR reactions were pooled and purified using
AMPure XP beads (Agencourt Bioscience). DNA amplicon libraries were prepared
from purified DNA using a Nextera XT DNA Sample Prep Kit with Nextera XT
Index Kit (Illumina) as per the manufacturer’s instructions. PCR conditions were as
follows: 95 C for 3 min, followed by 10 cycles of 9 C for 30s, 5 C for 45 s, and
72 C for 60 s, and a final step of 72 C for 5min. PCR products were purified with
AMPure XP beads and quantified with a Qubit 2.0 Fluorometer. Amplicon libraries
were normalized to 2 nM, pooled in equal volumes, denatured in 0.2 N NaOH, and
diluted with hybridization buffer according to Nextera XT protocol. Paired-end
sequencing (300 ×300 bp) of libraries at 15 pM final concentration was performed
on an Illumina Miseq instrument using the manufacturer’s reagents and according
to the manufacturer’s instructions.
Analyses of 16S rDNA amplicon data. Paired-end reads (300 ×300 bp) were
merged with usearch (3) (‘-fastq_mergepairs -fastqminovlen 100-fastqmaxdiffs
8-fastq_allowmergestagger -fastq_truncqual 3’), followed by trimming of primers
with Mothur41 (‘trim.seqs pdiffs=0’) and read quality control with usearch
(‘-fastq_filter -fastq_trunclength 350 -fastq_maxee 1’). Dereplication, removal of
singleton and/or chimaeric reads, and clustering was done with the uparse pipeline
based on 97% sequence identity42,43. Taxonomic assignment was done with the RDP
classifier implemented in Mothur with the SILVA training data set as the template
( Our approach is available
online at
Data availability. The authors declare that the data supporting the findings of this
study are available within the article and its Supplementary Information. DNA
sequence data are available in the NCBI database (
(SRA accession SRP076566, Bioproject accession: PRJNA32558).
34. MacFarlane, D. S., Cherr y, J. A., Gillham, R. W. & Sudicky, E. A. Migration of
contaminants in groundwater at a landfill: a case study: 1. Groundwater flow
and plume delineation. J. Hydrol. 63, 1–29 (1983).
35. Sheriff, R. E. & Geldart, L. P. Exploration Seismology (Cambridge Univ. Press,
36. Cassidy, N. J. Ground Penetrating Radar: Theory and Applications
(ed. Jol, H. M.) 41–72 (Elsevier, 2009).
37. Thomson, N. R. & Johnson, R. L. Air distribution during in situ air sparging:
an overview of mathematical modeling. J. Hazards Mat. 72, 265–282 (2000).
38. Porter, K. G. & Feig, Y. S. The use of DAPI for identifying and counting aquatic
microflora. Limnol. Oceanogr. 25, 943–948 (1980).
39. Sihota, N. J., Mayer, K. U., Toso, M. A. & Atwater, J. F. Methane emissions and
contaminant degradation rates at sites affected by accidental releases of
denatured fuel-grade ethanol. J. Contam. Hydrol. 151, 1–15 (2013).
40. Takahashi, S., Tomita, J., Nishioka, K., Hisada, T. & Nishijima, M. Development
of a prokaryotic universal primer for simultaneous analysis of bacteria and
archaea using next-generation sequencing. PLos ONE 9, e105592 (2014).
41. Schloss, P. D. et al. Introducing mothur: open-source, platform-independent,
community-supported software for describing and comparing microbial
communities. Appl. Environ. Microbiol. 75, 7537–7541 (2009).
42. Edgar, R. C. Search and clustering orders of magnitude faster than BLAST.
Bioinformatics 26, 2460–2461 (2010).
43. Edgar, R. C. UPARSE: highly accurate OTU sequences from microbial
amplicon reads. Nat. Methods 10, 996–998 (2013).
© 2017 Macmillan Publishers Limited, part of Springer Nature. All rights reserved.
... The buoyant properties of FPG, caused by its low density compared to water, can cause FPG to migrate in complex pathways irrespective of groundwater flow [1]. In response to these FPG characterization difficulties and the lack of observation-supported studies, recent studies have employed geophysical [6][7][8] or light transmission [9] methods to image the migration pathways of free methane gas from a controlled injection point, within sandy aquifers. These studies observed lateral spreading of free-phase methane that was strongly controlled by layers with relatively low permeability-even by thin lenses with slightly finer grain size. ...
... These studies observed lateral spreading of free-phase methane that was strongly controlled by layers with relatively low permeability-even by thin lenses with slightly finer grain size. Free-phase methane was observed to migrate laterally further than expected by groundwater advection alone [6,7], and FPG plumes were temporally persistent below capillary barriers [6][7][8][9]. ...
... These studies observed lateral spreading of free-phase methane that was strongly controlled by layers with relatively low permeability-even by thin lenses with slightly finer grain size. Free-phase methane was observed to migrate laterally further than expected by groundwater advection alone [6,7], and FPG plumes were temporally persistent below capillary barriers [6][7][8][9]. ...
Full-text available
Geophysical imaging of free-phase gas (FPG) within aquifers is an emerging method for understanding the mechanisms controlling stray gas migration from oil and gas wells. Crystal Geyser is an unsealed and partially cased well that transports stray CO2 gas to the shallow subsurface. Accumulations of subsurface CO2 FPG near Crystal Geyser have been inferred, but the actual location and dimensions remained unclear. Here, the subsurface FPG distribution surrounding Crystal Geyser was characterized by interpreting 2D electrical resistivity images with previous drilling records and field mapping. An approximately 70-metre-wide FPG plume was located laterally between Crystal Geyser’s conduit and the Little Grand Wash Fault. The FPG plume spanned the vertical extent of approximately 20 to 55 metres below the ground surface, located within the Slick Rock Member sandstone with the relatively low permeability Earthy Member silty sandstone acting as a caprock. The FPG plume was identified from an anomalously high resistivity zone within the Slick Rock Member that was not caused by lateral lithofacies changes or fault displacement. The conceptual FPG migration pathways beneath Crystal Geyser are presented, based on the interpreted FPG distribution from the electrical resistivity images combined with previous site characterization and the principles of buoyant FPG migration. FPG accumulates within the Slick Rock Member by buoyant up-dip migration beneath siltstone capillary barriers of the Earthy Member. FPG leaks to the ground surface within high permeability preferential pathways along the Little Grand Wash Fault and the conduit of Crystal Geyser.
... Well integrity failures, including surface casing vent flow (SCVF) and GM outside the outermost (or surface) casing, represent safety, environmental, and financial liabilities to the upstream oil and gas industry and negatively affect the oil and gas industry's social license (Alboiu and Walker 2019;Cahill et al. 2017;. Wells detected with SCVF or GM cannot be decommissioned legally in Canada; therefore, appropriate GM detection informs operational decision making on remedial cementing, with important environmental and social consequences, and financial implications [Alberta Energy Regulator (AER) 2014; Schiffner et al. 2021;Trudel et al. 2019]. ...
... Temporal variation in gas concentrations and effluxes may be driven by episodic and pulsed movement of gas in the saturated zone (Cahill et al. 2017;Van De Ven et al. 2020) and due to changing atmospheric conditions (Kuang et al. 2013;Oliveira et al. 2018). Barometric pressure changes are also known to induce variable effluxes and may cause atmospheric gases to flow into the soil during rising barometric pressures due to a pressure imbalance between atmospheric and soil gases (Abbas et al. 2010;Forde et al. 2019). ...
... A similar field-scale gas injection experiment has been performed with CO 2 at our field site, 9 whereas a CH 4 injection field experiment has been recently carried out in a Canadian aquifer. 10 The distribution of injected CO 2 or CH 4 in these shallow aquifers was monitored by δ 13 C measurements. 9,10 The migration of the CO 2 plume at our field site was reported to be slow. ...
... 36−38 When the injection and, subsequently, supply of gas is halted, these channels collapse, as observed in a CH 4 injection experiment. 10 At the field site, H 2 was injected with over pressure at about 18 m bgl to generate a coherent gas phase and dissolved H 2 plume (cf. section SI-1 of the Supporting Information). ...
... Recent research has focused on understanding specific geochemical and hydrogeological aspects of natural gas migration, from well bore integrity (Wigston et al. 2019), to the study of the migration and fate of injected gas in shallow aquifers (Forde et al. 2019;Cahill et al. 2017;2018;. While repairing a petroleum well with SCVF remains an important engineering challenge (Watson 2004;Watson and Bachu 2009;, geochemical methods (Tilley and Muehlenbachs 2012;Darrah et al. 2014;Jackson et al. 2013;Mayer et al. 2015;McIntosh et al. 2019) can be used along with wireline well logging methods (Aslanyan et al. 2015) to identify the depth at which natural gas enters the well annulus. ...
Full-text available
Surface casing vents divert natural gas migration along oil and gas boreholes to bypass groundwater, with the gas venting to the atmosphere. While this strategy is designed to protect groundwater, it constitutes a source of greenhouse gases to the atmosphere. In instances where gas leakage occurs, the characterization of the molecular and isotopic composition of natural gas emitted from surface casing vent flows can be used to assist in identifying the gas source. We compare concentration measurements of non-hydrocarbon gases (within natural gas) of samples analyzed by laboratory-based gas chromatography (N 2 , Ar, CO 2 and O 2 ) and magnetic sector noble gas mass spectrometry (He, Ar and Kr) with field measurements conducted using a field portable quadrupole mass spectrometer (miniRUEDI). The standard deviation of miniRUEDI concentration results was within plus/minus one standard deviation of samples measured using laboratory-based GC (N 2 , O 2 , Ar and He) and magnetic sector noble gas mass spectrometry (He, Ar). Additional laboratory-based determination of isotope ratios of methane and argon (δ ¹³ C CH4 , δ ² H CH4 , and ⁴⁰ Ar/ ³⁶ Ar) enabled a comparison between information provided by the analysis of reactive gases compared with noble gas isotopes. Gases from different sources displayed quantifiable differences in δ ¹³ C CH4 and δ ² H CH4 , but these changes may or may not be distinguished if only one sampling event is conducted. By comparison, ⁴⁰ Ar/ ³⁶ Ar further enabled the differentiation of various gas sources. The objective of this paper is to discuss the advantages and trade-offs of the three different analysis methods considered, and the feasibility of their application in different environmental monitoring scenarios.
... The coexistence of these guilds was previously observed in a methane-rich shallow aquifer (37) as well as in methylotrophic microbiomes of mesocosms (38). Aerobic methane oxidizers of the genus Crenothrix (39) were the fourth most abundant and occurred in almost half of the samples (45 of 109, Fig S8). ...
Full-text available
Groundwater ecosystems are globally widespread yet still poorly understood. We investigated the age, aqueous geochemistry, and microbiology of 138 groundwater samples from 87 monitoring wells (<250m depth) located in 14 aquifers in the Canadian Prairie. Geochemistry and microbial ecology were tightly linked revealing large-scale aerobic and anaerobic hydrogen, methane, nitrogen, and sulfur cycling carried out by diverse microbial communities. Older groundwaters contained on average more cells (up to 1.4×10 ⁷ /mL) than younger groundwaters. Organic carbon-rich strata featured some of the highest abundances, challenging current estimates of global groundwater population sizes. Substantial concentrations of dissolved oxygen (n=57; 0.52±0.12 mg/L [mean ± SE]; 0.39 mg/L [median]) in older groundwaters could support aerobic lifestyles in subsurface ecosystems at an unprecedented scale. Metagenomics, oxygen isotope analyses and mixing models indicated that microbial ″dark oxygen″ contributed to the dissolved oxygen pool in subsurface ecosystems commonly assumed to be anoxic.
... The extent of GM is influenced by a multitude of factors including aquifer heterogeneity, presence of low permeability layers, natural gas leakage rates and depth of release, and the attenuation capacity for natural gas in the aquifer (Molofsky et al., 2021), including subsurface microbial CH 4 oxidation (Lyman et al., 2017(Lyman et al., , 2020Schout et al., 2019). While CH 4 dissolution can result in a significant amount of gas retention in an aquifer (Chao et al., 2021;Schout et al., 2020), degassing and free phase GM can allow gas to escape into the vadose zone and subsequently into the atmosphere (Cahill et al., 2017;Van De Ven & Mumford, 2020). In some cases, vadose zone GM has been identified with no or limited CH 4 emissions detected at the ground surface (Lyman et al., 2017;McMahon et al., 2018;Schout et al., 2019), indicating a high potential for attenuation. ...
Full-text available
Plain Language Summary Natural gas leaking from imperfectly sealed oil and gas wells can lead to explosive conditions in soil gas and methane emissions to the atmosphere. To understand the impacts and improve monitoring practices of fugitive gas migration, there is a need to characterize processes that control gas transport and fate in the unsaturated zone. We simulated subsurface wellbore leakage by injecting natural gas into thick unsaturated deposits in a region of petroleum development in Western Canada. The response was monitored by measuring methane and carbon dioxide concentrations and emissions on the ground surface and in soil gas in the unsaturated zone. Our results show that numerous compounding processes influence methane concentrations in the unsaturated zone and emissions to the atmosphere. While microbially mediated reactions consume methane in the unsaturated zone, variations in soil grain sizes and changes in barometric pressure strongly influence gas transport, which can lead to high episodic emissions in unpredictable locations. To accurately detect, quantify and assess fugitive gas migration at oil and gas well sites, adequate site characterization and continuous, spatially dense monitoring are necessary.
... CH 4 is nontoxic but can pose an explosion hazard above 10 mg/L. 2,3 Additionally, CH 4 migration can stimulate the growth of methanotrophic microorganisms 61,62 and subsequently produce reducing conditions in aquifers that solubilize iron, manganese, and toxic trace elements such as arsenic. 3 Thus, we calculated the highest expected increase in [CH 4 ] associated with UOGD in hotspots. ...
Unconventional oil and gas development (UOGD) sometimes impacts water resources, including incidents of methane (CH4) migration from compromised wells and spills that degrade water with salts, organics, and metals. We hypothesized that contamination may be more common where UOGD overlaps with legacy coal, oil, and gas extraction. We tested this hypothesis on ∼7000 groundwater analyses from the largest U.S. shale gas play (Marcellus), using data mining techniques to explore UOGD contamination frequency. Corroborating the hypothesis, we discovered small, statistically significant regional correlations between groundwater chloride concentrations ([Cl]) and UOGD proximity and density where legacy extraction was extremely dense (southwestern Pennsylvania (SWPA)) but no such correlations where it was minimal (northeastern Pennsylvania). On the other hand, legacy extraction of shallow gas in SWPA may have lessened today's gas leakage, as no regional correlation was detected for [CH4] in SWPA. We identify hotspots where [Cl] and [CH4] increase by 3.6 and 3.0 mg/L, respectively, per UOG well drilled in SWPA. If the [Cl] correlations document contamination via brines leaked from wellbores, impoundments, or spills, we calculate that thallium concentrations could exceed EPA limits in the most densely developed hotspots, thus posing a potential human health risk.
... Stray gas migration (GM), the unintended release of natural gas in the subsurface resulting from operations and infrastructure for energy development, is a significant environmental concern that can lead to safety risks, potential reduction of groundwater quality, and greenhouse gas emissions (Brantley et al. 2014;Cahill et al. 2017; Council of Canadian Academies 2014; Forde et al. 2019;Kelly et al. 1985;US EPA 2016). Environmental monitoring for GM at suspected or impacted sites most commonly relies on groundwater monitoring, where aqueous samples are collected and analyzed for dissolved gas concentrations of common components of natural gas (e.g., methane [C1], ethane [C2], propane [C3] and heavier hydrocarbon gases), as well as other potential indicators of impacted groundwater including major ions, redox sensitive species and total dissolved gas pressure. ...
Full-text available
Total dissolved gas pressure (PTDG) measurements are useful to measure accurate in‐situ dissolved gas concentrations in groundwater, but challenged by in‐well degassing. Although in‐well degassing has been widely observed, its cause(s) are not clear. We investigated the mechanism(s) by which gas‐charged groundwater in a recently pumped well becomes degassed. Vertical PTDG and dissolved gas concentration profiles were monitored in the standing water column (SWC) of a groundwater well screened in a gas‐charged aquifer for seven days before and fifteen days after pumping. Prior to pumping, PTDG values remained relatively constant and below calculated bubbling pressure (PBUB) at all depths. In contrast, significant increases in PTDG were observed at all depths after pumping was initiated, as fresh groundwater with elevated in‐situ PTDG values was pumped through the well screen. After pumping ceased, PTDG values decreased to below PBUB at all depths over the fifteen‐day post‐pumping period, indicating well degassing was active over this time frame. Vertical profiles of estimated dissolved gas concentrations before and after pumping provided insight into the mechanism(s) by which in‐well degassing occurred in the SWC. During both monitoring periods, downward mixing of dominant atmospheric and/or tracer gases, and upwards mixing of dominant groundwater gases were observed in the SWC. The key mechanisms responsible for in‐well degassing were i) bubble exsolution when PTDG exceeded PBUB as gas‐charged well water moves upwards in the SWC during recovery (i.e., hydraulic gradient driven convection), ii) micro‐advection caused by the upward migration of bubbles under buoyancy, and iii) long‐term, thermally‐driven vertical convection. This article is protected by copyright. All rights reserved.
Full-text available
Petroleum resource development generates a legacy of energy wells that must be decommissioned effectively as we transition towards NetZero. Unfortunately, some decommissioned wells (DWs) can suffer integrity failure resulting in release of fugitive natural gas into the surrounding soils and atmosphere. After decommissioning there are typically no ongoing assessments to confirm well integrity, meaning integrity status remains uncertain into the future. Furthermore, factors affecting fugitive natural gas migration in surficial soils around DWs are poorly recognized, inferring integrity assessment and monitoring strategies are lacking. To better understand the integrity status of DWs, identify soil properties controlling fugitive gas migration and help develop more effective monitoring and detection methodologies, we undertook field investigations at six DWs in England involving surficial CH4 measurements, sub-surface soil-gas and sediment sampling and dynamic flux-chamber measurements. We found no evidence of integrity failure at any site. However, the composition and structure of soils in which examined DWs are embedded suggest fugitive gas migration to surface may be severely limited; potentially mitigating methane emissions while making integrity assessment at the surface challenging. Overall, the integrity status of DWs in England is poorly constrained and we show surficial soil properties must be characterised and considered to effectively constrain the fate of fugitive gas and in order to design effective field assessment and monitoring methods.
Full-text available
We report observations on the dynamics of bacterial communities in response to methane stimulus in laboratory microcosm incubations prepared with lake sediment samples. We first measured taxonomic compositions of long-term enrichment cultures and determined that, although dominated by Methylococcaceae types, these cultures also contained accompanying types belonging to a limited number of bacterial taxa, methylotrophs and non-methylotrophs. We then followed the short-term community dynamics, in two oxygen tension regimens (150 μM and 15 μM), observing rapid loss of species diversity. In all microcosms, a single type of Methylobacter represented the major methane-oxidizing partner. The accompanying members of the communities revealed different trajectories in response to different oxygen tensions, with Methylotenera species being the early responders to methane stimulus under both conditions. The communities in both conditions were convergent in terms of their assemblage, suggesting selection for specific taxa. Our results support prior observations from metagenomics on distribution of carbon from methane among diverse bacterial populations and further suggest that communities are likely responsible for methane cycling, rather than a single type of microbe.The ISME Journal advance online publication, 21 October 2014; doi:10.1038/ismej.2014.203.
Three-dimensional numerical simulations are used to provide insight into the behavior of methane as it migrates from a leaky decommissioned hydrocarbon well into a shallow aquifer. The conceptual model includes gas-phase migration from a leaky well, dissolution into groundwater, advective-dispersive transport and biodegradation of the dissolved methane plume. Gas-phase migration is simulated using the DuMux multiphase simulator, while transport and fate of the dissolved phase is simulated using the BIONAPL/3D reactive transport model. Methane behavior is simulated for two conceptual models: first in a shallow confined aquifer containing a decommissioned leaky well based on a monitored field site near Lindbergh, Alberta, Canada, and secondly on a representative unconfined aquifer based loosely on the Borden, Ontario, field site. The simulations show that the Lindbergh site confined aquifer data are generally consistent with a 2 year methane leak of 2–20 m3/d, assuming anaerobic (sulfate-reducing) methane oxidation and with maximum oxidation rates of 1 × 10−5 to 1 × 10−3 kg/m3/d. Under the highest oxidation rate, dissolved methane decreased from solubility (110 mg/L) to the threshold concentration of 10 mg/L within 5 years. In the unconfined case with the same leakage rate, including both aerobic and anaerobic methane oxidation, the methane plume was less extensive compared to the confined aquifer scenarios. Unconfined aquifers may therefore be less vulnerable to impacts from methane leaks along decommissioned wells. At other potential leakage sites, site-specific data on the natural background geochemistry would be necessary to make reliable predictions on the fate of methane in groundwater.
To assess potential future impacts on shallow aquifers by leakage of natural gas from unconventional energy resource development it is essential to establish a reliable baseline. Occurrence of methane in shallow groundwater in Alberta between 2006 and 2014 was assessed and was ubiquitous in 186 sampled monitoring wells. Free and dissolved gas sampling and measurement approaches yielded comparable results with low methane concentrations in shallow groundwater, but in 28 samples from 21 wells methane exceeded 10mg/L in dissolved gas and 300,000ppmv in free gas. Methane concentrations in free and dissolved gas samples were found to increase with well depth and were especially elevated in groundwater obtained from aquifers containing coal seams and shale units. Carbon isotope ratios of methane averaged -69.7±11.1‰ (n=63) in free gas and -65.6±8.9‰ (n=26) in dissolved gas. δ(13)C values were not found to vary with well depth or lithology indicating that methane in Alberta groundwater was derived from a similar source. The low δ(13)C values in concert with average δ(2)HCH4 values of -289±44‰ (n=45) suggest that most methane was of biogenic origin predominantly generated via CO2 reduction. This interpretation is confirmed by dryness parameters typically >500 due to only small amounts of ethane and a lack of propane in most samples. Comparison with mud gas profile carbon isotope data revealed that methane in the investigated shallow groundwater in Alberta is isotopically similar to hydrocarbon gases found in 100-250meter depths in the WCSB and is currently not sourced from thermogenic hydrocarbon occurrences in deeper portions of the basin. The chemical and isotopic data for methane gas samples obtained from Alberta groundwater provide an excellent baseline against which potential future impact of deeper stray gases on shallow aquifers can be assessed.
Methane and brine leakage rates and associated time scales along the cemented casing of a hypothetical decommissioned shale-gas well have been assessed with a multi-phase flow and multi-component numerical model. The conceptual model used for the simulations assumes that the target shale formation is 200 m thick, overlain by a 750 m thick caprock, which is in turn overlain by a 50 m thick surficial sand aquifer, the 1000 m geological sequence being intersected by a fully-penetrating borehole. This succession of geological units is representative of the region targeted for shale gas exploration in the St. Lawrence Lowlands (Québec, Canada). The simulations aimed at assessing the impact of well casing cementation quality on methane and brine leakage at the base of a surficial aquifer. The leakage of fluids can subsequently lead to the contamination of groundwater resources and/or, in the case of methane migration to ground surface, to an increase in greenhouse-gas emissions. The minimum reported surface casing vent flow (measured at ground level) for shale gas wells in Quebec (0.01 m3/day) is used as a reference to evaluate the impact of well casing cementation quality on methane and brine migration. The simulations suggest that an adequately cemented borehole (with a casing annulus permeability kc ≤ 1 mD) can prevent methane and brine leakage over a time scale of up to 100 years. However, a poorly cemented borehole (kc ≥ 10 mD) could yield methane leakage rates at the base of an aquifer ranging from 0.04 m³/day to more than 100 m³/day, depending on the permeability of the target shale gas formation after abandonment and on the quantity of mobile gas in the formation. These values are compatible with surface casing vent flows reported for shale gas wells in the St. Lawrence Lowlands (Quebec, Canada). The simulated travel time of methane from the target shale formation to the surficial aquifer is between a few months and 30 years, depending on cementation quality and hydrodynamic properties of the casing annulus. Simulated long-term brine leakage rates after 100 years for poorly cemented boreholes are on the order of 10−5 m3/day (10 ml/day) to 10−3 m3/day (1 l/day). Based on scoping calculations with a well-mixed aquifer model, these rates are unlikely to have a major impact on groundwater quality in a confined aquifer since they would only increase the chloride concentration by 1 mg/l, which is significantly below the commonly recommended aesthetic objective of 250 mg/l for chloride. This article is protected by copyright. All rights reserved.
The environmental impacts of shale-gas development on water resources, including methane migration to shallow groundwater, have been difficult to assess. Monitoring around gas wells is generally limited to domestic water-supply wells, which often are not situated along predominant groundwater flow paths. A new concept is tested here: combining stream hydrocarbon and noble-gas measurements with reach mass-balance modeling to estimate thermogenic methane concentrations and fluxes in groundwater discharging to streams and to constrain methane sources. In the Marcellus Formation shale-gas play of northern Pennsylvania (U.S.A.), we sampled methane in 15 streams as a reconnaissance tool to locate methane-laden groundwater discharge: concentrations up to 69 µg L-1 were observed, with four streams ≥5 µg L-1. Geochemical analyses of water from one stream with high methane (Sugar Run, Lycoming County) were consistent with Middle Devonian gases. After sampling was completed, we learned of a state regulator investigation of stray-gas migration from a nearby Marcellus Formation gas well. Modeling indicates a groundwater thermogenic methane flux of about 0.5 kg d-1 discharging into Sugar Run, possibly from this fugitive gas source. Since flow paths often coalesce into gaining streams, stream methane monitoring provides the first watershed-scale method to assess groundwater contamination from shale-gas development.
Recent studies in northeastern Pennsylvania report higher concentrations of dissolved methane in domestic water wells associated with proximity to nearby gas-producing wells.1,2 We test this possible association by using Chesapeake Energy's baseline dataset of over 11,300 dissolved methane analyses from domestic water wells, densely arrayed in Bradford and nearby counties (Pennsylvania), and near 661 pre-existing oil and gas wells. The majority of these, 92%, were unconventional wells, drilled with horizontal legs and hydraulically fractured. Our dataset is hundreds of times larger than datasets used in prior studies. In contrast to prior findings, we found no statistically significant relationship between dissolved methane concentrations in groundwater from domestic water wells and proximity to pre-existing oil or gas wells. Previous analyses used small sample sets compared to the population of domestic wells available, which may explain the difference in prior findings compared to ours.